Proving Ground V7 — Manual

Import a Rhino model, read the verdicts, calibrate the design

The student manual for Proving Ground V7 (Four rectangles). It walks the work in order — prepare a Rhino model, import it, check the import, set the site, read the verdicts, calibrate — and then explains every number the tool shows: every verdict rule, every force sheet, all 52 Readout keys, energy and carbon, goals, and what each reading is and is not good for.

The reference tables are drawn from the tool's own code, so they say what V7 actually computes today.

1Start here

What V7 is, how to open it, and the one rule for reading it.

Proving Ground is a massing bench. You give it a simple building — a plinth, a room, a roof, openings — and it answers, live, what the sun, the wind, the mass and the view are doing to it. V7 · Four rectangles is the version module 2 uses: it opens on the 4 Rectangles game (plinth, room, roof, and the fourth rectangle — the focus, a table or a seating pit) and reads the sun on the focus and the view from a seat at it.

  1. Open it with the link: toolkit.allmeans.works/tools-3d/crucible#v7. The plain Proving Ground link opens V7 too (since 2026-09-30), so either address works. You are in V7 when the header reads V7 · Four rectangles.
  2. No account is needed. Every reading is computed in your browser and your Rhino file is never uploaded. You need to sign in only to load a coordinate you saved in the Surveyor, or to brainstorm goals with Claude.
  3. There are two ways to work. With the sliders: the 4 Rectangles jig, where every length is in feet. With your own Rhino model: import it, read it, change it in Rhino, and import it again. Everything below works for both except where it says jig only.
The one rule

Trust the comparison, not the absolute. Every number here is one stated calculation run over the shape you drew — direct sun from a clear sky, one room, one fixed wall build-up. It is very good at saying which of two schemes is better, and why. It is not an energy model, a climate-based daylight study (the Daylight box is a daylight-factor screen under an overcast sky; Daylighter is the study) or a structural check, and its absolute figures are magnitudes, not facts to put in front of a client. Section 15 says, force by force, what each number leaves out.

2Prepare your Rhino model

The engine scores what is MODELLED. It cannot see a window in a hole, a focus in a slab that happens to be higher, or a wall's thickness in one surface. Get these right in Rhino and the import needs no fixing.

The layers — the role is the layer

Make these 9 top-level layers and model on sub-layers inside them (name the sub-layers anything). This is the same layer tree Proving Ground's own Rhino export writes, so a model can go back and forth without re-mapping. A sorting prefix is fine — _Roof, 01_Walls and 1. Plinth all read — but the group word has to be at the top of the tree: under a parent layer such as Project::Roof, it is not.

The top-level layers Proving Ground reads, what goes on each, and what the engine does with it
LayerEngine roleWhat goes on itWhat it does to the numbers
RoofroofThe roof — slabs, pitches, the ridge.Envelope: counts in skin area, heat loss and roof carbon; shades the sun.
WallswallThe walls, with the openings TRIMMED out of them.Envelope: skin area, heat loss, wall carbon; shades the sun; takes the wind.
PlinthfloorThe plinth as a CLOSED solid (and any floor slab).The floor the sun patches land on, the mass, and slab carbon (priced by volume).
FocusfloorThe fourth rectangle only — the table top block or the seating-pit floor. Put “table” or “pit” in the layer or object name.Plinth material and mass; the plane “sun on the focus” is measured on; the seat every Readout view is read from.
GlazingglazingOne flat surface in every opening, one per wall face — glass or not.Becomes an opening. Every sun, air, view and glass number reads from these and nothing else.
GroundgroundTerrain you model yourself, as a closed solid.Shades the sun; not envelope. Closed solids here add to mass volume.
ShadeshadeLouvers, fins, brise-soleil, awnings, free-standing overhangs.Shades the sun; not envelope, no carbon.
StructurestructureRafters, beams, joists, columns, trusses.Shades the sun; not envelope, so a rafter is never charged as roof carbon twice.
Notationsnone — not modelledText, dimensions, grids, anything that is not the building.Left out of the model entirely.

A layer outside these names gets its role guessed from words in its name, and a name the guesser does not know becomes opaque wall. So Hearth is read as wall, Door as wall, Fins as wall. Use the layers above and nothing is guessed.

The rules, in the order they bite

  1. Every opening is a flat surface on the Glazing layer — and the hole is trimmed in the wall as well. The engine reads an opening ONLY from a Glazing surface; a hole trimmed out of a wall is just a smaller wall, and every sun, air, view and glass number reads 0. Put one flat surface in each hole, one per wall face (a corner window is two surfaces). Trim the wall behind it too: an untrimmed wall blocks the sun rays — sun on the focus reads 0 with no warning, and if the glass sits inside the wall's thickness, the window's own sun readings drop as well.
  2. An opening with no glass is modelled the same way. Draw the surface, then either name it with (open), unglazed, no glass or open air in Rhino, or switch it to Open — no glass in the Openings — glazed or open group after import. Naming it in Rhino is safer: it stays open every time you re-import.
  3. Say how each window opens. Every air number counts only the part of an opening that opens — a fixed light moves no air, a sliding sash at most half. Put the sash type in the object's name in Rhino (fixed, casement, sliding, double-hung, awning, hopper, louvre, jalousie, tilt-only, trickle), or set opens as for each opening in the Openings — glazed or open group after import. Unset, an opening counts as wholly open. Sun, view, glass area and carbon keep the whole pane either way (table in section 10).
  4. Give each pane its own object name (in Rhino's Properties). Openings on one layer are told apart by object name; without one they are numbered in file order, and a re-import can shuffle which pane is which.
  5. A gap is not an opening. A wall that stops short of a corner, or floats off the plinth, lets nothing through as far as the engine is concerned. If a gap should count as a doorway, draw a Glazing surface across it and make it open. The drawn air closes a seam under 10 in (a wall a little short of the roof or the plinth) as a modelling slip, the same at every Quality; a wider gap reads as a hole, and the room is then not drawn.
  6. The plinth is a CLOSED solid. Mass volume, thermal mass and slab carbon measure volume. A plinth drawn as one surface reads 0 (flagged ⚑ no mass solid) — or, if it floats above the model's lowest point, a wrong number with no flag. Extrude it. Its top face must point up.
  7. Walls and roof: thin closed solids, not single surfaces. Sun, wind, air and views read a single surface correctly, but skin area, heat loss and carbon halve the summed face area (the engine assumes thin closed boxes), so a single-surface wall counts at half its real area.
  8. The focus is its own object on the Focus layer. Keep the word focus off every other layer and object. Put table or pit in its name, or the tool decides by height: at or above the plinth top is a table, below is a pit. Model a table as a block standing on the plinth. Model a pit as the pit floor surface (pointing up), and cut the pit out of the plinth solid — otherwise the plinth sits over the pit floor and shades it.
  9. Units and position. Set the Rhino file's units (feet or inches are fine — they are read from the file). Model near the origin with +Y as north and +X as east: the model keeps its Rhino position, and Rotate the building turns it about the origin.
  10. Save with render meshes. View the model shaded once before saving, and do not use SaveSmall. Without render meshes the importer rebuilds surfaces itself, and a trimmed face comes back UNTRIMMED — the window holes fill in and a thick glass box becomes six openings.
  11. Clean the file. Hidden and locked layers are imported too — delete what is not the building, or put it on Notations. Explode blocks. Curves, text, dimensions and hatches are ignored.
Checklist before you export from Rhino
  • Top-level layers are the 9 above; nothing important sits on a hidden layer.
  • Every opening: a hole trimmed in the wall AND a flat surface on Glazing, one per wall face, each with its own object name.
  • Open (unglazed) openings named (open); windows named by how they open (fixed, casement, sliding…).
  • Plinth, walls and roof are closed solids; the plinth's top faces up.
  • The focus is one object on Focus, named table or pit; a pit is cut out of the plinth.
  • Units set; model near 0,0 with +Y north; viewed shaded, then saved.

3Import it

From the Rhino file to a scored model in about a minute.

  1. Open V7. In the left rail, under The massing, press ⤓ Import your Rhino model — or drop the file straight onto the drawing. It reads .3dm, .obj, .stl, .ply and .glb (a .3dm is best: it carries its layers and units). Only the first file of a multi-file drop is read.
  2. In the import panel, check Units & orientation. Model units is read from a .3dm (metres, centimetres, millimetres, feet, inches); any other file starts at metres, so set it. The Extent line under it gives the model's size east–west, north–south and in height, in feet: if the height is ten or twelve times wrong, change the units; if the model lies on its side, tick Model is Y-up (a .3dm never needs it).
  3. Read the Layers → roles table. Every layer has a row, a triangle count and a role you can change. ▣ means the role was read from your layer tree; ◇ means it was guessed from the name — check every ◇ row. The line under the table counts how many objects each way claimed.
  4. Read the report: N parts · N triangles · N openings read. If openings read is 0, stop and go back to Rhino (rule 1 above). Open What this import can’t tell you for the list of what this import cannot measure.

What changes when you import

  • Your model replaces the 4 Rectangles jig and its sliders disappear. There is nothing to drag — you change the design in Rhino and import again. Back to the program brings the jig back with the slider values you left.
  • Still working on an import: every reading, verdict and sheet; the site (location, weather, landform); Rotate the building; the Openings — glazed or open switches (glazed or open, and how far each opens); the site dials (wind, temperatures, view target, material); the Readout; goals (the cards score live); Save this study; the change ledger; exports.
  • Not on an import (jig only): Sweep a dial and Improve toward goals (they move sliders, and an import has none); Start from; the roof warnings; and the surfaces Rhino export (an import can go back out as a mesh only).
  • The model is set down on the ground: its lowest point is moved to z = 0 — unless you ticked Use it as the site’s landform, instead of the location’s, when it stands on the file's own terrain where you modelled it (section 6). That is why buried fraction and soil contact read 0 on an import. Rotate the building, the site, and your glazed/open and opens-as choices are all kept when you import a new version (they follow each opening's name).

4Check the import

Before you read a single verdict, make sure the zeros are design facts and not modelling gaps.

The ⚑ flag. In the Readout, a number that reads 0 or “—” because the model lacks what it reads from carries a small flag beside it. Hover it, or reach it with the keyboard, for the sentence that says what to model. A 0 with no flag is a measurement.

The Readout's modelling-gap flags: the chip, what it means, and which rows carry it
FlagWhat the tool saysRows it appears on
⚑ no openingsNot a design fact: no opening was modelled, so this reads 0. The engine reads an opening ONLY from a surface on the Glazing layer (or one named glazing / glass / window); a hole trimmed out of a wall is a smaller wall. Model a flat surface in each hole — one per wall face — on the Glazing layer; if there is no glass in it, switch it to OPEN in the Openings group (V6, V7) or put 'unglazed' in its name.Winter solar gain · Summer solar gain · Overheating ratio · Winter gain · self-shaded · Summer gain · self-shaded · Winter sun kept · Summer sun cut · Equator-facing gain · Winter sun on the floor · Summer sun on the floor · Winter sun on the focus · Summer sun on the focus · Ventilation provision · Stack ventilation · Stack height · Air path you can't shut · Air changes you can't shut · Solar fraction · Ventilation relief · Air changes · View to target · View openness · Sky view · Opening area · Glass area · Open (no glass) area · Glazing ratio · Glazing carbon
⚑ no floorNothing to read: there is no up-facing floor surface to sample. Model the plinth on the Plinth layer (a surface or, better, a closed solid) and it becomes the plan this is read on.Winter sun on the floor · Summer sun on the floor
⚑ no focusNothing to read: no focus was found. Put the fourth rectangle — the table top or the seating-pit floor — on a layer called Focus (or name the solid with the word), and this reads from its own plane.Winter sun on the focus · Summer sun on the focus
⚑ no mass solidNot a design fact: no CLOSED floor or ground solid, so the mass volume is 0. Thermal mass integrates volume, and a plinth drawn as a single surface has none — model it as a closed solid (an extrusion or a box).Mass volume · Thermal mass · Slab carbon
⚑ no weather fileNo weather loaded, so this cannot be counted. Load a year of weather — search a place, pick a studio site, or drop an .epw weather file, whichever this version offers — and it is counted from that year's 8,760 hours.Passive comfort (year) · HVAC hours (year)

When the numbers look wrong

Symptoms after an import, their causes, and the fix
You seeWhyFix
0 openings read; the opening rows flagged ⚑ no openings; a Heads up banner in the ReadoutNo surface on the Glazing layer — the holes are only trimmed in the walls.A flat surface in each hole on Glazing, one per wall face.
Openings read, but sun on the focus is 0 (and winter gain low) — no flagThe wall behind the glass was not trimmed, so it blocks the sun rays — always for the focus, and for the window itself when the glass sits inside the wall's thickness.Trim the hole in the wall too.
Opening area about double, with thin extra openingsA thick glass box, saved without render meshes, read one face at a time.View shaded and save again, or model glass as single flat surfaces.
Window holes filled in; a note says trimmed faces were read as their full patchThe file was saved without render meshes.View the model shaded in Rhino, then save (not SaveSmall).
Rows marked ◇, and the building reads as all wallLayers outside the contract (Layer 01, Default…).Rename the top-level layers, or set each ◇ row's role.
Sun on the focus flagged ⚑ no focusNo Focus layer or name, the focus joined to the plinth, or its top face pointing down.A separate object on Focus, top face up.
Sun on the floor flagged ⚑ no floor; footprint equals the bounding boxNo up-facing plinth surface (missing, reversed, or guessed as wall).Plinth layer; check its direction (Dir), or make it a closed solid.
Mass volume, thermal mass and slab carbon flagged ⚑ no mass solidThe plinth is a single surface.A closed solid (Extrude, Box, Cap).
Mass volume implausibly large, no flagOpen floor or ground surfaces above the model's lowest point.Closed solids only on Plinth, Focus and Ground.
The model lies on its sideA Y-up file (.obj, .glb from other software).Tick Model is Y-up.
The Extent is 3.3×, 12×, 25× or 100× wrongWrong or missing units (the default is metres).Change Model units.
The model is off the landform, or swings wide when rotatedIt was modelled far from Rhino's origin.Move it to 0,0 in Rhino and re-import.

5Find your way around

On a wide screen V7 is three columns: the rail on the left, the model in the middle, the readings on the right.

Left rail — the inputs

  • The massing — Start from (4 Rectangles in V7), ⤓ Import your Rhino model, the jig's sliders (or, after an import, the import panel).
  • Openings — glazed or open — each opening Glazed — a pane or Open — no glass, and how far it opens as (whole, casement, sliding, fixed…), with the total Area that opens.
  • The ground — Where it stands — your coordinate or a studio site (the five studio sites, Load coordinate — saved in the Surveyor, a pasted coordinate, and the key map), Ground extent, Seat the landform and Rotate the building.
  • Sun — a place search, Or drop a weather file (.epw), Latitude, and Sweep a dial (jig only).
  • Air — Prevailing wind from, Wind speed, Inside − outside ΔT. Mass — the material, Day–night outdoor swing. View — View target, Eye height above the floor. Energy — the design days and Internal gains.

Middle — the model, and everything under it

  • Above the model: Save this study, and once you have studies, ◀ Prev study / Next study ▶.
  • The model stays on screen as you scroll. Bottom right: 3D · Plan · N · E · S · W, parallel or perspective (in perspective, 2-pt keeps verticals vertical — a two-point perspective), ⟳ Turntable (just left of Parallel: the view turns round the model by itself, once every 24 seconds, until you press it again — a drag holds it and sets the height it turns at), ⊕ Orbit: model and ⌖ Recenter the view. A left-drag turns the model about its own centre; double-click a point (or press the orbit button and click) to turn about that point instead. Right-drag pans; the wheel zooms toward the cursor. Top left, only when the roof breaks the rules: ⚠ The roof breaks the game (in V7 the ridge governs, so this warns in feet when the eave passes below the ground or the plinth, or leaves less than 6′-6″ of headroom).
  • Under the model, in this order: Saved views · Date + time · Viewport display · the Water and Daylight boxes · Studies (after your first study) · Verdicts · then the force pills Sun · Air · Mass · View · Shadows · Year · Climate · Earth · Balance with the sheet for the pill you pick (⤓ SVG and ⤓ PNG export it).

Right column — four lenses

  • Site forces (opens first): What just changed, The fourth rectangle — the focus, What each reading is good for — how far to trust it, and when to use a professional tool, then one panel per force.
  • Energy & carbon: the winter and summer energy balance, carbon, wind load, earth and form.
  • Readout: all 52 keys, with Show methods.
  • Goals: turn an intention into a testable goal, and (jig only) Improve toward goals.

Hover or keyboard-focus any number for how it was computed and where its data comes from — every stat, verdict card and Readout row carries both. On a phone the model comes first, then the rail, then the readings.

6Set the site first

Calibrating against the default site calibrates against nowhere. Set the place before you judge anything — it changes the sun, the wind, the temperatures and which verdicts appear at all.

V7 opens on a placeholder site. Until you change it, these are the numbers every reading uses:

The site V7 opens with
SettingOpens at
Latitude42° N — no place, no weather file, flat ground
NorthFixed: +Y is true north. Turn the building with Rotate the building, not the north arrow.
Prevailing wind from270° (west), 11 mph
Inside − outside ΔT10.8 °F — drives the stack
View target / eye height180° (south) / 5′-3″ standing
Winter / summer design day32 °F / 82 °F
Day–night outdoor swing18 °F
Mass materialDense concrete
Date + timeMar 20, 2 pm sun time — the drawing only (see below)

Give it a real place

  1. A studio site: under Where it stands — your coordinate or a studio site, press Tar Creek, Butte, Onondaga Lake, Colorado Smelter or Palmerton. It moves the latitude, brings a measured year of weather (and the wind from it), and fetches the landform.
  2. Your own spot: save it in the Surveyor with ★ Save coordinate, then here pick it under Load coordinate — saved in the Surveyor and press Load (signed in). Or paste a latitude and longitude into the box and press Go.
  3. Anywhere: type a place into the Sun group's search, or drop an .epw weather file. A dropped .epw sets the latitude, the design days, the swing and the wind — its prevailing direction and average speed, into Prevailing wind from and the speed dial (the Air sheet's rose shows all of the site's winds).
  4. Then walk the ground: drag the box on the key map, or use the shift sliders. The landform is fetched again where you put it; the weather is not (a mile is inside one weather cell).

The ground — two ways

The box at the top of The ground, Where the ground comes from — two ways, switches between them. Either way, set the location as above too: it sets the sun, the weather and the wind, and a Rhino file carries no coordinates.

  1. ① The location (the default, and how V7 opens): the landform is fetched around the place you set — USGS 3DEP where it is measured. Show more ground (the line under the picker says this too): pick a wider square in Ground extent — 200 ft, 400 ft, 800 ft, ¼ mi (how V7 opens) or ½ mi. The landform, and the aerial photo or topo map on it, are fetched again around the site. Wider takes in the ridge that shades you; the patch is cut from the measured grid at up to 97 samples a side, so at ½ mi it is as coarse as the 3DEP source (about 26 ft) and a wider square is coarser under the slab.
  2. ② Your own ground, from Rhino. Say you exported the site from the Surveyor (its Rhino kit), trimmed out the piece around your site in Rhino, and did your 4 Rectangles studies on it. Press ⤓ Import ground from Rhino and pick the file. The layers that read as ground are ticked for you — the Surveyor's Terrain, or a Ground layer — and you can tick or untick any. Check the units (the Surveyor's kit is in feet, and says so). The jig (or whatever design is on the bench) stays, and now stands on your ground, centred on it — or, if you drew your rectangles on the Plinth / Focus / Walls / Roof layers in the same file, centred where you drew them. Remove the Rhino ground takes it away; the ② ① switch goes back and forth without re-reading anything.
  3. Your building AND its ground, from one file. Import the whole model (⤓ Import your Rhino model) and, under The ground in this file, tick Use it as the site’s landform, instead of the location’s. The building then stands on the terrain exactly where you modelled it — the same height, the same slope under it — instead of being set down at z = 0 on its own. It is off unless you tick it.
What a Rhino ground can and cannot do

It is in the score exactly like a fetched landform: it shades every sun ray, the Earth card reads its slope, cut and fill and water, and Seat the landform moves it up or down (0 = where you modelled it). The tool reads the largest SQUARE your file covers everywhere (up to ½ mi), sampled about every foot on a small piece and every 28 ft or so on a big one; it never fills in ground that is not there — a piece too small, or not reaching round the building, is refused and the message says by how much. The top surface wins: anything under an overhang is lost.

It has no coordinates, so the aerial photo, the topo map and the Surveyor’s map layers are not drawn on it, and Ground extent does not change it: to show more ground, trim a bigger piece in Rhino and import it again (a smaller piece for less).

Trimming the Surveyor’s surfaces kit? Its terrain is one NURBS surface. After you Split or Trim it, view it shaded once and save normally (not SaveSmall): a trimmed surface saved with no render mesh is read as the WHOLE surface it was cut from, and the box says so under the file name. The Surveyor’s mesh kit has no such catch.

What the site unlocks

Without a weather file there are six verdict cards (Sun, Air, Mass, View, Energy, Carbon). A weather file adds Year and Climate, the ventilation year and the measured-climate Readout rows. A landform adds Earth, and puts the terrain in every sun ray.

The Date + time bar moves the sun in the drawing only. No verdict and no Readout number uses it: they all use the same fixed suns — Dec 21, the equinox and Jun 21, at 8, 10, 12, 2 and 4 o'clock sun time — and V7's drawn air reads the wind and ΔT dials until you switch it to the date.

Switch the drawn air to the date (Air → The drawn air follows: it starts on The dials below; pickThe date + hour). At the day and sun hour on the bar it reads the weather year's hours within a fortnight and an hour either side: their median outside air, the wind most of them blow from, and the room held at ASHRAE 55's adaptive comfort temperature — heated to 68 °F (20 °C) when the month before averaged under 50 °F (10 °C). So January draws a stronger stack and its own wind, and a hot July afternoon, outside warmer than the room, draws the stack running backwards — in at the high openings, out at the low ones. The air readout under the model then says what that air does to a person (a breeze worth a few degrees in a warm room, a draught by ISO 7730 in a cool one) and to the heat (what the openings carry out of a heated room). With no weather year the outside air is a placeholder curve through the Winter and Summer day dials. Switch it to the dials to draw the air at the design wind and ΔT instead.

7Read the verdicts

The verdict strip is the fastest honest read of a scheme: one card per force, worst first. Read it before you touch the Readout.

In V7 the strip sits under the model, in the box titled Verdicts. Its header counts the forces read and the warnings. Above the cards one sentence names what is fighting you — “One thing is fighting you: Air — …” — or says “Nothing is fighting you.”

The three verdict severities
MarkMeans
▲ warningThe scheme fails this force on the tool's teaching threshold. Fix it, or be able to say why not.
● noteWorth a look: a partial trade, a lightweight response, a fact about the place.
✓ goodMeets the teaching threshold. The number still matters — a ✓ is not a maximum.
  • Order. Warnings, then notes, then good. Within one mark the order is fixed — sun, air, mass, view, year, climate, earth, energy, carbon — so “the biggest two” in the headline are the first two in that order, not the two largest.
  • Each card shows the force, the mark, a headline, and a number line (imperial first, metric in parentheses).
  • Hover or keyboard-focus a card for four things: Why (the cause), Try (a move, when there is one), How this is computed, and What this leaves out. Screen readers get the same text.
  • Click a card to open that force's sheet under the strip. Energy and Carbon switch the right column to Energy & carbon instead.
  • No card for the focus, the seat, Shadows or Balance — the focus is read in its own panel (section 11), and Shadows and Balance are drawings.

The thresholds are teaching thresholds, not codes: a line in the sand that makes a trade visible. Section 9 lists every rule each card can fire.

8Calibrate: the loop

Calibrating a design means making one move at a time and reading what it did to every force — including the ones you were not aiming at.

The tools

  • What just changed (top of Site forces): after every move — one slider drag counts as one move, read once the numbers settle — it lists what moved, which way, and whether that is BETTER, WORSE or just CHANGED, with a sentence such as “Since Plinth width: winter kept down 6 points, air changes up 0.3 ACH.” A number is listed when it moves 2 % or more, or past its own threshold (table below). Percentages move in points.
  • Compare against: The last move (the default), or any saved study — the ledger then measures everything from that study, so you see the sum of all your moves since. Mark this as the baseline freezes the current design as Your mark and measures from there until you pick something else in Compare against.
  • Save this study (or Save, top right — the same thing): keeps the design, the site dials and all the ledger's numbers as a study, with thumbnails — and the whole bench it was looked at on: every viewport's display, the camera and its lens, the date and hour, and how the ground was drawn. Studies live in this browser, not your account: up to 36, the oldest dropped first, shared with V6. Clearing the browser's site data clears them — export what matters. A weather file you dropped is not stored; drop it again.
  • Saved views: Add view stores a camera and Hold keeps it fixed while the model changes, so every study is drawn from the same place.
  • Jig only: Sweep it — winter kept vs summer cut rebuilds the design at nine values of one dial and draws winter kept against summer cut on the Sun sheet, with the best trade ringed. Improve toward goals searches every slider toward your goals (section 14).

The loop, step by step

  1. Set the site (section 6) and the material. Everything after this is measured against that place.
  2. Fix the camera: Saved views → Add view, then Hold.
  3. Save the starting point: Save this study, and rename it in the Studies box (“01 as imported”).
  4. Read the strip worst first. For the top card: hover it for Why and Try, then click it and read its sheet. Check its What this leaves out line before you believe it.
  5. Make ONE move that answers it.
    • Jig: drag one slider — or sweep it first to find the value.
    • Rhino model: change it in Rhino, save, and drop the new file on the import panel. Your rotation, site and OPEN switches carry over.
    • Either: a site dial — Rotate the building, View target, the material, an opening glazed or open, or how far it opens.
  6. Read What just changed. Did the target number move the way you wanted? What else went WORSE? A move that fixes the sun and breaks the air is a trade, not a fix — decide it on purpose.
  7. For an imported model, compare against the saved study rather than the last move. A re-import changes everything at once, and the ledger's “Since …” names the last control you touched, which may be an older one.
  8. Keep what earns its place: Save this study again and name it after the move (“02 west fin”). Use ◀ Prev study / Next study ▶ to flip between them, and pick any of them in Compare against.
  9. Repeat until the warnings are gone or you can defend each one. Then check the survivors against section 15 before you present them.

What turns each card ✓ — the calibration targets

What each verdict card needs to read good
CardReads ✓ whenNeeds
SunSummer cut 40 % or more AND winter kept 80 % or more — and no skylight, side-facing or half-side-facing pane carrying half the summer sun.Always shown.
AirNo dead mechanism, 4 to 10 air changes an hour, the plan within reach of the air, and — with a weather file — purging at least half the mild hours and cross-flow working at least 40 % of the year.Always shown. Two of its rules need a weather file.
MassLess than half the mass inert and a decrement factor of 0.8 or less — a lag pinned at the 6-hour ceiling counts as good (the mass is heavy). Mass is never a warning.Always shown. It is never a warning.
ViewA pane faces the view, at least a quarter of the floor shares it, and the view sits less than 90° from the winter sun.Always shown.
YearThe month the massing shades most is a cooling month (Apr–Sep north of the equator), at 15 % or more, and every heating month under 15 %.A weather file (a studio site, a loaded coordinate, a searched place, or a dropped .epw).
ClimateNever — it describes the place, so it is always a note.A weather file. It is always a note: it describes the place, not your design.
EarthThe footprint on the grid, no channel or standing water under it, a slope of 15 % or less, and net earthwork within about 65 yd³.A landform — a studio site, a loaded or pasted coordinate, a searched place, or a Surveyor survey all bring one. On flat ground there is no Earth card.
EnergySummer float of 9 °F or less AND the sun paying back 20 % or more of the winter loss.Always shown.
CarbonCarbon per square foot of the room's own floor under the lower band (a ranking mark; it is not a pass against LETI, which counts more).Always shown.

The carbon card's lower and upper bands are 27.9 kgCO₂e/ft² (300 kgCO₂e/m²) and 55.7 kgCO₂e/ft² (600 kgCO₂e/m²), counting the making of materials only (A1–A3) — so they rank schemes; they do not rate them against a published target.

What the ledger tracks

A number is listed when it changes by 2 % or more of its own value, or by at least the amount in the last column — whichever comes first. A percentage moves in points: winter kept going from 95 % to 89 % is “down 6 points”.

The numbers What just changed compares, their better direction, and how far each must move to be listed
NumberUnitBetterOr when it moves at least
Winter kept%higher2 points
Summer cut%higher2 points
Air changesACHhigher0.2 ACH
Stack flowcfmhigher50 cfm
Cross flowcfmhigher50 cfm
Useful night-flush hourshhigher24 h
Open-window hours purgedhhigher24 h
Open-window hours stillhlower24 h
Working massft³higher20 ft³
Decrement factorratiolower0.02
Time laghhigher0.2 h
Floor sharing the view%higher2 points
Mean view qualitysrhigher0.02 sr
View against sun°lower5 °
Annual on the glasskWh/m²·yrneither — reads CHANGED10 kWh/m²·yr
Sun hours at the probehneither — reads CHANGED20 h
Cutyd³lower5 yd³
Fillyd³lower5 yd³
Slope under the footprint%lower1 points
Catchment arrivingft²lower500 ft²
Winter heat demandkBtu/daylower5 kBtu/day
Solar fraction%higher2 points
Summer float°Flower0.5 °F
Embodied carbonkgCO₂elower500 kgCO₂e
Carbon intensitykgCO₂e/ft²lower1 kgCO₂e/ft²

Numbers that need a weather file or a landform drop out of the comparison when either side lacks them. The focus readings are not in the ledger — watch them in the focus panel, or make them goals.

A worked example — the landing state

Open V7 fresh and read the strip: the Air card is a warning, “Cross-ventilation is dead at this wind.” Hover it: the wind comes from the west (270°) and no opening faces west, so there is no windward inlet — the air it counts is all stack, rising out of the skylight. Its Try line: turn the building, or put an opening on the windward face.

  1. Save the study as “01 landing”.
  2. Move one aperture onto the west wall (its position slider runs round the room, south → east → north → west), or turn the building so an opening faces west.
  3. Read What just changed: cross flow rises from zero; now see what it cost — winter kept, summer cut, winter heat demand — and whether the Air card flipped to a note about the rate instead.
  4. Save it as “02 west inlet”, then compare against “01 landing” to see the whole trade at once.

9Every verdict rule

Each card checks its rules top to bottom and shows the first that matches. The card's own wording is fuller; this is what triggers it.

Sun — Always shown. Clicking it opens the Sun sheet.

The Sun verdict's rules
MarkWhenThe card saysTry
● noteThe model has no openings.No openings, so no solar gain to read.Model the openings (on an import: a surface on the Glazing layer in every hole).
● noteNo sampled sun (8 am–4 pm on the two solstices) reaches any opening — north glass does this north of the equator.Steady, glare-free light but no free winter heat.Put glass on the equator-facing side if you want winter heat.
▲ warningOne skylight takes half or more of the summer sun.Names the skylight: it faces the high summer sun, and there is no wall above it to hang an overhang from.Make it smaller, turn it into a clerestory facing the equator, or put an outside louvre over it.
▲ warningOne pane takes half or more of the summer sun and faces east, west, north-east or north-west (east, west, south-east or south-west south of the equator).Names the pane: an overhang cannot fix a low morning or afternoon sun, which comes in under it.A vertical fin on that pane, less glass there, or turn it toward the equator.
▲ warningOne pane takes half or more of the summer sun and faces south-east or south-west (north-east or north-west south of the equator).An overhang only half-works there: it catches the midday part and misses the morning or afternoon part.A shallow overhang plus a vertical fin on the sunny side, or less glass there.
✓ goodSummer cut is 40 % or more AND winter kept is 80 % or more.The classic overhang trade: summer sun cut, winter sun kept.—
▲ warningWinter kept is under 50 % (a note instead when the weather year is cooling-dominated).Something — an overhang, a wing, a wall or, with ground loaded, the hill — is shading most of the winter sun.Shallow or lift the overhang; the overhang study shows where summer cut stops costing winter.
● noteSummer cut is under 15 % and more summer sun than winter sun reaches the glass.The summer sun is coming in almost unshaded.A horizontal overhang above the south glass.
● noteAnything else.A partial trade — some summer sun shaded, most winter sun through.Start with the pane that carries the most summer load (the card names it).

Air — Always shown. Two of its rules need a weather file. Clicking it opens the Air sheet.

The Air verdict's rules
MarkWhenThe card saysTry
● noteOne mechanism is dead (the wind cannot cross the room, or there is no stack) but the other alone still gives more than 1 air change an hour.Names the dead one and says the other is carrying the room — a note, not a warning.An opening on the windward or leeward face lets the wind help; a roof vent above the inlets buys the stack.
▲ warningA mechanism is dead and the other cannot make up for it: no openings; or (with two or more openings) none faces the wind, or none is on the leeward side; or the inlets and outlets are within about 1′-8″ of the same height; or there is no inside–outside temperature difference.Names what is dead and why, in the engine's own sentence — for example “Cross-ventilation is dead at this wind”.No windward opening: turn the building or put an opening on the windward face. No leeward opening: add one. No height difference: a roof vent or clerestory above the inlets.
● noteAir changes are over 10 an hour.The rate is a ceiling, not a rate: each opening is counted as open as its “opens as” setting says, and that is the whole opening unless you set it.Set how each window really opens in the Openings group (a sliding or double-hung sash is 50 %, a fixed light 0), then read it again.
▲ warningAir changes are 1 an hour or fewer.Fresh air only — no useful cooling.The card sizes it for you: how much must OPEN on the smaller side of the wind pair (or the stack pair) to flush the room at 4 air changes an hour — sometimes a sash that opens further, not a bigger window.
● noteThe plan is deeper, along the wind, than the air reaches: about 5 ceiling heights for cross-ventilation, 2½ for one-sided.The far side of the plan is still air, however good the rate reads.Shorten the plan along the wind, raise the ceiling, or add a roof outlet.
● noteWeather file: over the year's mild hours (68–79 °F outside, at least 200 of them), the openings reach 4 air changes an hour less than half the time.Purges only part of the year's open-window hours — the design-day rate is one good hour, not the year.More outlet area or height if the stack leads; a roof outlet for calm hours; otherwise open a second facade to the wind.
● noteMore than 1 and under 4 air changes an hour.Some fresh air, but not enough to flush out a hot day's heat (4–10 is the usual purge range).Enlarge the SMALLER opening of each pair, or add height between a low inlet and a high outlet.
● noteWeather file: the cross-flow works (at least half its best) for under 40 % of the year's hours.The scheme is tuned to one wind direction and the site's measured wind is somewhere else most of the year.Give the side the card names an opening too.
✓ goodAnything else: 4 to 10 air changes an hour with no other note.The rate, and what the stack (warm air rising) and the wind would each give on their own.—

Mass — Always shown. It is never a warning. Clicking it opens the Mass sheet.

The Mass verdict's rules
MarkWhenThe card saysTry
● noteHalf or more of the mass is inert — deeper than the daily heat can reach (about 5¾ in for dense concrete).That share of your mass is doing nothing on a daily cycle.Thin the slab toward the penetration depth and spread the mass over more exposed floor — surface stores heat, not thickness.
✓ goodThe time lag hits the model's 6-hour ceiling.Heavy mass: the model cannot lag any further, so read the decrement factor, which still carries information.—
● noteThe decrement factor is over 0.8 — more than 80 % of the outdoor swing gets inside.A lightweight response: the inside follows the outside.More EXPOSED mass, or a heavier material.
✓ goodAnything else.The mass damps the daily swing to the figure shown.—

View — Always shown. Clicking it opens the View sheet.

The View verdict's rules
MarkWhenThe card saysTry
● noteNo pane faces the View target from anywhere on the floor.No pane faces the view.Turn a pane toward the view, or add one on that face.
▲ warningLess than a quarter of the standing positions see at least 0.1 sr of glass toward the view.Only that share of the floor gets the view — either the view glass is too small or far, or most of the glass looks elsewhere.Enlarge or spread the glass on the view face, narrow the plan toward it, or turn a pane toward it.
● noteThe view and the winter sun are 90° or more apart.Every pane has to be argued for twice: glass toward the view is glass away from the sun.Keep the view glass modest and where it is used most; let the south glass do the heating.
✓ goodAnything else.The share of the floor that shares the view, and how far the view sits from the winter sun.—

Year — A weather file (a studio site, a loaded coordinate, a searched place, or a dropped .epw). Clicking it opens the Year sheet.

The Year verdict's rules
MarkWhenThe card saysTry
▲ warningThe month your massing shades most is a heating month (Oct–Mar north of the equator) and it blocks 15 % or more then (a note instead when the weather year is cooling-dominated).Your shading peaks in the heating season — something is in the low winter sun's way.Open the sun calendar for the shaded pane, find the week the shadow crosses it, and move what casts it.
✓ goodThe shading peaks at 15 % or more in a cooling month and stays under 15 % in every heating month.Shading peaks in the cooling season, where you want it.—
● noteThe shading peaks in a cooling month but still blocks 15 % or more in a heating month.Shading is heavy all year — some of it costs winter sun.Find the heating-season month on the curve and trim what shades then.
● noteThe shading never reaches 15 % in any month.The year is almost unshaded.If summer overheating matters here, there is room for an overhang.

Climate — A weather file. It is always a note: it describes the place, not your design. Clicking it opens the Climate sheet.

The Climate verdict's rules
MarkWhenThe card saysTry
● noteHeating degree-days are more than 1.5 × cooling degree-days.A cold-dominated climate: winter sun kept, heat loss and mass matter most.Weight winter kept over summer cut when you read the overhang study.
● noteCooling degree-days are more than 1.5 × heating degree-days.A hot-dominated climate: shading, cross-ventilation and night flushing matter most.Read summer cut, air changes and night-flush hours first.
● noteAnything else.A two-season climate: a device that helps one season and hurts the other is a real trade.—

Earth — A landform — a studio site, a loaded or pasted coordinate, a searched place, or a Surveyor survey all bring one. On flat ground there is no Earth card. Clicking it opens the Earth sheet.

The Earth verdict's rules
MarkWhenThe card saysTry
● noteThe footprint covers no point of the terrain grid.Slope, water and earthwork cannot be read under it.Widen the Ground extent, or move the massing onto the plot.
▲ warningA drainage channel runs through the footprint.The building sits in the water's way.Move the plan off the channel, or regrade a swale around it.
▲ warningWater would stand 4 in or deeper under the footprint — about the terrain data's own vertical error.There is a low point under the slab.A drain under the slab, or seat the massing on higher ground.
● noteThe ground under the footprint slopes more than 15 % on average.A flat pad means a deep cut uphill and fill downhill.Step the section with the ground, or use Seat the landform to balance cut and fill.
● noteThe pad's net earthwork is more than about 65 yd³ in or out.The pad imports (mostly fill) or exports (mostly cut) that much material.Raise or lower the seat until cut and fill balance.
● noteShallower standing water under the footprint (more than about ⅜ in, under 4 in).A slight low point — within the terrain data's error, so check it on a survey before designing around it.—
✓ goodAnything else.A gentle site: slope, and no water through the footprint.—

Energy — Always shown. Clicking it opens the Energy & carbon lens on the right.

The Energy verdict's rules
MarkWhenThe card saysTry
▲ warningOn a typical summer day, windows open, the room floats more than 9 °F above the outdoor daily average.Gains arrive faster than the building can shed them.Cut the summer sun (an overhang, less west glass) before adding ventilation.
● noteIt would float more than 9 °F with the windows shut, and the Air rate it leans on is over 10 air changes an hour (a rate the Air card already reads as a ceiling).Cool only with the windows open.Cut the summer sun at the source first, then give each opening its sash type and read it again.
✓ goodA typical winter day needs no heat at all (the sun and internal heat cover the loss, or the day is already 68 °F).No heating needed — summer is the season to design for.—
● noteThe sun pays back less than 20 % of the winter loss.The winter heat demand is mostly the fabric's.More south glass with winter sun kept, a more compact form, or more working mass.
✓ goodAnything else.What the sun pays and how far summer floats.—

Carbon — Always shown. Clicking it opens the Energy & carbon lens on the right.

The Carbon verdict's rules
MarkWhenThe card saysTry
▲ warningCarbon per square foot of the room's own floor (every floor surface when no room is enclosed) is above the upper band.Names the part carrying most of it — in the jig, usually the slab.Thin the slab toward the mass penetration depth (the Mass card says how much is doing nothing).
● noteCarbon per square foot of the room's floor is between the two bands.Names the largest share.—
✓ goodCarbon per square foot of the room's floor is under the lower band.A compact skin and a slab no thicker than it needs to be.—

The air bands

The Air card and sheet name the rate with the engine's own bands. A passive scheme is sized to purge at 4 air changes an hour; cross-ventilation reaches about 5 ceiling heights into a plan and one-sided ventilation about 2.5.

Air-change bands
Air changes an hourBand
up to 0.35below a dwelling's minimum fresh-air rate
0.35 – 1fresh air only — no useful cooling
1 – 4some relief, but not enough to flush out a day's heat
4 – 10enough to flush out a day's heat (the purge / night-flush range)
over 10very high — check the openings are real and operable

10The force sheets

Under the verdicts, the pills open one drawing each; the right column's Site forces lens has the matching panel of numbers. The verdict is the headline — the sheet is where you find the move.

Sun — which opening, not how much

Every opening is sampled on a 3 × 3 grid across its pane, at the ten Dec 21 and Jun 21 suns (8 am – 4 pm), with a ray against the massing and the landform. The panel: Winter sun kept, Summer sun cut, the pane that earns the most winter sun and the one that takes the most summer sun, and a table per opening — winter and summer gain (square metres of glass squarely facing the sun) and summer ÷ winter. A dash means no sun of that season reaches the pane at all: an orientation fact, not a shading one. The sheet draws each opening's winter and summer bars (outline = available, fill = collected) and, after Sweep it — winter kept vs summer cut, the overhang study curve. Nothing after 4 pm is sampled, so a west window's evening problem is understated.

Air — the mechanism, not the total

Each opening is named windward or leeward (for the wind) and inlet or outlet (for the stack, by height; roof openings are always outlets). The two flows are reported separately and combined in quadrature (BS 5925). Every area here is the part of the opening that opens, set by opens as. Effective open area is the most useful row: the openings work in series, so the smaller side governs — doubling the bigger one buys almost nothing. Also: To purge by the wind alone / by the stack alone (the area the smaller opening needs to reach 4 air changes an hour), One-sided rate (a room with all its windows on one wall), Plan depth against reach, and with a weather file Night-flush hours, Works across the year (the share of the year's hours the wind comes from a useful direction) and the ventilation year — every hour of the file run through an EnergyPlus-style network: the share of the mild hours (68–79 °F outside) the openings purge, the hours nothing moves, whether wind or stack leads, and the median rate. The sheet adds a plan of the openings, the two circuits with the governing side marked, and a rose of the cross-flow at all 16 wind directions.

How far each kind of sash opens — the share every air number multiplies an opening's area by (typical figures, not a measured window):

How far each kind of opening opens
Opens asShare that opensWhy
Whole opening100 %a door, a folding or sliding-away wall, an unglazed hole, or any sash that clears its whole frame — what every air number assumed before this control existed
Casement, side-hung90 %a side-hung sash swung to 90° clears nearly its whole frame; the frame and stop take the rest
Louvres / jalousie80 %blades fully open: their edges and the frame take about a fifth
Sliding / double-hung50 %one sash slides over the other, so at most half the glass is ever uncovered
Awning / hopper, top- or bottom-hung40 %a top- or bottom-hung sash held at about 30°: the gap along the free edge and the two side wedges — the effective-area rule of thumb for top-hung lights (CIBSE AM10 §4, BRE Digest 399)
Tilt-only / trickle15 %a tilt-and-turn on its tilt, or a night-vent stop: the wedge at the head and two narrow side gaps
Fixed0 %a fixed light: daylight and a view, no air at all

View layers → Air (in Viewport display) draws the air in the model — round the building and through the room — and adds a readout: the air through the room, each opening in or out, the neutral plane in a calm (openings below it breathe in, above it out), the breeze across the floor and at the focus, What that breeze is worth (ASHRAE 55: about 2 °F cooler at 1.3 mph, 4 °F at 2.7 mph), and where each opening's air goes. It is potential flow — read the path, not the evenness. Run CFD runs a real CFD simulation in your browser (15 seconds to three minutes) and adds The room's air: air-change time, the age of the room's air, and air-exchange efficiency (100 % piston flow, 50 % evenly stirred, under 50 % short-circuiting). The flow totals are the Air sheet's, not the drawing's.

The wind outside is drawn as three families of line, each its own box under The lines. Street: lines started on a curtain upwind that spans the building from just above the ground to a little over the roof, each drawn from one building height upwind of it to two and a half downwind — the wind parting round it and over it. Along walls: 32 lines started just off the building’s solid faces, windows or none — the wind stopping against a blank wall, running along it and round the corners (after a CFD run on every face; in potential flow on the windward faces only, because behind the building potential flow draws a mirror of the front, which real wind never does). Eddy, after a CFD run only: 24 heavier lines on a pale casing, started where the air behind the building runs back toward it, against the wind, and followed round for 3 turns. It is the average over the run: at any moment gusts cross it and it sheds and re-forms. Potential flow has no eddy, so it draws none. The lines through the room are drawn outside only for a building height and a half of travel before the inlet and after the outlet.

That readout opens with Read the drawing: one row per mark in the picture — the dashes, their colour, the rings on the openings, the wind arrow, the places no dash goes — each saying what it measures (the same figure the stats under it print), what it means for a person in the room at this place and date (a draught or a cooling credit, heat carried out of a heated room or brought in on a hot afternoon, a placeholder wind or the sector the year really blows from), and the move. Hover a row and the model lights its mark; the colour row draws the standing points the breeze is read at. With the drawn air on the date, the two buttons under the rows take the Date + time bar to a January morning and a July afternoon — read both before you decide an opening is right.

Mass — how much of it is working

Only the first few inches of a slab take part in a daily cycle — the penetration depth. The panel: Mass actually working (ft³), Inert on a daily cycle (%), Time constant, Decrement factor (the share of the outdoor swing that gets inside — heavy 0.2–0.4, light 0.7–1.0), Time lag (hours the indoor peak trails the outdoor), and Indoor swing. The sheet draws the 24-hour swing outside and in, and the slab in section with the working depth marked. Pick the material in the rail:

Mass materials and how deep a daily temperature swing reaches into each
MaterialDensityDaily penetration depthSource
Dense concrete144 lb/ft³ (2300 kg/m³)5.7 in (0.145 m)CIBSE Guide A, dense cast concrete
Brickwork109 lb/ft³ (1750 kg/m³)4.7 in (0.120 m)CIBSE Guide A, fired clay brick
Limestone137 lb/ft³ (2200 kg/m³)5.7 in (0.144 m)CIBSE Guide A, sedimentary stone
Rammed earth125 lb/ft³ (2000 kg/m³)4.9 in (0.124 m)Standards Australia HB 195, stabilised rammed earth
Concrete block87 lb/ft³ (1400 kg/m³)4.6 in (0.117 m)CIBSE Guide A, medium-density block
Cross-laminated timber30 lb/ft³ (480 kg/m³)2.7 in (0.068 m)CIBSE Guide A / EN 16351, softwood CLT
Water (a mass wall)62 lb/ft³ (1000 kg/m³)2.5 in (0.063 m)the reference case — the highest volumetric capacity a building can hold

View — a direction, and its conflict with the sun

Read from up to 81 standing positions on the floor inside the walls, not one point. A position shares the view when at least 0.1 sr of glass faces the view from it (about a 3 ft × 3 ft window seen squarely from 10 ft). The panel: Floor that shares the view, View quality, best → worst, which way the glass faces, View against sun (the angle between the view target and the winter sun — past 90° every pane has to be argued for twice) and Glass serving both. The sheet is a rose of where the glass looks, with the view and the sun on one dial. (The Readout's view rows use a different eye: one seat at the focus.)

Shadows

Plan shadows of the massing on Dec 21, the equinox and Jun 21 at 8, 10, 12, 2 and 4 o'clock, winter darkest, with how far the winter-noon shadow reaches north. A drawing; it has no verdict card.

Year — the overhang study as a curve (weather file)

Solar energy on the glass month by month, from the weather file's own sky (kWh/m², metric by house rule): outline = the open sky, fill = what your massing lets through, and a strip below of the share it blocks each month. A good overhang blocks most in summer and least in December. Beside it: the eight compass faces of the opaque skin, open against actual. The sun calendar needs no weather file: pick an opening and it draws every day of the year across and every half hour down — night, shaded, or in sun — with sunlit hours a year and the best and worst month. The sun-path dial paints the same opening's shading on the sky dome.

Climate (weather file)

Each month's coldest and hottest hour, and its mean, against the 68–79 °F comfort band; the share of hours the place is cold, comfortable or hot on its own; heating and cooling degree-days counted against the band's edges, not the almanac's 65 °F base (so never put them beside published HDD65 figures); and the strategy mix — what share of the year passive solar, ventilation, mass and evaporative cooling could cover here, against what this design captures.

Earth (landform)

Slope under the footprint and which way the ground faces, relief under the building, the pad cut · fill of a level pad at the plinth's base (with its error bar), net earth in or out, the catchment draining through, channels through the footprint and standing water under it. The sheet is a contour plan with the drainage lines; under it, a site section along the fall line with cut and fill hatched. Seat the landform swings cut against fill. Screening resolution: a landform study, not a grading plan.

Balance

Sun, air, mass and view side by side on four gauges — summer ÷ winter, air changes, share of the swing damped, share of the floor with the view — deliberately not added up. There is no single score, on purpose.

11The focus — the fourth rectangle

V7's own reading: a table or a seating pit, built of the plinth's material, placed inside the room.

The fourth rectangle — the focus sits in the Site forces lens. It reads Rect D off the built model:

The focus panel's readings
ReadingWhat it is
Table / PitIts plan area (ft²), and how far a table rises above the plinth top or a pit drops below it.
Mass added / Mass taken outPlan area × rise or depth, in ft³ of plinth. Already inside Mass volume and slab carbon — nothing is counted twice.
Winter sun on the focusThe share of the fourth rectangle — the table top or the seating-pit floor — in direct winter sun. This one is MEASURED: every cell of each sun patch fires a ray back at the sun, so the pit's own rim, the walls and the roof all count.
Summer sun on the focusThe same measured patch in summer. A hearth wants winter sun and summer shade: this pair says that as two numbers.
The view from the seatView to target and view openness (sr) from a SEATED eye at the centre of the focus, 3′-11″ above the pit floor or the slab the table stands on.
  • In the jig, point 4 places the focus in the room's clear span (u, v in %), with Focus width, Focus depth, Focus · z (up for a table, down for a pit) and Focus rotation. The program keeps it legal: a 1′-6″ walkway to every wall, at most 60 % of the room each way, a table at least 3 ft under the eave, and a pit that never cuts through the last 6″ of plinth — so at the default 2 ft plinth, a −1.5 ft pit is already the deepest; a deeper pit needs a taller plinth. z = 0 means no focus.
  • On an import, the focus is whatever is on the Focus layer (section 2). The panel's numbers work; its little plan drawing may be blank, because it looks for the jig's part names.
  • There is no focus verdict card. To calibrate the focus, make it a goal: in Goals, Warm the focus and Shade the focus draft targets on the two sun readings.

12The Readout, key by key

Every number the shared engine returns — 52 keys in nine groups, plus up to 9 measured-climate rows with a weather file. The Readout prints them in imperial units; press Show methods to print each one's method under it, and open Method notes from this run for what this particular run did (for example, that the terrain is in the sun rays).

Each entry: the label and unit as the Readout prints them, which way is usually better (a cue, not a law), what it needs from the model, what it reads from, what it tells you, what moves it, and — in the fold — the tool's own method note and data sources, quoted.

☀ Sun

Winter solar gain

idxHigher is usually better

Needs: openings (Glazing surfaces)

Reads from: The openings · The sun's path · Your massing

What it tells you:
How much glass faces the low winter sun squarely, averaged over five hours of Dec 21 (8 am to 4 pm, sun time). The unit “idx” is square metres of glass turned straight at the sun — a comparison number, not energy.
What moves it:
More or bigger glass facing south; take away whatever shades it in winter.
In V7:
The hover note says this ignores self-shading. In V7 it does not: the tool re-casts every sun ray against your massing and the landform, so with flat ground it equals “Winter gain · self-shaded”.
The tool's own method note, and where its data comes from

How much direct winter sun reaches your openings — more means more free heat on a cold day. For each opening: its area × how squarely the sun strikes it, averaged over 5 sample hours on the winter solstice (8 am, 10 am, noon, 2 pm and 4 pm sun time; Dec 21 in the northern hemisphere, Jun 21 in the southern). Unit 'idx': the opening area the sun sees face-on, counted in square metres — 1 idx = 10.8 ft² (1 m²). A number for comparing schemes, not heat. No standard sets a target: compare your own schemes. Whether shading counts depends on the tool showing it: some ray-test each opening's centre against your model and any loaded landform, so an overhang or a hill lowers it; others show the unshaded figure.

  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.
  • The sun's path: Sun positions computed from the site's latitude and the date and hour — geometry, not a measurement: no clouds are in it. The latitude is the Site dial, or the place you searched or loaded.
  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.

Summer solar gain

idxLower is usually better

Needs: openings (Glazing surfaces)

Reads from: The openings · The sun's path · Your massing

What it tells you:
The same count for Jun 21 — the sun you usually want to keep out.
What moves it:
An overhang over the south glass; less east and west glass; a smaller or shaded skylight.
In V7:
Also ray-cast against the massing and the landform in V7.
The tool's own method note, and where its data comes from

The same measure on the summer solstice — usually heat you want to keep OUT, so lower is better in most climates. Same unit (idx), same 5 sample hours and the same shading rule as Winter solar gain.

  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.
  • The sun's path: Sun positions computed from the site's latitude and the date and hour — geometry, not a measurement: no clouds are in it. The latitude is the Site dial, or the place you searched or loaded.
  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.

Overheating ratio

×Lower is usually better

Needs: openings (Glazing surfaces)

Reads from: The openings · The sun's path · Your massing

What it tells you:
Summer gain ÷ winter gain. Below 1, the form takes in more winter sun than summer sun — the balance you want in a heating climate. A reading of 99 is a flag, not a ratio: summer sun reaches the glass and winter sun does not.
What moves it:
Anything that raises winter gain or cuts summer gain.
The tool's own method note, and where its data comes from

Summer sun ÷ winter sun on the openings (the two rows above), so no unit. Below 1 the form lets in more winter sun than summer sun — the passive-solar aim (a rule of thumb, not a standard). A balance cue, not a cooling load. Reads 99 when summer sun reaches the openings but no winter sun does: a divide-by-zero flag, not a measurement. Shaded exactly as the two rows above.

  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.
  • The sun's path: Sun positions computed from the site's latitude and the date and hour — geometry, not a measurement: no clouds are in it. The latitude is the Site dial, or the place you searched or loaded.
  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.

Winter gain · self-shaded

idxHigher is usually better

Needs: openings (Glazing surfaces)

Reads from: The openings · The sun's path · Your massing · The ground

What it tells you:
Winter gain after each opening is checked against your own massing's shadow (one ray from the centre of each pane). This is the number the energy balance uses.
What moves it:
Shallower or higher overhangs; move a wing or wall out of the low winter sun.
The tool's own method note, and where its data comes from

Winter sun on the openings after your model's own solid parts (overhangs, fins, walls, roof) cast their shadows: one ray from each opening's centre toward each of the 5 winter sample suns. Same unit (idx) as Winter solar gain. This is the winter sun the heat balance uses (Winter heat demand, Solar fraction). Winter sun kept = this ÷ the same sum with nothing in the way. Where Winter solar gain is shown unshaded, the gap between the two rows is what your shading removed. If a tool hands this test none of your model's parts, it equals the unshaded sum — and then Winter sun kept reads 100 % and Summer sun cut 0 %, which says nothing about your shading. Whether a loaded landform's shadow counts here too depends on the tool.

  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.
  • The sun's path: Sun positions computed from the site's latitude and the date and hour — geometry, not a measurement: no clouds are in it. The latitude is the Site dial, or the place you searched or loaded.
  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.
  • The ground: The landform under the site — USGS 3DEP where it is measured, AWS terrain tiles outside the US, a flat plane when none is loaded. The ground banner says which.

Summer gain · self-shaded

idxLower is usually better

Needs: openings (Glazing surfaces)

Reads from: The openings · The sun's path · Your massing · The ground

What it tells you:
Summer gain after the same shadow check. A deep south overhang cuts this while barely touching the winter figure.
What moves it:
Deeper overhang over south glass; fins on east and west glass.
The tool's own method note, and where its data comes from

Summer sun on the openings after the same shadow test. A deep overhang over a sun-facing window cuts THIS while leaving the winter figure almost untouched, because the high summer sun is blocked and the low winter sun slips under — the classic passive move. Same unit (idx) and the same shadow rule as the row above; this is the summer figure the heat balance uses. Whether a loaded landform's shadow counts here too depends on the tool.

  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.
  • The sun's path: Sun positions computed from the site's latitude and the date and hour — geometry, not a measurement: no clouds are in it. The latitude is the Site dial, or the place you searched or loaded.
  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.
  • The ground: The landform under the site — USGS 3DEP where it is measured, AWS terrain tiles outside the US, a flat plane when none is loaded. The ground banner says which.

Winter sun kept

%Higher is usually better

Needs: openings (Glazing surfaces)

Reads from: The openings · The sun's path · Your massing · The ground

What it tells you:
Of the winter sun that would reach your openings with nothing in the way, the share your own massing lets through. 100 % means nothing of yours shades the glass in winter.
What moves it:
Pull back overhangs, wings or courtyard walls that sit between the low sun and the glass.
In V7:
The Readout checks one point per pane. The Sun sheet and the Sun verdict sample a 3 × 3 grid across every pane, so their “winter kept” can differ by a few points — trust the Sun sheet for shading design.
The tool's own method note, and where its data comes from

Share of the winter sun that still reaches the openings after shading: the self-shaded winter figure ÷ the same sum with nothing in the way. 100 % = nothing blocks the low winter sun, which is the aim in a heating climate (a passive-solar rule of thumb). It uses one ray per window centre at the 5 sample hours, so it jumps in steps as you move an overhang; a check that tests several points across each pane can read differently on a partly shaded window, and is the better figure. The same quantity the Brise tool reports for one shading device, here for the whole form. Reads 0 when no winter sun reaches any opening at all: nothing to keep. Shaded exactly as Winter gain · self-shaded. Whether a loaded landform's shadow counts here too depends on the tool.

  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.
  • The sun's path: Sun positions computed from the site's latitude and the date and hour — geometry, not a measurement: no clouds are in it. The latitude is the Site dial, or the place you searched or loaded.
  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.
  • The ground: The landform under the site — USGS 3DEP where it is measured, AWS terrain tiles outside the US, a flat plane when none is loaded. The ground banner says which.

Summer sun cut

%Higher is usually better

Needs: openings (Glazing surfaces)

Reads from: The openings · The sun's path · Your massing · The ground

What it tells you:
Of the summer sun that would reach your openings with nothing in the way, the share your own massing blocks. It moves in steps, because each season is only five sampled hours.
What moves it:
Deepen the overhang, add fins, or shade the skylight. The overhang study (Sweep it) draws this against winter kept.
In V7:
Same one-point-per-pane caveat as winter sun kept.
The tool's own method note, and where its data comes from

Share of the summer sun blocked before it reaches the openings: 1 − (self-shaded ÷ unshaded summer figure). Higher = better shading. Deepen an overhang and this rises while Winter sun kept barely moves — the classic passive move, read as two numbers. With only 5 sample hours and one ray per window it moves in visible STEPS and stops rising once every sampled hour is already blocked. A cue for comparing forms, not a shading schedule: a real shading study runs every hour of a weather year (Ladybug or ClimateStudio). Shaded exactly as Summer gain · self-shaded. Whether a loaded landform's shadow counts here too depends on the tool.

  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.
  • The sun's path: Sun positions computed from the site's latitude and the date and hour — geometry, not a measurement: no clouds are in it. The latitude is the Site dial, or the place you searched or loaded.
  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.
  • The ground: The landform under the site — USGS 3DEP where it is measured, AWS terrain tiles outside the US, a flat plane when none is loaded. The ground banner says which.

Equator-facing gain

idxHigher is usually better

Needs: openings (Glazing surfaces)

Reads from: The openings · The sun's path · Your massing

What it tells you:
Gain through only the openings that face within 45° of south (north, below the equator), averaged over all fifteen sampled suns — the whole year, not just winter.
What moves it:
Put more of the glass on the south side.
The tool's own method note, and where its data comes from

Sun reaching only the openings that face the equator — within 45° of due south in the northern hemisphere, due north in the southern. That orientation takes low winter sun and is the easiest to shade in summer. Averaged over all 15 sample suns (the 5 hours on each solstice and on the equinox), so it is not a winter-only figure. Same unit (idx) and the same shading rule as Winter solar gain.

  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.
  • The sun's path: Sun positions computed from the site's latitude and the date and hour — geometry, not a measurement: no clouds are in it. The latitude is the Site dial, or the place you searched or loaded.
  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.

Whole-skin sun

idxLower is usually better

Reads from: Your massing · The sun's path

What it tells you:
Direct sun landing on all the opaque skin — walls, roof and plinth — averaged over the year's fifteen suns. More sun on the skin is more heat to shed in summer.
What moves it:
A more compact form; less roof and west wall turned to the sun; shade the skin.
The tool's own method note, and where its data comes from

Direct sun on the whole outer skin — walls, roof and plinth faces, not the openings — averaged over all 15 sample suns. Unit idx = square metres of skin the sun sees face-on. Lower = less sun heating the walls and roof; in a cold climate that sun is welcome, so 'lower is better' is only a default. An index for comparing schemes, not an area. The heat balance does not use it (it counts sun through the openings only). Shaded by the same rule as Winter solar gain; where the faces are not ray-tested, every face turned toward the sun counts, even one hidden behind another part of the model.

  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.
  • The sun's path: Sun positions computed from the site's latitude and the date and hour — geometry, not a measurement: no clouds are in it. The latitude is the Site dial, or the place you searched or loaded.

▤ Sun indoors

Winter sun on the floor

%Higher is usually better

Needs: openings (Glazing surfaces), an up-facing plinth

Reads from: The openings · The floor · The sun's path · Your massing

What it tells you:
The share of the floor plan standing in a patch of direct sun, averaged over the five winter hours. In the 4 Rectangles game the “floor” is the whole plinth top, terrace included.
What moves it:
Wider glass, lower sills, or turn the room toward the sun.
In V7:
Each patch is the window's shape slid down the sun ray. Interior walls and the far wall do not trim it, so a low sun can throw a patch further than a real room would let it go.
The tool's own method note, and where its data comes from

Share of the floor standing in a patch of direct sun, averaged over the 5 winter sample hours. The sun rows above count sun arriving AT the glass; this counts where it LANDS. Widen the glass or lower the sill and it rises; turn the room away from the sun and it collapses. How: each window whose centre sees the sun has its outline slid along the sun's direction onto the floor; walls and partitions in between do NOT trim the patch here (the focus rows below do check every point). The floor is every up-facing floor surface — a plinth or terrace outside the walls included — so this is a share of that whole floor, not necessarily of the room. Reads 0 when the engine has no floor surface to sample. Direct sun only — not daylight. How bright a room is, is measured by sDA (spatial daylight autonomy, IES LM-83), which is the Daylighter tool's job. No standard sets a target for this number.

  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.
  • The floor: The up-facing floor surfaces — the Plinth layer, or any solid with the floor role. Without one there is no plan to sample.
  • The sun's path: Sun positions computed from the site's latitude and the date and hour — geometry, not a measurement: no clouds are in it. The latitude is the Site dial, or the place you searched or loaded.
  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.

Summer sun on the floor

%Lower is usually better

Needs: openings (Glazing surfaces), an up-facing plinth

Reads from: The openings · The floor · The sun's path · Your massing

What it tells you:
The same patch in summer — usually sun you do not want indoors.
What moves it:
An overhang: it can take this close to zero while the winter figure barely moves.
The tool's own method note, and where its data comes from

The same sun-patch share in summer — the season you usually don't want it. An overhang that barely touches the winter figure can take this most of the way to zero. 'Lower is better' is a default, not a law: unwanted in a hot climate, welcome in a cold one. Same method and limits as the winter row. Not a cooling load — no glass properties, no mass, no schedule.

  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.
  • The floor: The up-facing floor surfaces — the Plinth layer, or any solid with the floor role. Without one there is no plan to sample.
  • The sun's path: Sun positions computed from the site's latitude and the date and hour — geometry, not a measurement: no clouds are in it. The latitude is the Site dial, or the place you searched or loaded.
  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.

Winter sun on the focus

%Higher is usually better

Needs: openings (Glazing surfaces), a focus

Reads from: The openings · The focus · The sun's path · Your massing

What it tells you:
The share of the fourth rectangle — the table top or the seating-pit floor — in direct winter sun. This one is MEASURED: every cell of each sun patch fires a ray back at the sun, so the pit's own rim, the walls and the roof all count.
What moves it:
Move the focus into the path of a south window; lower that window's sill; make the pit shallower (its rim shades it); a table catches more sun than a pit.
In V7:
Reads 0 with a “no focus” flag if the model has no focus. On an import, put the focus on the Focus layer.
The tool's own method note, and where its data comes from

Share of the FOCUS — the surface your model names as its focus, such as a table top or a seating-pit floor — standing in direct sun, averaged over the 5 winter sample hours. Read on the focus's OWN plane (a pit floor sits lower, so a window's sun patch lands further from the glass), and RAY-TESTED, not just projected: every point in a window's patch fires a ray at the sun, so the pit's rim, the walls and the roof all cast their shadows — for example, at 42° N a 1′-6″-deep pit loses about 3 ft (1 m) of its floor to the rim's shadow at a winter-solstice noon. A table top above the sill gets only the light through the part of the window above it. A calculation on your model, not a measurement; cloudless direct sun only — no sky light, no reflections. Reads 0 when the model names no focus.

  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.
  • The focus: The fourth rectangle — a solid on the Focus layer (or named focus): the table top, or the seating-pit floor.
  • The sun's path: Sun positions computed from the site's latitude and the date and hour — geometry, not a measurement: no clouds are in it. The latitude is the Site dial, or the place you searched or loaded.
  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.

Summer sun on the focus

%Lower is usually better

Needs: openings (Glazing surfaces), a focus

Reads from: The openings · The focus · The sun's path · Your massing

What it tells you:
The same measured patch in summer. A hearth wants winter sun and summer shade: this pair says that as two numbers.
What moves it:
A deeper overhang over the window that lights the focus, or tuck the focus under the eave side.
The tool's own method note, and where its data comes from

The same ray-tested patch on the focus in summer, the season you usually don't want it — a deep overhang, or a pit tucked under the eave side, takes it toward zero. A hearth wants winter sun and summer shade; this pair says that as two numbers. Same method and limits as the winter row; reads 0 when the model names no focus.

  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.
  • The focus: The fourth rectangle — a solid on the Focus layer (or named focus): the table top, or the seating-pit floor.
  • The sun's path: Sun positions computed from the site's latitude and the date and hour — geometry, not a measurement: no clouds are in it. The latitude is the Site dial, or the place you searched or loaded.
  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.

↯ Wind

Windward area

ft²Lower is usually better

Reads from: Your massing · The wind dials

What it tells you:
The area of the building facing the prevailing wind. A shelter cue: less area, less wind load and less cold wind on the skin.
What moves it:
Turn the narrow end to the wind; lower or berm the windward side.
In V7:
It counts the inside faces of thin walls as well as the outside, so it runs high. Compare schemes with it; do not quote it.
The tool's own method note, and where its data comes from

How much of the skin the prevailing wind hits: every skin face turned toward the wind × how squarely it faces it, summed — walls, pitched roof slopes and plinth edges; only near-flat roofs and floors (under about 18°) are left out. Where walls and roofs are modelled as closed solids, the inside face of the far wall and the underside of a roof slope also 'face' the wind, so it reads up to roughly twice the real windward area — compare massings with it, don't read it as an area. Smaller = more sheltered. A shelter cue, not a pressure map: real pressures vary across a facade and come from wind-tunnel data or CFD (computational fluid dynamics, a simulation of the moving air); Cp, the pressure coefficient, is the number those supply.

  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.
  • The wind dials: Site → wind speed and direction: the DESIGN wind every score reads. Set from the place's ERA5 wind rose (its mean speed and prevailing direction) when you search a place or pick a studio site, from the survey on a Surveyor Sync, or aimed at the hot, open-window or whole-year hours by the buttons under them — otherwise the values you set. The drawn air reads them only while it follows the dials; following the date, it takes the date's wind instead.

Wind load proxy

lbfLower is usually better

Reads from: Your massing · The wind dials

What it tells you:
A rough sideways wind force on the building (½ × air density × wind speed² × windward area), in pounds-force. A comparison for lateral load, not a structural figure.
What moves it:
Same as windward area; it also rises with the Wind speed dial.
The tool's own method note, and where its data comes from

A rough wind push on the building, in pounds-force (lbf): ½ × air density (0.0765 lb/ft³, 1.225 kg/m³, sea level) × wind speed² × the windward area above — as if the wind stopped dead against every face counted there. Uses one wind speed, unchanged at every height: no gusts, no increase with height, no pressure coefficients, no suction on the far side. Compare massings with it; never size anything with it — structural wind loads come from ASCE 7 and a structural engineer.

  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.
  • The wind dials: Site → wind speed and direction: the DESIGN wind every score reads. Set from the place's ERA5 wind rose (its mean speed and prevailing direction) when you search a place or pick a studio site, from the survey on a Surveyor Sync, or aimed at the hot, open-window or whole-year hours by the buttons under them — otherwise the values you set. The drawn air reads them only while it follows the dials; following the date, it takes the date's wind instead.

≈ Air

Ventilation provision

cfmHigher is usually better

Needs: openings (Glazing surfaces)

Reads from: The openings · Your massing · The wind dials · The ΔT dial

What it tells you:
Air the openings can move through the building, in cubic feet per minute: the stack flow (warm air rising out of high openings) and the wind flow (in on the windward side, out the leeward) combined. Each opening counts only the part of it that opens — its “opens as” setting, the whole opening unless you set it.
What moves it:
Enlarge the SMALLER of each inlet–outlet pair; add a high outlet (a skylight or clerestory); add an opening that faces the wind; or let a sash open further (a casement clears about 90 %, a sliding sash 50 %).
The tool's own method note, and where its data comes from

How much outdoor air the openings could move through the building with every one of them open, in cubic feet per minute (cfm). Two drivers, each by BS 5925's formulas (the British code of practice for natural ventilation): warm air rising (the stack — from the height between low inlets and high outlets and the inside − outside ΔT) and wind blowing through (cross-flow — windward openings against leeward ones, from the wind speed and one assumed pressure difference, ΔCp 0.7). They are combined as √(stack² + wind²), the approximation the ASHRAE Handbook of Fundamentals gives for the two acting together. Low inlets with high or roof outlets raise it. One room, no partitions or doors, and each opening counted as open as far as its sash type allows (the whole pane unless a sash type says less) — read it as a ceiling. Before anyone relies on it for comfort, a mechanical engineer checks it with a room-by-room airflow model.

  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.
  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.
  • The wind dials: Site → wind speed and direction: the DESIGN wind every score reads. Set from the place's ERA5 wind rose (its mean speed and prevailing direction) when you search a place or pick a studio site, from the survey on a Surveyor Sync, or aimed at the hot, open-window or whole-year hours by the buttons under them — otherwise the values you set. The drawn air reads them only while it follows the dials; following the date, it takes the date's wind instead.
  • The ΔT dial: Site → the inside−outside temperature difference that drives the stack. A design assumption you set — no weather file changes it — and what every score reads. The drawn air reads it only while it follows the dials; following the date, it takes the date's room minus outside instead, which can be negative.

Stack ventilation

cfmHigher is usually better

Needs: openings (Glazing surfaces)

Reads from: The openings · The ΔT dial

What it tells you:
The stack flow on its own — how much air buoyancy alone moves.
What moves it:
A taller gap between low inlets and high outlets; bigger outlets; a larger inside–outside temperature difference.
In V7:
Despite the “idx” label this is a flow in cubic metres per second: multiply by 2,119 for cubic feet per minute.
The tool's own method note, and where its data comes from

Airflow from warm air rising alone (the stack effect), in cubic feet per minute (cfm). BS 5925's formula: 0.61 × the openings' combined effective area × √(2 × g × stack height × ΔT ÷ the absolute indoor temperature (68 °F, 293 K)). The openings are split into low inlets and high outlets at their average height (a roof opening is always an outlet); inlets and outlets act in series, so the smaller side governs the effective area. It needs BOTH a height difference and an inside–outside temperature difference (ΔT) — a setting you choose; no weather file sets it. Each opening counted as open as far as its sash type allows (the whole pane unless one says less); one room, no partitions.

  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.
  • The ΔT dial: Site → the inside−outside temperature difference that drives the stack. A design assumption you set — no weather file changes it — and what every score reads. The drawn air reads it only while it follows the dials; following the date, it takes the date's room minus outside instead, which can be negative.

Stack height

ftHigher is usually better

Needs: openings (Glazing surfaces)

Reads from: The openings

What it tells you:
The height from the average inlet up to the average outlet. Stack ventilation needs this gap.
What moves it:
Raise the outlets (a roof opening, a clerestory) or lower the inlets.
The tool's own method note, and where its data comes from

Height from the average low (inlet) opening to the average high (outlet) opening — the 'chimney' warm air rises through. Taller = more stack flow (the flow grows with its square root). Openings are split at their area-weighted average height, and a roof opening is always an outlet. Openings all at one height give a stack of 0.

  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.

Air path you can't shut

cfmNeither way is better — it depends on the design

Needs: openings (Glazing surfaces)

Reads from: The openings · The wind dials · The ΔT dial

What it tells you:
Air through the openings you switched to OPEN (no glass) alone. You cannot close a hole, so this air moves on a January night as well as a July one. 0 when every opening is glazed.
What moves it:
Glaze an opening (Openings group) or make the open ones smaller.
The tool's own method note, and where its data comes from

Airflow, in cfm, through the UNGLAZED openings alone — the ones with no glass in them, which cannot be closed on a winter night. The same BS 5925 pair as Ventilation provision (stack and cross-flow, combined as √(stack² + wind²)) run over those openings only, then never allowed below BS 5925's single-opening rules (wind 0.025 × area × wind speed; stack ⅓ × 0.61 × area × √(g × opening height × ΔT ÷ the absolute indoor temperature (68 °F, 293 K))), so one lone hole never reads as free. Reads 0 when every opening is glazed. This is the term that makes unglazing cost heat: it is added to the SEALED heat loss, where an openable window adds only its pane's U-value × area. Not a comfort number and not a draught; one room, same limits as the rows above.

  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.
  • The wind dials: Site → wind speed and direction: the DESIGN wind every score reads. Set from the place's ERA5 wind rose (its mean speed and prevailing direction) when you search a place or pick a studio site, from the survey on a Surveyor Sync, or aimed at the hot, open-window or whole-year hours by the buttons under them — otherwise the values you set. The drawn air reads them only while it follows the dials; following the date, it takes the date's wind instead.
  • The ΔT dial: Site → the inside−outside temperature difference that drives the stack. A design assumption you set — no weather file changes it — and what every score reads. The drawn air reads it only while it follows the dials; following the date, it takes the date's room minus outside instead, which can be negative.

Air changes you can't shut

1/hLower is usually better

Needs: openings (Glazing surfaces)

Reads from: The openings · Your massing · The wind dials · The ΔT dial

What it tells you:
That same unshuttable air as air changes per hour. If it is close to the total air changes, the scheme has no ventilation control at all.
What moves it:
As above.
The tool's own method note, and where its data comes from

Air changes per hour through the unglazed openings alone: the path above ÷ the engine's estimate of the enclosed volume (see Enclosed volume). Compare it with Air changes: if the two are close, the scheme has no ventilation CONTROL — what it gets in July it also gets in January. 0 when every opening is glazed.

  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.
  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.
  • The wind dials: Site → wind speed and direction: the DESIGN wind every score reads. Set from the place's ERA5 wind rose (its mean speed and prevailing direction) when you search a place or pick a studio site, from the survey on a Surveyor Sync, or aimed at the hot, open-window or whole-year hours by the buttons under them — otherwise the values you set. The drawn air reads them only while it follows the dials; following the date, it takes the date's wind instead.
  • The ΔT dial: Site → the inside−outside temperature difference that drives the stack. A design assumption you set — no weather file changes it — and what every score reads. The drawn air reads it only while it follows the dials; following the date, it takes the date's room minus outside instead, which can be negative.

⚡ Energy

Winter heat demand

kBtu/dayLower is usually better

Reads from: Your massing · The openings · The design days · The construction · The sun's path

What it tells you:
Heat the room still needs on an average winter day to stay at 68 °F, after the sun and the people inside have done what they can. One room, one fixed wall build-up: a ranking, not a bill.
What moves it:
More south glass with the winter sun kept; a more compact form; less glass facing north; glaze any open openings.
The tool's own method note, and where its data comes from

Heat still needed on a TYPICAL winter day, in thousands of Btu per day (kBtu/day). Heat lost through walls, roof, glass, ground and air leakage at the gap between 68 °F (20 °C) inside and the winter-day temperature, minus the share of sun and internal heat the building can actually use (ISO 13790's utilisation method — more mass, more of it usable). The winter day is the average of the three winter months of a loaded weather year where the tool reads it from one; otherwise it is a placeholder, 32 °F (0 °C), that some tools let you change. Either way it is NOT the much colder 99 % 'design day' an engineer sizes a furnace with. One room and one assumed construction (wall U-0.056, roof U-0.039, window U-0.32 Btu/h·ft²·°F; SHGC 0.6 = the share of sun the glass lets through as heat). Ranks schemes; a real heating load comes from an ACCA Manual J calculation or an EnergyPlus model.

  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.
  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.
  • The design days: Site → Energy: the winter and summer day, the day–night swing and the internal gains. The first three are set from the place's measured year when you search a place, pick a studio site or load a weather file (its Dec–Feb and Jun–Aug mean dry-bulb, and the seasonal beam sun 8 am–4 pm); otherwise — and always for internal gains — they are the dials you set.
  • The construction: One fixed build-up per role in lib/crucible/energy.ts — a U-value each for wall, roof, floor and glass, and 0.3 ACH of leakage. An envelope sent from the Detailer replaces the wall, roof and floor.
  • The sun's path: Sun positions computed from the site's latitude and the date and hour — geometry, not a measurement: no clouds are in it. The latitude is the Site dial, or the place you searched or loaded.

Solar fraction

0–1Higher is usually better

Needs: openings (Glazing surfaces)

Reads from: Your massing · The openings · The design days · The construction · The sun's path

What it tells you:
The share of that winter day's heat loss the sun pays back (0 to 1). Heavier buildings bank more of a sunny noon, so mass raises it too.
What moves it:
More winter sun through the glass; more exposed mass; less heat loss.
The tool's own method note, and where its data comes from

How much of the typical winter day's heat loss the usable sun pays back, as a share from 0 to 1 (0.25 = a quarter). The sun is the self-shaded winter figure × SHGC 0.6 (1.0 through an opening with no glass) × the direct-sun strength — a clear-sky 190 Btu/h·ft² (600 W/m²), or a loaded weather year's winter average where the tool reads one — × 8 hours; then only the share the building can bank without overheating counts (ISO 13790's utilisation: a heavy building banks midday sun a light one would overheat on). Higher = the sun does more of the heating. No standard sets a target.

  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.
  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.
  • The design days: Site → Energy: the winter and summer day, the day–night swing and the internal gains. The first three are set from the place's measured year when you search a place, pick a studio site or load a weather file (its Dec–Feb and Jun–Aug mean dry-bulb, and the seasonal beam sun 8 am–4 pm); otherwise — and always for internal gains — they are the dials you set.
  • The construction: One fixed build-up per role in lib/crucible/energy.ts — a U-value each for wall, roof, floor and glass, and 0.3 ACH of leakage. An envelope sent from the Detailer replaces the wall, roof and floor.
  • The sun's path: Sun positions computed from the site's latitude and the date and hour — geometry, not a measurement: no clouds are in it. The latitude is the Site dial, or the place you searched or loaded.

Heat loss rate

Btu/h·°FLower is usually better

Reads from: Your massing · The openings · The construction

What it tells you:
How much heat the sealed building loses for every degree between inside and outside (Btu per hour per °F): walls, roof, glass, the ground contact, background leakage, and the air through any opening with no glass.
What moves it:
Less glass, a more compact form. Switching an opening to OPEN raises it — a hole leaks far more than a pane.
The tool's own method note, and where its data comes from

Heat the building loses every hour for each °F it is warmer inside than out (Btu per hour per °F) — multiply by the temperature difference to get the heat flow. Smaller = better insulated and tighter. The sum of: walls + roof + GLAZED openings (area × U-value, one assumed construction: whole-wall U-0.056 Btu/h·ft²·°F, about R-18 counting every layer, and roof U-0.039, about R-26. For comparison, the 2021 IECC energy code (International Energy Conservation Code) caps a new home in climate zone 5, most of Illinois, at U-0.045 for a frame wall and U-0.024 for a ceiling, so both assumed assemblies lose more heat than a new code-built home. The familiar code R-values — R-60 in an attic, R-20 plus R-5 of continuous insulation in a wall — are the insulation alone, not the whole assembly); the below-grade skin and slab at 40 % (the ground is milder than winter air — an assumed factor); 0.3 air changes an hour of background leakage (assumed; a blower-door test measures the real one); and the air path through any UNGLAZED opening, because you cannot shut a hole. Switching an opening to OPEN drops that pane's U × area and adds a much bigger air term, so this goes UP.

  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.
  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.
  • The construction: One fixed build-up per role in lib/crucible/energy.ts — a U-value each for wall, roof, floor and glass, and 0.3 ACH of leakage. An envelope sent from the Detailer replaces the wall, roof and floor.

Summer float

°F (a difference)Lower is usually better

Reads from: Your massing · The openings · The design days · The construction · The sun's path · The wind dials · The ΔT dial

What it tells you:
How many degrees the inside runs above the outdoor average on a warm afternoon, with the openings ventilating. An overheating cue.
What moves it:
Cut summer sun first (overhang, less west glass); then more ventilation.
The tool's own method note, and where its data comes from

Degrees warmer than the outdoor average the inside floats on a summer afternoon with the openings ventilating and no air conditioning — a difference, not a temperature. (Sun on the openings, self-shaded, × SHGC × the direct-sun strength, + internal gains) ÷ the heat the building sheds per degree through its skin and its open windows. No equipment, no humidity, one moment. Lower = cooler. An overheating cue for comparing schemes, not a cooling load; whether a room overheats is an hour-by-hour EnergyPlus or IES model checked against ASHRAE 55's comfort limits.

  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.
  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.
  • The design days: Site → Energy: the winter and summer day, the day–night swing and the internal gains. The first three are set from the place's measured year when you search a place, pick a studio site or load a weather file (its Dec–Feb and Jun–Aug mean dry-bulb, and the seasonal beam sun 8 am–4 pm); otherwise — and always for internal gains — they are the dials you set.
  • The construction: One fixed build-up per role in lib/crucible/energy.ts — a U-value each for wall, roof, floor and glass, and 0.3 ACH of leakage. An envelope sent from the Detailer replaces the wall, roof and floor.
  • The sun's path: Sun positions computed from the site's latitude and the date and hour — geometry, not a measurement: no clouds are in it. The latitude is the Site dial, or the place you searched or loaded.
  • The wind dials: Site → wind speed and direction: the DESIGN wind every score reads. Set from the place's ERA5 wind rose (its mean speed and prevailing direction) when you search a place or pick a studio site, from the survey on a Surveyor Sync, or aimed at the hot, open-window or whole-year hours by the buttons under them — otherwise the values you set. The drawn air reads them only while it follows the dials; following the date, it takes the date's wind instead.
  • The ΔT dial: Site → the inside−outside temperature difference that drives the stack. A design assumption you set — no weather file changes it — and what every score reads. The drawn air reads it only while it follows the dials; following the date, it takes the date's room minus outside instead, which can be negative.

Ventilation relief

°F (a difference)Higher is usually better

Needs: openings (Glazing surfaces)

Reads from: The openings · Your massing · The design days · The construction · The wind dials · The ΔT dial

What it tells you:
The degrees the openings buy back on that afternoon compared with the same building sealed shut — what your ventilation is worth.
What moves it:
Anything that raises ventilation provision.
The tool's own method note, and where its data comes from

Degrees the openings buy back on that summer afternoon: the float with every opening sealed minus the float with them ventilating — what the stack and cross-flow provision is worth in temperature. A difference, not a temperature. Higher = the openings do more. Same assumptions as Summer float and Ventilation provision.

  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.
  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.
  • The design days: Site → Energy: the winter and summer day, the day–night swing and the internal gains. The first three are set from the place's measured year when you search a place, pick a studio site or load a weather file (its Dec–Feb and Jun–Aug mean dry-bulb, and the seasonal beam sun 8 am–4 pm); otherwise — and always for internal gains — they are the dials you set.
  • The construction: One fixed build-up per role in lib/crucible/energy.ts — a U-value each for wall, roof, floor and glass, and 0.3 ACH of leakage. An envelope sent from the Detailer replaces the wall, roof and floor.
  • The wind dials: Site → wind speed and direction: the DESIGN wind every score reads. Set from the place's ERA5 wind rose (its mean speed and prevailing direction) when you search a place or pick a studio site, from the survey on a Surveyor Sync, or aimed at the hot, open-window or whole-year hours by the buttons under them — otherwise the values you set. The drawn air reads them only while it follows the dials; following the date, it takes the date's wind instead.
  • The ΔT dial: Site → the inside−outside temperature difference that drives the stack. A design assumption you set — no weather file changes it — and what every score reads. The drawn air reads it only while it follows the dials; following the date, it takes the date's room minus outside instead, which can be negative.

Indoor swing (energy balance)

°F (a difference)Lower is usually better

Reads from: Your massing · The design days · The construction

What it tells you:
The indoor day-to-night temperature range after the mass has damped the outdoor range. Smaller is steadier.
What moves it:
More exposed mass (a slab the room can see); less heat loss.
The tool's own method note, and where its data comes from

How many °F the inside swings between night and day once the mass evens out the outdoor day–night swing (a loaded weather year's where the tool reads it from one; otherwise a placeholder, 18 °F (10 °C), that some tools let you change). A difference, not a temperature. A simple one-mass model — a resistor–capacitor ('RC') circuit pushed by a smooth 24-hour cycle — with the same assumption as the time constant below: concrete, and 60 % of ALL the floor and ground solid working, whatever the plinth is really made of. Smaller = steadier.

  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.
  • The design days: Site → Energy: the winter and summer day, the day–night swing and the internal gains. The first three are set from the place's measured year when you search a place, pick a studio site or load a weather file (its Dec–Feb and Jun–Aug mean dry-bulb, and the seasonal beam sun 8 am–4 pm); otherwise — and always for internal gains — they are the dials you set.
  • The construction: One fixed build-up per role in lib/crucible/energy.ts — a U-value each for wall, roof, floor and glass, and 0.3 ACH of leakage. An envelope sent from the Detailer replaces the wall, roof and floor.

Time constant (energy balance)

hHigher is usually better

Reads from: Your massing · The construction

What it tells you:
How many hours the building “remembers” a temperature: the heat its mass can hold ÷ how fast it loses heat.
What moves it:
More mass, or less heat loss.
The tool's own method note, and where its data comes from

How many hours the building 'remembers' a temperature: heat the mass can store ÷ the Heat loss rate. Longer = steadier, and the heat balance can use more of the free sun. This is the estimate the energy balance uses: concrete (about 31 Btu/ft³·°F, 2.11 MJ/m³·K) with 60 % of ALL the floor and ground solid working, whatever the plinth is really made of — though a day's heat reaches only the first few inches, so on a thick plinth it overstates the working mass. A comparison number, not a prediction.

  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.
  • The construction: One fixed build-up per role in lib/crucible/energy.ts — a U-value each for wall, roof, floor and glass, and 0.3 ACH of leakage. An envelope sent from the Detailer replaces the wall, roof and floor.

Air changes

1/hHigher is usually better

Needs: openings (Glazing surfaces)

Reads from: The openings · Your massing · The wind dials · The ΔT dial

What it tells you:
How many times an hour the openings can replace the building's air. Fresh air needs under 1; cooling a room by flushing it needs 4–10; above 10 usually means big windows counted as wholly open — set how each one really opens.
What moves it:
As for ventilation provision.
The tool's own method note, and where its data comes from

Air changes per hour (ACH): how many times an hour the openings could replace all the air inside — Ventilation provision ÷ the engine's estimate of the enclosed volume (see Enclosed volume; where that over-counts the room, this reads LOW). Each opening counts as open as far as its sash type allows (the whole pane unless one says less). Read it as a ceiling for flushing heat out, not a comfort guarantee. As a guide: under 0.35 is below the old ASHRAE 62 minimum for a home; roughly 4–10 is the range usually quoted for purging heat or night-flushing mass (a rule of thumb from natural-ventilation design guidance, not a code figure).

  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.
  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.
  • The wind dials: Site → wind speed and direction: the DESIGN wind every score reads. Set from the place's ERA5 wind rose (its mean speed and prevailing direction) when you search a place or pick a studio site, from the survey on a Surveyor Sync, or aimed at the hot, open-window or whole-year hours by the buttons under them — otherwise the values you set. The drawn air reads them only while it follows the dials; following the date, it takes the date's wind instead.
  • The ΔT dial: Site → the inside−outside temperature difference that drives the stack. A design assumption you set — no weather file changes it — and what every score reads. The drawn air reads it only while it follows the dials; following the date, it takes the date's room minus outside instead, which can be negative.

Passive comfort (year)

0–1Higher is usually better

Needs: a weather file

Reads from: The weather file · Your massing · The openings

What it tells you:
The share of the weather file's 8,760 hours your design keeps comfortable without heating or cooling, given how well its openings, mass and winter sun use each passive strategy. A bin count on the psychrometric chart, not an hour-by-hour simulation.
What moves it:
Whichever strategy the Climate sheet's strategy mix says the climate offers most. Needs a weather file.
The tool's own method note, and where its data comes from

Share of the year's 8,760 hours this design stays comfortable without heating or cooling equipment, from 0 to 1 (0.45 = 45 %). Each hour is placed on a psychrometric chart (temperature + humidity) and sorted into simplified versions of Givoni's bioclimatic zones — already comfortable (68–81 °F, not muggy), fixable by winter sun, by moving air, by thermal mass or by evaporative cooling, or needing equipment (the idea behind UCLA's Climate Consultant; the boundaries here are approximate teaching values, not validated ones). A passive zone's hours count only as far as THIS geometry delivers that strategy, against three assumed targets: 10 air changes an hour of the enclosed volume for ventilation, a Thermal mass ratio of 0.45, and 5 idx of Winter solar gain. A count for ranking schemes, not an hour-by-hour heat simulation — whether a room will be comfortable needs EnergyPlus or IES and ASHRAE 55. Reads — until weather is loaded.

  • The weather file: The loaded .epw's 8,760 hourly rows — ERA5 reanalysis via Open-Meteo for a searched place or studio site, the Surveyor's .epw on a Sync, or a file you dropped. The Site panel names which.
  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.
  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.

HVAC hours (year)

hLower is usually better

Needs: a weather file

Reads from: The weather file · Your massing · The openings

What it tells you:
The hours left over for heating or cooling equipment (the rest of the year). Needs a weather file.
What moves it:
The same moves as passive comfort.
The tool's own method note, and where its data comes from

Hours a year the climate forces onto heating or cooling equipment (HVAC — heating, ventilating and air conditioning) after this geometry's passive strategies do what they can: 8,760 × (1 − Passive comfort). A count of hours, not energy use or equipment size. Same model and limits as Passive comfort. Reads — until weather is loaded.

  • The weather file: The loaded .epw's 8,760 hourly rows — ERA5 reanalysis via Open-Meteo for a searched place or studio site, the Surveyor's .epw on a Sync, or a file you dropped. The Site panel names which.
  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.
  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.

◉ Views

View to target

srHigher is usually better

Needs: openings (Glazing surfaces)

Reads from: The openings · Your massing · The view dial

What it tells you:
How much window you see looking toward the View target, as a solid angle in steradians (sr) — how big the glass looks from where you sit. 0.1 sr is about a 3 ft × 3 ft window seen squarely from 10 ft.
What moves it:
Bigger glass on the face toward the view, or bring the eye closer to it.
In V7:
With a focus, the eye is SEATED at the focus, 3′-11″ above the pit floor or the slab the table stands on, and the Eye height dial does not apply. The View verdict and View sheet use standing positions across the whole floor instead.
The tool's own method note, and where its data comes from

How much window you see looking toward the view direction you set, from ONE eye. Where the model names a focus, that eye is SEATED at it — 3′-11″ (1.2 m) above the pit floor, or above the slab the table stands on — and the eye-height setting does not move it; otherwise it stands at the middle of the plan, at the eye height set above the floor. Unit sr (steradian) measures how much of your field of vision something fills: a 3′ × 3′ window seen square-on from 10 ft is about 0.1 sr; the whole sky dome is 6.28 sr. Each window counts area × cos(angle) ÷ distance², weighted by how nearly it lies toward the view direction — a quick estimate that overstates a window you sit right beside. No standard sets a target. Only what your model's windows offer: no trees, hills or neighbours outside (a viewshed study on surveyed context answers that).

  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.
  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.
  • The view dial: Site → the view you want to face and the eye height. Both are your settings.

View openness

srHigher is usually better

Needs: openings (Glazing surfaces)

Reads from: The openings · Your massing · The view dial

What it tells you:
All the wall-opening area you can see from the eye, in any direction (sr).
What moves it:
More or bigger wall openings; the seat closer to them.
In V7:
Read from the seated eye at the focus in V7.
The tool's own method note, and where its data comes from

All the wall windows you see from that same eye, whichever way they face, in sr (steradians — see View to target; about 0.1 sr is a 3′ × 3′ window from 10 ft). Skylights are counted separately below. Higher = more open. Same quick area × cos ÷ distance² estimate; nothing outside the model is tested.

  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.
  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.
  • The view dial: Site → the view you want to face and the eye height. Both are your settings.

Sky view

srHigher is usually better

Needs: openings (Glazing surfaces)

Reads from: The openings · Your massing · The view dial

What it tells you:
How much sky you see up through roof openings (sr).
What moves it:
A bigger skylight, or put the focus under it.
In V7:
Read from the seated eye at the focus in V7.
The tool's own method note, and where its data comes from

Sky seen through roof openings (skylights) from that same eye, in sr (steradians — see View to target), weighted toward straight up. Sky view, not daylight: how bright the room is needs a daylight study (sDA and ASE, the IES LM-83 metrics — Daylighter, or ClimateStudio).

  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.
  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.
  • The view dial: Site → the view you want to face and the eye height. Both are your settings.

⏚ Earth & mass

Buried fraction

0–1Neither way is better — it depends on the design

Reads from: Your massing

What it tells you:
The share of the building's height below ground level (z = 0 in the model).
What moves it:
Sink the form into the ground.
In V7:
Measured against the model's own zero, not the landform. The 4 Rectangles jig stands on z = 0, and an import is set down with its lowest point at z = 0, so both read 0.
The tool's own method note, and where its data comes from

How much of the building's height sits below the model's ground level (normally height 0), as a share from 0 to 1 (0.25 = a quarter). Bed the form in to raise it; perch it clear to drop it. Earth-sheltering steadies temperatures and shelters the walls from wind. Steady geometry, not a ground-heat simulation.

  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.

Soil contact

ft²Neither way is better — it depends on the design

Reads from: Your massing

What it tells you:
The area of skin below ground level — the part coupled to the earth.
What moves it:
As buried fraction; same caveat.
The tool's own method note, and where its data comes from

Skin area below ground level — the surface in contact with soil, which loses heat more slowly than a face in winter air (the heat balance applies only 40 % of the temperature difference to the buried walls and the slab). Each buried skin is counted once: a wall modelled as a two-faced solid has its raw face sum halved — a rough correction.

  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.

Mass volume

ft³Higher is usually better

Needs: a closed plinth solid

Reads from: Your massing

What it tells you:
The volume of the floor and ground solids — in the 4 Rectangles game, the plinth, plus a table or minus a pit. The heat bank behind the time constant and the indoor swing, and the volume slab carbon prices.
What moves it:
Plinth size and height. A thicker plinth adds volume, but see the Mass sheet: most of a thick slab does no daily work.
In V7:
Needs a CLOSED plinth solid on an import. A plinth drawn as one surface reads 0 (flagged) or, if it floats above the model's lowest point, a wrong number with no flag.
The tool's own method note, and where its data comes from

Solid volume of floor and ground material — usually the plinth. It is the heat store behind the Time constant and Indoor swing, and the volume Slab carbon prices. In a model built of solids only CLOSED ones count (a plinth drawn as one surface has no volume). It grows with the plinth; the ratio below divides it by enclosed volume, so widening the plinth moves THIS value even when the ratio holds still. It counts the whole slab, though a day's heat only reaches the first few inches.

  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.

Thermal mass

ratioHigher is usually better

Needs: a closed plinth solid

Reads from: Your massing

What it tells you:
Mass volume × 2 ÷ enclosed volume. The × 2 is not explained anywhere in the engine; use this row only to compare schemes. The Mass sheet's “working mass” is the better number.
What moves it:
As mass volume.
The tool's own method note, and where its data comes from

Floor and ground solid volume × 2.0 ÷ enclosed volume — no unit. The 2.0 is a fixed weighting inherited from the original Eco-Architect engine (it roughly stands for concrete's heat capacity, about 30 Btu/ft³·°F, 2 MJ/m³·K), not a property of your material — so this ratio cannot tell concrete from timber, and it counts the whole slab, far deeper than a day's heat reaches. Passive comfort counts 0.45 or more as full thermal-mass credit (an assumed teaching value). Mass steadies the indoor swing most in DRY climates with a big day–night range.

  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.

◱ Form

Footprint

ft²Neither way is better — it depends on the design

Reads from: Your massing

What it tells you:
The up-facing floor area — in the jig, the whole plinth top including the terrace. A table top is left out; a pit floor is counted.
What moves it:
Plinth width and depth; the plinth solid on an import.
The tool's own method note, and where its data comes from

The floor area the engine reads: the up-facing floor surfaces, seen from above — a plinth or terrace outside the walls included, so it is NOT the room inside the walls. With nothing tagged floor, the bounding box's plan area is used. Every 'per floor area' number — Carbon intensity, internal gains, the slab's heat loss — is spread over this area.

  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.

Enclosed volume

ft³Neither way is better — it depends on the design

Reads from: Your massing

What it tells you:
Footprint × full height, plinth and ridge included — a box stand-in for the air inside. A pitched roof and a terrace both make it read high.
What moves it:
Plinth size; ridge height.
The tool's own method note, and where its data comes from

The engine's estimate of the air inside — not a measured room volume. For a model built of solids it is a BOX: Footprint × the model's full height (lowest point to highest), so a pitched roof over-counts and a courtyard under-counts; a parametric gable counts the room inside its walls plus the roof void instead. Air changes per hour divide by it (an over-count makes them read low) and background leakage multiplies it (so that heat loss reads high).

  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.

Envelope area

ft²Neither way is better — it depends on the design

Reads from: Your massing

What it tells you:
The opaque skin (walls, roof and plinth), counted once. Louvers, terrain and rafters shade the sun but are not skin, so they are left out.
What moves it:
A more compact form.
In V7:
The engine halves the summed face area because it expects thin CLOSED boxes. An imported wall or roof drawn as a single surface reads at HALF its real area — here, in heat loss and in carbon.
The tool's own method note, and where its data comes from

The outer skin: wall, roof and floor surfaces only — louvres, terrain and rafters cast shadows but are not skin. Each skin is counted once: a wall or roof modelled as a closed two-faced solid has its face sum halved (a rough correction: a thick solid wall is over-corrected). It is what Glazing ratio and Surface : volume divide by.

  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.

Opening area

ft²Neither way is better — it depends on the design

Needs: openings (Glazing surfaces)

Reads from: The openings

What it tells you:
Every opening, glazed or not — the area ventilation, views and sun all read.
What moves it:
Aperture width and height dials; the Glazing surfaces on an import.
The tool's own method note, and where its data comes from

Every opening in the skin, glazed or not — the area the ventilation, view and sun rows read. One face of each opening. Equal to Glass area unless an opening is set to OPEN (no glass). In a drawn or imported model the engine sees an opening ONLY where a surface sits on the Glazing layer (or is named glazing / glass / window); a hole trimmed out of a wall is just a smaller wall.

  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.

Glass area

ft²Neither way is better — it depends on the design

Needs: openings (Glazing surfaces)

Reads from: The openings

What it tells you:
The openings with glass in them. This is the area that loses heat as glass and carries glazing carbon.
What moves it:
As opening area, minus anything switched to OPEN.
The tool's own method note, and where its data comes from

The area with an actual PANE in it — openings set to OPEN (no glass) are left out. This is the area that carries a U-value (heat loss) in the energy balance and a double-glazed unit + aluminium frame in the carbon, which is why it is counted apart from Opening area.

  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.

Open (no glass) area

ft²Lower is usually better

Needs: openings (Glazing surfaces)

Reads from: The openings

What it tells you:
Openings switched to OPEN (no glass): a breezeway, a loggia, a doorway. They let the full sun in and air through, and cannot be shut.
What moves it:
The Openings group, opening by opening.
In V7:
The energy balance treats the building as one heated room at 68 °F. Once you unglaze an opening, winter heat demand and heat-loss rate describe a room that is no longer heated — sun, air, view and carbon still read true.
The tool's own method note, and where its data comes from

Opening area minus glass area — the openings set to OPEN (no glass). No pane, so no U-value and no glazing carbon; all of the sun comes through instead of 60 % of it; and it is a permanent air path (see Air path you can't shut). 'Lower is better' is a default, not a law: a breezeway, a loggia or a screened porch wants this area, an insulated room does not. The energy balance is ONE HEATED ROOM held at 68 °F (20 °C), so for a space that was never going to be heated, Winter heat demand and Heat loss rate stop meaning anything the moment you unglaze it (the sun, ventilation, view and carbon figures still do).

  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.

Glazing ratio

0–1Neither way is better — it depends on the design

Needs: openings (Glazing surfaces)

Reads from: The openings · Your massing

What it tells you:
Glass area ÷ opaque skin area (0 to 1).
What moves it:
More or less glass.
The tool's own method note, and where its data comes from

Glass area ÷ the whole outer skin (walls + roof + the plinth/floor surfaces, each counted once), from 0 to 1 (0.10 = 10 %). This is NOT the window-to-wall ratio energy codes limit — for commercial buildings the IECC and ASHRAE 90.1 prescriptive paths cap glass at roughly 30–40 % of the WALL area — because the roof and floor are in the denominator here, so it reads much lower than a window-to-wall ratio for the same building. Panes only: an opening set to OPEN drops out of it but stays a hole in the wall.

  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.
  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.

Surface : volume

1/ftLower is usually better

Reads from: Your massing

What it tells you:
Skin ÷ enclosed volume, per foot. Lower is more compact and loses less heat for the air it encloses.
What moves it:
Fewer, larger, squarer volumes.
The tool's own method note, and where its data comes from

Compactness: outer skin area ÷ enclosed volume — square feet of skin per cubic foot enclosed, shown as 1/ft. Lower = more compact, so less skin to lose heat through for the space you enclose; a cube beats a thin slab, and a courtyard or a spread-out plan raises it. It divides by the engine's volume estimate (see Enclosed volume), so compare your schemes with each other rather than with a published figure.

  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.

⌬ Carbon

Embodied carbon

kgCO₂eLower is usually better

Reads from: Your massing · The openings · Carbon factors

What it tells you:
The carbon released MAKING the materials (stages A1–A3): slab, walls, roof and glass, from the ICE v4.0 database. In kgCO₂e — the unit every product declaration uses. Foundations, transport, repair and demolition are not counted.
What moves it:
A thinner plinth first — in the jig the slab is usually most of it; then less glass.
The tool's own method note, and where its data comes from

Greenhouse gas released MAKING the materials, in kgCO₂e (kilograms of carbon-dioxide equivalent — every greenhouse gas counted as the CO₂ that would warm the planet as much). Stages A1–A3 only: extraction, transport to the factory, manufacture. Material quantities from your model × ICE v4.0 factors, with ONE assumed build-up per part: reinforced-concrete slab, timber-stud wall, timber roof, double glazing in aluminium frames. Each skin is counted once (a closed two-faced wall solid is halved — a rough correction). Leaves out foundations, services, transport to site, construction, replacement and end of life, so it runs LOW against a whole-building count. Compare your schemes with it; a real assessment is a life-cycle assessment (LCA, EN 15978) in Tally or One Click LCA. Kept in kilograms on purpose: every EPD (environmental product declaration), ICE, EC3 and LETI count that way.

  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.
  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.
  • Carbon factors: ICE v4.0 (Circular Ecology) cradle-to-gate factors per kilogram, over ONE fixed build-up per role — RC slab, timber stud wall, timber roof, double glazing in aluminium. An envelope sent from the Detailer replaces the wall and roof build-ups.

Carbon intensity

kgCO₂e/ft²Lower is usually better

Reads from: Your massing · The openings · Carbon factors

What it tells you:
Embodied carbon ÷ floor area, so schemes of different sizes compare. Shown per ft², with per m² beside it on the Energy & carbon lens.
What moves it:
As embodied carbon.
In V7:
The floor area is the whole plinth top, terrace included, so it is carbon per square foot of plinth, not of room.
The tool's own method note, and where its data comes from

Embodied carbon ÷ the floor area the engine reads (Footprint), which includes any plinth or terrace outside the walls — NOT the room's gross internal area (GIA) that LETI and RICS divide by, so it can read well below a per-GIA figure. It also counts less of the building (A1–A3) than LETI's upfront-carbon targets for homes (A1–A5: about 46 kgCO₂e per ft² (500 per m²) for 2020, and 28 per ft² (300 per m²) by 2030) — so never set it beside a LETI target. Compare your own schemes with each other.

  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.
  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.
  • Carbon factors: ICE v4.0 (Circular Ecology) cradle-to-gate factors per kilogram, over ONE fixed build-up per role — RC slab, timber stud wall, timber roof, double glazing in aluminium. An envelope sent from the Detailer replaces the wall and roof build-ups.

Envelope carbon

kgCO₂e/ft²Lower is usually better

Reads from: Your massing · Carbon factors

What it tells you:
Wall and roof carbon ÷ their area — the build-up the skin wears, per square foot. Slab and glass are not in it.
What moves it:
It changes with the wall-to-roof proportion, or when the Detailer sends a real assembly.
The tool's own method note, and where its data comes from

Wall + roof carbon ÷ their skin area: the carbon per square foot of the build-up this massing wears — the same quantity the Footprinter tool reports for ONE assembly. Weighted by area across wall and roof, so it moves when the wall-to-roof ratio does, not only when the materials do. A1–A3 only. Leaves out the slab and glazing, which are not priced per area here.

  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.
  • Carbon factors: ICE v4.0 (Circular Ecology) cradle-to-gate factors per kilogram, over ONE fixed build-up per role — RC slab, timber stud wall, timber roof, double glazing in aluminium. An envelope sent from the Detailer replaces the wall and roof build-ups.

Wall carbon

kgCO₂eLower is usually better

Reads from: Your massing · Carbon factors

What it tells you:
One fixed timber-stud wall build-up × wall area, net of openings.
What moves it:
Less wall; a Detailer assembly.
The tool's own method note, and where its data comes from

An assumed timber-stud wall (studs, OSB sheathing, 5½″ glass-wool insulation, plasterboard, timber cladding — ICE v4.0 factors) × wall skin area, net of openings, each skin counted once. A1–A3 only.

  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.
  • Carbon factors: ICE v4.0 (Circular Ecology) cradle-to-gate factors per kilogram, over ONE fixed build-up per role — RC slab, timber stud wall, timber roof, double glazing in aluminium. An envelope sent from the Detailer replaces the wall and roof build-ups.

Roof carbon

kgCO₂eLower is usually better

Reads from: Your massing · Carbon factors

What it tells you:
One fixed timber roof build-up × roof area, net of the skylight.
What moves it:
Less roof — smaller overhangs, a lower pitch.
The tool's own method note, and where its data comes from

An assumed timber roof (rafters, OSB deck, 8″ glass-wool insulation, galvanized steel sheet — ICE v4.0 factors) × roof skin area, net of roof openings, each skin counted once. A1–A3 only.

  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.
  • Carbon factors: ICE v4.0 (Circular Ecology) cradle-to-gate factors per kilogram, over ONE fixed build-up per role — RC slab, timber stud wall, timber roof, double glazing in aluminium. An envelope sent from the Detailer replaces the wall and roof build-ups.

Slab carbon

kgCO₂eLower is usually better

Needs: a closed plinth solid

Reads from: Your massing · Carbon factors

What it tells you:
Plinth and ground solid volume × reinforced concrete (about 12 kgCO₂e per ft³). Volume-based, so every inch of plinth height counts.
What moves it:
Thin the plinth toward the mass penetration depth (the Mass sheet shows it).
In V7:
Needs a closed plinth solid on an import (flagged “no mass solid” otherwise).
The tool's own method note, and where its data comes from

Reinforced-concrete slab: floor/ground solid VOLUME × (C25/30 ready-mix concrete + about 5 lb/ft³ (80 kg/m³) of rebar), ICE v4.0 factors. Priced by volume, so every inch of a thicker plinth costs carbon — even the part deeper than a day's heat reaches. Always priced as concrete.

  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.
  • Carbon factors: ICE v4.0 (Circular Ecology) cradle-to-gate factors per kilogram, over ONE fixed build-up per role — RC slab, timber stud wall, timber roof, double glazing in aluminium. An envelope sent from the Detailer replaces the wall and roof build-ups.

Glazing carbon

kgCO₂eLower is usually better

Needs: openings (Glazing surfaces)

Reads from: The openings · Carbon factors

What it tells you:
Double glazing plus aluminium frame × glass area. An opening switched to OPEN carries none.
What moves it:
Less glass.
The tool's own method note, and where its data comes from

Double-glazed unit (two panes of float glass) + an aluminium frame × GLASS area. Openings set to OPEN (no glass) carry none of it — no pane and no frame to make. The hole still comes off the wall and roof areas, so unglazing an opening lowers the total by exactly its area × the glazing build-up and changes nothing else in the carbon.

  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.
  • Carbon factors: ICE v4.0 (Circular Ecology) cradle-to-gate factors per kilogram, over ONE fixed build-up per role — RC slab, timber stud wall, timber roof, double glazing in aluminium. An envelope sent from the Detailer replaces the wall and roof build-ups.

☁ Measured climate

Only with a weather file. Passive comfort (year) and HVAC hours (year) also appear here, a second time, with the same values as in the Energy group. Irradiance stays in kWh/m² on purpose: every solar dataset and tool uses it.

Glazing radiation (yr)

kWh/m²Neither way is better — it depends on the design

Reads from: The weather file · The openings · Your massing · The ground

What it tells you:
Solar energy landing on the openings over a year, from the weather file's own sky (sun plus an evenly bright sky), with your massing and the landform in the way. kWh/m² stays metric on purpose: every solar tool and dataset uses it.
What moves it:
Orientation, overhangs, and what stands in front of the glass.
The tool's own method note, and where its data comes from

Solar energy on the openings over the loaded weather year, in kWh/m² (SI by house rule; 1 kWh/m² ≈ 317 Btu/ft²): the sky split into 145 patches (the Tregenza sky), each given its share of that year's direct and diffuse sun, clouds included, shaded by your model and any loaded landform. Diffuse light is spread evenly over the sky, a simplification Ladybug's Perez sky refines.

  • The weather file: The loaded .epw's 8,760 hourly rows — ERA5 reanalysis via Open-Meteo for a searched place or studio site, the Surveyor's .epw on a Sync, or a file you dropped. The Site panel names which.
  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.
  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.
  • The ground: The landform under the site — USGS 3DEP where it is measured, AWS terrain tiles outside the US, a flat plane when none is loaded. The ground banner says which.

Glazing radiation (winter)

kWh/m²Higher is usually better

Reads from: The weather file · The openings · Your massing · The ground

What it tells you:
The same for the winter months (Dec–Feb in the northern hemisphere) — the free heat actually on offer at this site.
What moves it:
Keep the low winter sky clear of the south glass.
The tool's own method note, and where its data comes from

Solar energy on the openings over the weather year's three winter months (Dec–Feb; Jun–Aug south of the equator), in kWh/m² (1 kWh/m² ≈ 317 Btu/ft²), from that year's sky with its clouds, shaded by your model and any loaded landform. The free heat on offer here in the heating season; compare schemes with it.

  • The weather file: The loaded .epw's 8,760 hourly rows — ERA5 reanalysis via Open-Meteo for a searched place or studio site, the Surveyor's .epw on a Sync, or a file you dropped. The Site panel names which.
  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.
  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.
  • The ground: The landform under the site — USGS 3DEP where it is measured, AWS terrain tiles outside the US, a flat plane when none is loaded. The ground banner says which.

Glazing radiation (summer)

kWh/m²Lower is usually better

Reads from: The weather file · The openings · Your massing · The ground

What it tells you:
The same for the summer months — the cooling load you invite.
What moves it:
Overhangs and fins; less east and west glass.
The tool's own method note, and where its data comes from

The same for the three summer months (Jun–Aug; Dec–Feb south of the equator): sun you usually want to keep out. Energy arriving at the glass — not a cooling load, which needs an hour-by-hour energy model.

  • The weather file: The loaded .epw's 8,760 hourly rows — ERA5 reanalysis via Open-Meteo for a searched place or studio site, the Surveyor's .epw on a Sync, or a file you dropped. The Site panel names which.
  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.
  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.
  • The ground: The landform under the site — USGS 3DEP where it is measured, AWS terrain tiles outside the US, a flat plane when none is loaded. The ground banner says which.

Envelope radiation (yr)

kWh/m²Lower is usually better

Reads from: The weather file · Your massing · The ground

What it tells you:
Solar energy on the opaque skin over a year, per square metre.
What moves it:
Compactness; shade the roof and the west wall.
The tool's own method note, and where its data comes from

Solar energy on the walls and roof (not the openings) over the loaded year, in kWh/m², averaged by area, shaded by your model and any loaded landform. The seasonal energy balance does not use it.

  • The weather file: The loaded .epw's 8,760 hourly rows — ERA5 reanalysis via Open-Meteo for a searched place or studio site, the Surveyor's .epw on a Sync, or a file you dropped. The Site panel names which.
  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.
  • The ground: The landform under the site — USGS 3DEP where it is measured, AWS terrain tiles outside the US, a flat plane when none is loaded. The ground banner says which.

Rose-weighted windward area

ft²Lower is usually better

Reads from: The weather file · Your massing

What it tells you:
Windward area weighted by how often the wind actually comes from each of 16 directions — every wind the site gets, not just the prevailing one. Calm hours count as zero.
What moves it:
Turn the long faces away from the directions the rose says are frequent.
The tool's own method note, and where its data comes from

Windward area weighted by the weather year's wind rose (16 directions) — every direction the wind came from that year, not only the prevailing one; calm hours push no wall. Same caveat as Windward area: where walls and roofs are modelled as closed solids it reads up to roughly twice the real windward area.

  • The weather file: The loaded .epw's 8,760 hourly rows — ERA5 reanalysis via Open-Meteo for a searched place or studio site, the Surveyor's .epw on a Sync, or a file you dropped. The Site panel names which.
  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.

Rose-weighted wind load

lbfLower is usually better

Reads from: The weather file · Your massing

What it tells you:
The rough sideways wind force, weighted the same way.
What moves it:
As rose-weighted windward area.
The tool's own method note, and where its data comes from

½ × air density × each direction's average speed² × its windward area, weighted by how often the wind came from it. A massing cue — structural wind loads come from ASCE 7.

  • The weather file: The loaded .epw's 8,760 hourly rows — ERA5 reanalysis via Open-Meteo for a searched place or studio site, the Surveyor's .epw on a Sync, or a file you dropped. The Site panel names which.
  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.

Heating hours

hLower is usually better

Reads from: The weather file · Your massing · The openings

What it tells you:
Cold hours your winter sun cannot buy back.
What moves it:
More winter sun kept; more mass.
The tool's own method note, and where its data comes from

Hours of the year too cold to be comfortable that winter sun through THIS design's openings (after shading) cannot fix — part of HVAC hours. A count of hours, not energy.

  • The weather file: The loaded .epw's 8,760 hourly rows — ERA5 reanalysis via Open-Meteo for a searched place or studio site, the Surveyor's .epw on a Sync, or a file you dropped. The Site panel names which.
  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.
  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.

Cooling hours

hLower is usually better

Reads from: The weather file · Your massing · The openings

What it tells you:
Hot hours your ventilation, mass and evaporative cooling cannot hold.
What moves it:
More ventilation; more exposed mass; less summer sun.
The tool's own method note, and where its data comes from

Hours too hot to be comfortable that this design's ventilation, thermal mass and evaporative cooling cannot fix — the rest of HVAC hours. A count, not a cooling load.

  • The weather file: The loaded .epw's 8,760 hourly rows — ERA5 reanalysis via Open-Meteo for a searched place or studio site, the Surveyor's .epw on a Sync, or a file you dropped. The Site panel names which.
  • Your massing: The solids on the bench — built from the massing dials, or read from your imported .3dm — by role: wall, roof, floor, glazing.
  • The openings: Every surface on the Glazing layer (or a solid named glazing / glass / window / skylight) is one opening, at its own size, height and facing — and, for the air, opened as far as its kind in the Openings group allows (a casement about 90 %, a fixed light not at all). Ones switched to OPEN count as holes with no pane.

Dehumidification exposure

0–1Lower is usually better

Reads from: The weather file

What it tells you:
The share of muggy hours — a moisture load no massing move fixes in this model. It is here so you know, not as a lever.
What moves it:
None in this tool.
The tool's own method note, and where its data comes from

Share of the year's hours too humid (muggy) to be comfortable. No change to the geometry fixes that in this model — drying the air takes equipment — so read it as a fact about the climate, not a design lever.

  • The weather file: The loaded .epw's 8,760 hourly rows — ERA5 reanalysis via Open-Meteo for a searched place or studio site, the Surveyor's .epw on a Sync, or a file you dropped. The Site panel names which.

13Energy & carbon

The right column's second lens: one room's heat balance on a winter and a summer day, then carbon, wind load, earth and form.

The winter day

Two bars on one scale, in kBtu a day. OUT is the heat lost at 68 °F inside: walls + roof, glazing, the ground (tempered), air leakage, and openings with no glass. IN is what the winter sun and the people and equipment inside supply — only the share the building can bank counts; the rest is drawn dashed as gains it can't use. Heating still needed is the gap; Sun pays back is the solar fraction. A paragraph above the bars says the same in words.

The summer day

A 24-hour trace of outside against inside, free-running (no cooling). Floats above outside (vents open) is the summer float; Ventilation buys back is what the openings are worth in degrees; Mass damps the swing to is the indoor swing and its lag; Openings can drive is the air changes. With a weather file: the share of the year held comfortable passively, and the hours left to heating and cooling equipment.

The fixed construction

One wall, roof, slab and window build-up is assumed for every scheme — unless the Detailer sends a real assembly.

The construction the energy balance assumes
PartValue
Wall U-value0.056 Btu/h·ft²·°F (0.32 W/m²K)
Roof U-value0.039 Btu/h·ft²·°F (0.22 W/m²K)
Slab U-value0.053 Btu/h·ft²·°F (0.3 W/m²K)
Glazing U-value0.317 Btu/h·ft²·°F (1.8 W/m²K)
Glass lets in (SHGC)0.6 of the sun — an opening with no glass lets in all of it
Clear-sky sun (no weather file)190 Btu/h·ft² (600 W/m²)
Held inside at68 °F (20 °C)
Background air leakage0.3 air changes an hour
Ground contact counts at0.4 of the air temperature difference

Carbon — making the materials only

Material quantities × the ICE v4.0 database's factors, stages A1–A3 (making the materials), in kgCO₂e — the unit every product declaration uses. Not counted: foundations, services, finishes, transport, construction, repair, demolition.

The carbon build-ups
Build-upLayersCarbon
Reinforced-concrete slabC25/30 ready-mix · rebar 80 kg/m³12.0 kgCO₂e/ft³ (423 kgCO₂e/m³)
Timber stud wall (studs+OSB+glass wool+plasterboard+cladding)studs + plates · 11 mm OSB · 140 mm glass wool · 12.5 mm plasterboard · 19 mm timber cladding1.45 kgCO₂e/ft² (15.6 kgCO₂e/m²)
Timber roof (rafters+OSB+glass wool+galvanized seam)rafters · 15 mm OSB deck · 200 mm glass wool · 0.6 mm galvanized steel2.47 kgCO₂e/ft² (26.6 kgCO₂e/m²)
DGU (2×4 mm float) + aluminium frame2×4 mm float glass · aluminium frame (≈2 kg/m²)4.29 kgCO₂e/ft² (46.2 kgCO₂e/m²)

The rest of the lens repeats Readout keys with metric in parentheses: carbon intensity, envelope carbon, windward area, wind load, buried fraction, soil contact, footprint, envelope area, glazing ratio and surface : volume (section 12).

14Goals and Improve

A goal turns an intention into a number to hit. The cards score live as you move anything.

  1. In Goals, under Forces — draft a goal from the form, press a move. It drafts a goal from the design as it stands: “Admit winter sun” asks for 1.25 × today's winter gain. A must-pass goal is a hard limit; the rest are targets to push toward.
  2. Move sliders (or re-import). A card fills as you close the gap and turns solid black with ✓ when it is met. “Minimise” and “maximise” goals are never “met” — they only show progress. A goal whose number is missing is never met.
  3. Jig only: Improve toward goals searches every slider with a genetic optimizer (NSGA-II: 40 designs × 30 generations) and proposes up to three designs in words — the biggest slider moves and which goals improve ▲ or worsen ▼. Apply moves the sliders there; Undo goes back one step. After Apply, trust the live goal cards and the verdicts over the candidate card's ▲ ▼ — the search scores with a simpler sun — no landform, and its sun-gain numbers skip the shadow check.
  4. Brainstorm goals with Claude (signed in) turns a sentence — “cozy in winter but never bakes in summer” — into proposed goals.

Every move on the deck, and the goal it drafts:

The goal moves
ForceMoveDraftsKind
☀ SunAdmit winter sunWinter solar gain >= 1.25 × nowmust-pass
☀ SunReject summer sunOverheating ratio <= 0.9 × nowmust-pass
☀ SunWarm the focusWinter sun on the focus >= 1.25 × nowtarget
☀ SunShade the focusSummer sun on the focus <= 0.5 × nowtarget
☀ SunFace the equatorEquator-facing gain >= 1.2 × nowtarget
☀ SunCalm the whole skinWhole-skin sun <= 0.9 × nowtarget
↯ WindShelter the windward faceWindward area <= 0.8 × nowtarget
↯ WindCut the wind loadWind load proxy <= 0.85 × nowtarget
≈ AirDrive the stackStack ventilation >= 1.25 × nowtarget
≈ AirOpen to the breezeVentilation provision >= 1.4 × nowmust-pass
⚡ EnergyCut the heatingWinter heat demand <= 0.75 × nowmust-pass
⚡ EnergyHeat with the sunSolar fraction >= 1.3 × nowtarget
⚡ EnergyTemperate summerSummer float <= 0.8 × nowtarget
⚡ EnergyRide the massIndoor swing (energy balance) <= 0.75 × nowtarget
◉ ViewsOpen the viewView to target >= 1.3 × nowtarget
◉ ViewsFrame the skySky view >= 1.2 × nowtarget
⏚ EarthBed into the groundBuried fraction >= 0.5must-pass
⏚ EarthAdd thermal massThermal mass >= 1.2 × nowtarget
⏚ EarthPerch above gradeBuried fraction <= 0.1target
◱ FormMake it compactSurface : volume <= 0.9 × nowtarget
◱ FormTread lightlyFootprint <= 0.85 × nowtarget
⌬ CarbonSpend carbon per square footCarbon intensity <= 0.9 × nowtarget
⌬ CarbonHold a carbon budgetEmbodied carbon <= 0.85 × nowmust-pass
⌬ CarbonLighten the skinWall carbon <= 0.9 × nowtarget

15What each reading is good for

The same words as V7's own panel, What each reading is good for — how far to trust it, and when to use a professional tool: what to design with, and what is left out and who checks it before a number goes in front of a client, a contractor or a jury.

What each reading is good for, and what somebody would have to check
ReadingUse it forWhat it leaves out, and who checks it
SunComparing shapes, orientations and overhang depths. The geometry is exact: rays go from nine points on every window toward the sun and test whether your building, or the loaded ground, blocks them — the same kind of ray test Ladybug runs. So deepening an overhang by six inches really does move the number.Screening estimate. The right physics, simplified. Good for comparing your schemes with each other; never quote it as the building's number. It samples the sun at only ten moments: 8am, 10am, noon, 2pm and 4pm sun time, on the shortest and on the longest day of the year. It counts only DIRECT sun from a cloudless sky: no cloudy days, no light from the rest of the sky, no reflections, and no sun before 8am or after 4pm (usually the west window's real problem). Its numbers are geometry — how much window faces the sun, and how squarely — not an amount of energy. Load a weather file and the Year sheet adds the energy that reaches each window, in kWh/m² a year. Who checks it: an hour-by-hour simulation over a whole year on the site's weather file, with the real glass in it — Ladybug and Honeybee (free, in Grasshopper), ClimateStudio, or EnergyPlus. Switch to that before anyone quotes a heating or cooling figure, sizes a shading device, or fixes an overhang on a real project.
AirRanking schemes, and seeing which effect a scheme leans on: wind pushing air through, or warm air rising and pulling it out (the stack). The verdict uses BS 5925, the British code of practice's standard hand method for natural ventilation, with the windward/leeward and inlet/outlet split drawn for you. In V6 and V7 the Air layer checks it two more ways. One is EnergyPlus's airflow-network equations: the method energy modellers use, here for one room at one moment. The other, when you press Run CFD, is a coarse CFD run (computational fluid dynamics: a computer simulation of the moving air itself).Standard method. The same calculation a professional tool runs, here on your simplified model. Good for design decisions; check it in the professional tool before it goes to a client or a permit. The flows, in cfm (cubic feet per minute), are the standard method. The air changes an hour on the card divide them by the engine's box volume — in 4 Rectangles the whole plinth top × the model's full height, two to three times the room's air — so the ACH reads low: treat it as a screening estimate. Its inputs are assumptions, not measurements. The wind is the Wind speed dial: a placeholder 11 mph (5 m/s) from the west until a weather year sets it, and even then an open-field reading from 33 ft up, used as if it reached your window unslowed. The stack runs on the Inside − outside ΔT dial, which you set. The wind's push is one standard figure for every shape. Each opening counts as open as far as its sash type allows (the whole pane until you pick one in the Openings group), with a sharp-edged hole's flow; an insect screen or a sash held part-way passes less. And the building is ONE room: a door, a partition or a corridor changes the answer completely. The Air layer's EnergyPlus method gives each wall its own push and slows the wind to window height; its CFD is a coarse, wind-only check (no warm air rising), not a CFD study. Who checks it: a mechanical engineer, with EnergyPlus's airflow network run hour by hour on the real rooms and doors (through Honeybee, OpenStudio or DesignBuilder), CONTAM (free, from NIST, the US National Institute of Standards and Technology) for several rooms, or a full CFD study (OpenFOAM, SimScale or DesignBuilder's CFD). Switch to that before anyone relies on natural ventilation to keep people comfortable, sizes a fan, or claims a building needs no air conditioning.
MassThe lesson itself: how much of a slab actually does any thermal work over a day, and how heavy construction behaves against light. How deep a day's heat soaks in comes from standard heat-conduction physics, with material figures from the standard tables of CIBSE (the UK's Chartered Institution of Building Services Engineers).Classroom proxy. A stated hand method for teaching. Trust the direction (better or worse), not the number. It treats the whole building as one lump of stored heat behind one path for heat loss, pushed by a smooth 24-hour temperature swing, and in the 4 Rectangles game it counts the whole plinth top, terrace included. It ranks; it cannot tell you the temperature in a room. Carpet or a rug over the slab, insulation under it, furniture, and how much water is still in the concrete can change the answer as much as the choice of material does, and none of them is modelled. The time lag stops at 6 hours, which is this simple model's ceiling, not a measurement. Who checks it: an hour-by-hour simulation that follows the heat through the real floor layers, including what sits on top of them — EnergyPlus or IES-VE, run by an energy modeller. Switch to that before you promise a room will stay cool without air conditioning, or quote an indoor temperature.
ViewOrientation, and whether a view still reaches you from across the room. The angles are calculated exactly (how much of your field of view each window fills, a solid angle), from up to 81 standing positions on a grid inside the walls, and where the view you want and the sun you do not want collide is stated in degrees.Screening estimate. The right physics, simplified. Good for comparing your schemes with each other; never quote it as the building's number. There is no world outside your model: no neighbouring buildings, trees or hills in the way, so it reads the windows, not what they look at. And this is not daylight. How BRIGHT a room is comes from a daylight simulation, as two numbers. sDA (spatial daylight autonomy) is the share of the floor that gets enough daylight — about 28 footcandles (300 lux) — for at least half the working hours of the year. ASE (annual sunlight exposure) is the share that gets too much direct sun. Those two, from IES LM-83 (a standard of the Illuminating Engineering Society), are what a LEED daylight credit asks for (LEED: the US Green Building Council's rating system). Who checks it: a viewshed on surveyed context (QGIS, or Ladybug with the neighbours modelled) for what you would really see; for how bright the room is, this toolkit's Daylighter for a first read, then ClimateStudio or Honeybee-Radiance. Switch to that before you promise a client a view, or claim a LEED views or daylight credit.
YearTwo questions it has the right shape for: what an overhang does across a whole year, and what a neighbour you model costs you in sun. It takes the weather year's own direct and diffuse sun, spreads it over a sky of 145 patches, and fires rays from a grid on every window at your massing and the loaded ground, month by month.Screening estimate. The right physics, simplified. Good for comparing your schemes with each other; never quote it as the building's number. The light from the open sky is spread evenly over it, which no real sky does — a real sky is brightest around the sun. Nothing bounces light back off the ground or a wall, and the glass is not modelled: the number is what arrives AT the glass. The weather is also just ONE year. A studio site, a searched place or a saved coordinate brings 2023 as ERA5 modelled it: a weather model fitted to observations on a grid about 15 miles (25 km) wide, not a measurement at a station. An .epw you drop is usually a TYPICAL year stitched together from one station's records. Neither is the extreme weather heating and cooling equipment are sized for. Who checks it: an hour-by-hour simulation with the real glass, under a sky model that knows where the sky is brightest (the Perez sky Ladybug, Honeybee and ClimateStudio use), on a station's weather file for the site. Switch to that before an annual solar or energy figure goes to a client, into an energy model, or into a code check.
ClimateA fair first fact about a place: how many hours of its weather year the outside air sits below, inside and above one comfort band (68–79 °F, 20–26 °C), before your building does anything at all.Screening estimate. The right physics, simplified. Good for comparing your schemes with each other; never quote it as the building's number. It reads the air temperature alone: humidity, the temperature of the surfaces around you, sun on your skin and any breeze are all left out, and each one changes whether a room feels comfortable. The band is this tool's own fixed band, not a comfort standard. Its degree-days count from the band's edges (68 and 79 °F), NOT from the 65 °F base every published heating or cooling degree-day uses, so never set them beside an almanac's figure. The psychrometric chart behind “held passively” uses a slightly wider box (68–81 °F, 20–27 °C) and drops muggy hours, so the two are close but not equal. And the weather is one year: ERA5's modelled 2023 from a studio site, a searched place or a saved coordinate, or the .epw you dropped. Who checks it: the comfort models of ASHRAE 55 (the comfort standard of ASHRAE, the American Society of Heating, Refrigerating and Air-Conditioning Engineers; its adaptive method is for rooms with windows people open), in Climate Consultant (free, from UCLA) or Ladybug's psychrometric chart, on a station's weather file. Switch to that before you claim a building will be comfortable without heating or air conditioning.
EarthSeeing how a massing sits on a landform. It shows how steep the ground is, which way it faces, and where water runs and stands. It also gives the rough size of the cut and fill a level pad would need, with an error bar. Slope, aspect and the water paths use the same algorithms GIS software (geographic information systems: digital mapping) uses — Horn's slope and D8 flow.Screening estimate. The right physics, simplified. Good for comparing your schemes with each other; never quote it as the building's number. The terrain is real measured data, at screening detail. In V6 and V7, and in V2 or V5 after a Surveyor sync, it is USGS 3DEP elevation (the U.S. Geological Survey's 3D Elevation Program; lidar where 3DEP has it), one height about every 26 ft (8 m). Otherwise, and outside the US, it is terrain tiles (made from 3DEP in the US), 13–33 ft (4–10 m) between samples. Or it is your own ground imported from Rhino (V6 and V7), as detailed as your file; the ground banner says which. At either spacing a retaining wall, a ditch or a curb is invisible to it. The cut and fill assumes vertical sides with no slopes to hold them back, no topsoil stripped off first, no rock, and no cost of carting soil anywhere. The water model sends every drop to its steepest neighbour: no soil type, nothing soaking in, no pipes. Who checks it: a licensed surveyor's topographic survey, a geotechnical report, and a civil engineer's grading and drainage design (usually drawn in Civil 3D). Switch to that before any earthwork is priced, a drain is sized, or a grading permit is drawn.
EnergyA steady seasonal balance. It ranks moves you can feel — more glass facing the sun, more mass, a tighter shape, a taller stack — and it gets the direction of each one right. The sun reaching the glass is ray-tested against your massing and the loaded ground.Classroom proxy. A stated hand method for teaching. Trust the direction (better or worse), not the number. This is not an energy model. It is one room with one fixed wall build-up (unless the Detailer sends yours). It uses a TYPICAL day per season — the average of the winter or the summer months — not the much colder and hotter design days that equipment is sized for. It has no heating or cooling equipment, no humidity, nobody switching lights on, no hour-by-hour weather, and no measured air leakage (it assumes 0.3 air changes an hour). Its floor area is every up-facing floor surface: in the 4 Rectangles game, the whole plinth top, terrace included. Who checks it: a full energy model with the real assemblies and the ASHRAE design conditions — EnergyPlus (through OpenStudio, Honeybee or DesignBuilder), IES-VE, or PHPP (the Passive House Planning Package) for a Passive House — plus a load calculation (Manual J, the Air Conditioning Contractors of America's method for a house). Switch to that before you size heating or cooling equipment, show energy-code compliance (the IECC, the International Energy Conservation Code, or ASHRAE 90.1), claim a LEED energy credit, or quote an energy bill.
CarbonRanking schemes on one footing — the factors for MAKING the materials in ICE v4.0 (the Inventory of Carbon and Energy, a published database), one fixed build-up per component — so a thinner slab or less glass shows up honestly, in kgCO₂e (kilograms of carbon-dioxide equivalent). It is the same quantity × factor arithmetic a professional carbon tool does, with generic factors.Screening estimate. The right physics, simplified. Good for comparing your schemes with each other; never quote it as the building's number. It counts only making the materials (what the standards call stages A1–A3). Not counted: foundations, services, finishes and fixings; getting anything to site or building with it; replacing a roof in thirty years; or taking the building down at the end. The walls and roof are an assumed wood-frame build-up unless the Detailer sent you a real one, and the slab is always reinforced concrete. This tool's good line, 27.9 kgCO₂e/ft² (300 kgCO₂e/m²), borrows the number of the 2030 target for homes set by LETI (the London Energy Transformation Initiative). But this total leaves so much out that a result under it is NOT a LETI pass: LETI also counts getting the materials to site and building with them. Nor is it a pass against a RIBA target (the Royal Institute of British Architects'), which counts the building's whole life. Who checks it: a whole-life assessment (EN 15978, the European standard for a whole-building life-cycle assessment, or the method of the RICS, the Royal Institution of Chartered Surveyors) in One Click LCA or Tally, using manufacturers' own product data (EPDs, environmental product declarations); EC3 (free) compares real products' EPDs. Switch to that before you claim a LEED life-cycle credit, meet a client's or a city's carbon target, or choose between real products.
The focus and the seatPlacing the table or the pit, and setting the sill above it. The sun on the focus is CALCULATED cell by cell, not sketched: every cell of a window's patch of sunlight fires a ray back at the massing and the loaded ground, so the pit's own rim, the walls and the roof all count. It is a calculation on your model; nothing here was measured on a site. The seated eye reads the view a person sitting at the focus actually gets, not the view from the middle of the room.Screening estimate. The right physics, simplified. Good for comparing your schemes with each other; never quote it as the building's number. It is still direct sun only, from a cloudless sky, at five moments a season (8am to 4pm sun time on the shortest and on the longest day), landing on the focus's own surface. That is right for WHERE the sun falls and wrong for how bright it is there. Who checks it: a daylight simulation on a real weather year — this toolkit's Daylighter for a first read, then ClimateStudio or Honeybee-Radiance — or a lighting designer's study. Switch to that before a hearth or a reading seat is promised its winter sun.

16Units and terms

Everything is imperial first; where a figure came from a metric source, the metric value follows in parentheses so you can check it.

Units and terms used in Proving Ground
TermMeaning
idxA comparison index. For the sun keys: square metres of glass turned squarely to the sun, averaged over the sampled hours.
sr (steradian)How big something looks from the eye — a solid angle. The whole sky is 6.28 sr; 0.1 sr is a 3 ft × 3 ft window seen from 10 ft.
cfmCubic feet of air per minute.
ACH, 1/hAir changes per hour: how many times an hour the room's air is replaced.
ΔTA temperature difference — °F here means degrees of difference, not a thermometer reading.
Btu/h·°FHeat lost per hour for each degree between inside and outside.
kBtu/dayThousands of Btu a day (1 kWh = 3.412 kBtu).
kWh/m²Solar energy per square metre — kept metric on purpose; every solar tool uses it.
kgCO₂eKilograms of CO₂-equivalent — kept metric on purpose; every carbon database uses it.
Penetration depthHow deep a daily temperature swing reaches into a material. Mass deeper than this does no daily work.
Decrement factorThe share of the outdoor day–night swing that arrives inside (0 to 1). Smaller is steadier.
Time lagHours the indoor temperature peak trails the outdoor one.
Time constantHours the building takes to forget a temperature: the heat it stores ÷ the heat it loses per hour.
Windward / leewardThe side the wind hits / the side sheltered from it.
Inlet / outletFor the stack: low openings take air in, high openings let it out.
Effective open area (A*)Two openings in series act like one smaller opening; the smaller of the two governs.
Degree-daysHere, counted against the comfort band's edges (68 / 79 °F) — not the usual 65 °F base, so not comparable to published figures.
A1–A3The carbon stages for making materials — raw material, transport to the factory, manufacturing.
Weather file (.epw)A year of hourly weather. A studio site or a searched place brings one measured by ERA5 (via Open-Meteo).

17Draw it your way

How the model is drawn — every layer's weight, blend, colour and line weight, the wind's colours, the marks, the cut — is the DISPLAY. Save one under a name and put it back with one pick. Nothing here changes a number.

Save a display, and put it back

Everything is in Viewport display. The dropdown in the box’s header lists your saved displays and the starting points below; pick one and the viewport is drawn that way. Set the box by hand and the dropdown says so; save what you have under a name (a blank name saves as Display 1, Display 2 …) and it is in the dropdown from then on. When one of your displays and a starting point draw the same picture, the dropdown names yours.

  • What a display keeps — everything the box draws with: the layers and their weights; every layer’s blend; the pen’s, hidden lines’ and silhouette’s ink colour and line weight, and the hidden lines’ dash, gap and opacity; the shaded model’s colour and shadow; the solar palette and the floor light; the wind’s colours, line weight, opacities and blend, and whether its lines and plane are on; the marks, tags, sun paths and the overlays’ size and weight; the section’s poché and ghost. It also keeps three things every viewport shares: how the air is drawn (the plane at a height or on the cut, what it shows, the slice and its resolution, the lines round the building and how many of them, the eddy behind it, the lines along its walls, the lines through the walls, the motion), the section cut itself, and the render’s sky, exposure and materials.
  • What it never keeps, because each one changes a number: the sun’s date and hour, what the solar paint measures, and the wind’s mechanism, flows, terrain, solver and quality.
  • In a split bench a display goes on the viewport you clicked into. Its cut is that viewport’s Cut: a display with no cut unticks it and leaves the other viewports cut as they were; one with a cut ticks it and moves the shared plane. How the air is drawn changes only if the display draws the air, and the render’s sky only if it recorded one — and when either reaches another viewport, the announcement says so.
  • Where they live: in this browser — not in your Work, and not on another computer. 12 at most, newest first (a13th drops the oldest, and says which); saving under a name you already used replaces that display, and says so. A display saved before September 27 still works: it puts back what it recorded and leaves everything newer as it is.

The starting points

1–5 are John’s own drawings of V7 — set from his screenshots of this box, setting for setting (5’s screenshot starts part-way down the box, so three of its settings were read off the picture). The rest are the drawings a studio asks for most. Each sets every layer, so picking one gives the same picture every time, with two exceptions: the air plane’s height stays where it is (it opens at the site’s eye height), and 1–3 leave the section cut as it is (4 and 5 cut in plan; the rest turn the cut off); none of them touches the render’s sky, exposure or materials. The last column is how many times it draws the model a frame (see below).

The Viewport display's starting points, what each is for, what it sets, and how many times it draws the model a frame
Starting pointWhat it is forWhat it setsExtra passes
1 · Burnt render on lightened sun, full ink, airRendered at full weight in COLOUR BURN over the solar paint in LIGHTEN, with pen, hidden lines and silhouette over the render, and the air drawn.solar 100% lighten · rendered 100% color-burn · pen 1.5px hidden 0.25px dash 6/4px · silhouette 1px · model 100% · air lines purples 1.4px1 sheet — the model drawn 2× a frame
2 · Luminosity render on the solar colour, full ink, airThe render's LIGHT only (luminosity) over the solar paint's colour at normal — the picture that keeps both — with the ink and the air.solar 100% · rendered 100% luminosity · pen 1.5px hidden 0.25px dash 6/4px · silhouette 1px · model 100% · air lines purples 1.4px1 sheet — the model drawn 2× a frame
3 · Overlay render over the air and its speed planeRendered at 85% in OVERLAY over the shaded model and the air, the speed plane on at eye height, the lines round the building and through the walls; no solar paint, no ink.solar off · rendered 85% overlay · pen off · silhouette off · model 100% · air lines purples 1.4px + planenone — one pass
4 · Plan cut: pen and air on a ghosted modelThe model's faces off, the pen alone at 1.5 px, the air and its speed plane, cut in plan at 50% with the cut poché'd and the removed half ghosted.solar off · rendered off · pen 1.5px · silhouette off · model 0% · section Z 50% poché ghost · air lines purples 1.4px + planenone — one pass
5 · Air slice on the plan cut, under a hue renderThe air's speed plane laid on the plan cut at 50%, poché'd and ghosted, with the render's hue laid over the whole drawing and the pen on the model.solar off · rendered 85% hue · pen 1.5px · silhouette off · model 100% · section Z 50% poché ghost · air lines purples 1.4px + plane1 sheet — the model drawn 2× a frame
Line drawing — full ink on an evenly lit modelThe pen, dashed hidden lines and a heavy silhouette over a white model under the flattest sun (shadow 0): the faces an even light grey, the ground shadow at its lightest; no paint, no render, no marks.solar off · rendered off · pen 1px hidden 0.5px dash 6/4px · silhouette 2.5px · model 100% · air offnone — one pass
Solar study — paint multiplied, inkedThe solar paint MULTIPLIED over the shading so the hue holds, a light pen and silhouette, the sun marks and the day arcs.solar 85% multiply · rendered off · pen 1px · silhouette 2px · model 100% · air offnone — one pass
Air study — see-through model, speed planeThe model at 35% so the room's air reads through the walls, heavier air lines and the speed plane, the wind and opening marks.solar off · rendered off · pen 1px · silhouette 2px · model 35% · air lines purples 2px + planenone — one pass
Shaded model onlyThe bench's own shaded model and its marks — every layer off.solar off · rendered off · pen off · silhouette off · model 100% · air offnone — one pass
The bench as it opensThe shaded model with the solar paint at 85%, its marks and the day arcs — what the page draws before anything is touched.solar 85% · rendered off · pen off · silhouette off · model 100% · air offnone — one pass

Move the layers, like Adobe’s panel

Every row of the Layers table but the shaded model has a grip (⠿) and ▲ ▼. Drag a row above or below another, or step it with the arrows — they are the keyboard’s, voice’s and a touch screen’s way to move it (“Move Pen up”). The shaded model is the locked Background: it stays at the bottom, and only the Rendered layer can go under it, where the render becomes the backdrop of the whole drawing. The order is part of a saved display, and every starting point puts the bench’s own order back. It is still a model, not a collage: a wall hides what is behind it whatever the order — the order says which layer is laid over which where both show. The bench’s own order draws the model once a frame; each layer you move out of it is drawn apart, on a sheet at its place, one more draw of the model a frame.

A blend on every layer

The layers stack, bottom to top: the page · the render (when it is set under) · the shaded model · the solar paint · the air · the render (when it is over) · hidden lines · silhouette · pen · the tags — the Layers table, read from the bottom up, with the pen on top of the ink. Every layer takes the render’s 16 blends — normal, darken, multiply, color-burn, lighten, screen, color-dodge, overlay, soft-light, hard-light, difference, exclusion, hue, saturation, color, luminosity — and the solar paint takes those and add (Photoshop’s Linear Dodge). A blend lays a layer on the picture UNDER it, so one mode reads differently over the shaded model, over the solar paint and over a render.

Black ink and the blends

The pen, the hidden lines and the silhouette are black until you give them a colour, and some modes do nothing with black. That is the arithmetic of the mode, not a broken layer:

lighten, screen, color-dodge, difference, exclusion — black is the mode’s “no change”, so a black line disappears everywhere.

hue, saturation, color — a black line takes the colour out of what is under it: a grey line over the solar paint or a render, and nothing on the white page.

overlay, soft-light — it disappears on the white page and is only a faint line over the model’s warm white; it darkens properly only over dark or strongly coloured paint. color-burn — it disappears where the page is pure white and is black everywhere else.

normal, darken, multiply, hard-light, luminosity — black.

Give the ink a colour and every mode has something to work with.

How many times the model is drawn. The model is drawn once a frame. Counting up from the bottom, every layer the graphics card can blend while it draws stays in that one pass — the shaded model at normal, the solar paint at normal, multiply, screen, add (or darken or lighten at 100 %), the air at normal — up to the first layer that cannot. From there up, and for every line layer laid over a render, layers are drawn apart on SHEETS the browser blends; neighbouring normal layers share one. Each sheet is one more draw of the whole model, every frame, in that viewport. With four viewports or on a slow laptop, fewer sheets is a smoother model. The Viewport PNG, the sun study, the layered SVG and the films are put together from the same sheets with the same blends, so a file is the picture on screen.

The wind: colours, weight, opacity, blend

The Air layer has two parts and each can be on by itself: the lines (the moving paths) and the plane (the speed at one height, or on the cut). Turn one on while the layer is off and the layer comes on with that part only; turn the last one off and the layer goes off. The line weight is the room’s path line in px (1.4 px to start), and every other air line is drawn in proportion to it. The lines and the plane each have their own opacity, and the layer has a blend like every other. Every colour ramp runs light → dark from still air to fast, so slow air fades toward the white sheet; the key over the model and the key in the Viewport PNG take the ramp you picked.

The wind's colour ramps, still air on the left, fast on the right
RampStill → fastWhat to know
Purplesthe default — the one hue the solar paint does not use
Greysink only — clear of every colour in the drawing, and it prints in black and white
Bluesshares its blues with the solar paint's cold end
Greensshares its greens with the middle of the solar paint
Redsshares its reds with the solar paint's hot end, and its dark end sits near the age plane's browns
Viridisperceptually uniform and colour-blind safe, but it spans the solar paint's yellows, greens and blues
Magmaperceptually uniform; its still end is nearly the white sheet, and it passes through the solar paint's oranges and reds

The solar paint has its own palettes, little sun → most sun:

The solar paint's palettes, little sun on the left, most on the right
PaletteLittle sun → mostWhat to know
Spectralthe default — Ladybug's convention, blue (little sun) to red (most); not perceptually even, and it shares hues with every wind ramp but Purples and Greys
Viridisperceptually uniform and colour-blind safe: equal steps of exposure look like equal steps
Magmaperceptually uniform; near-black in shade to pale yellow in full sun — reads like light itself
Heat (yellow–red)ColorBrewer YlOrRd: pale where little sun lands, deep red where most does — the 'hot spot' reading
Greysdark in shade to white in sun — ink only, prints in black and white, and leaves colour to the wind

18Taking it out

Studies live in your browser; the board is made in InDesign from what you export.

  • Viewport PNG — the picture on screen. For the board, set Set the viewport to a scale (in Saved views) or pick a Scale in the export row, and the plate comes out at a true architectural scale for InDesign.
  • Sun study (nine plates, PNG or layered SVG for Illustrator) and three animations (sun, turntable, both).
  • Layered package — the Viewport PNG taken apart for Illustrator: every layer its own layer, the lines (pen, hidden, silhouette, contours, sun arcs, the air’s streamlines) as strokes cut where the viewport hid them, the solar paint and the air’s plane as one filled shape per class of their key, tags as live text; the shaded model, the ground and the render as pictures. Each layer is also a transparent PNG in the zip, with a README.
  • Rhino .3dm (the massing) — surfaces / mesh, in feet, on the same 9 layers the importer reads. An imported model goes out as mesh only. Watertight STL at any model scale, for printing.
  • Openings CSV, Year CSV and Site-forces JSON — the numbers, for a spreadsheet.
  • Each force sheet has ⤓ SVG and ⤓ PNG under the pills; they export the sheet at the top of that pill.
  • Save this study keeps the design and where it stood: the location (the coordinate, the landform's centre, extent and seat, how the ground is drawn), the weather it was read with, and a Rhino ground if one was under it. Loading the study puts it back on that site — fetching the landform and the weather again if you have moved since, and re-fetching nothing if you have not. A weather file you dropped is not stored: its season dials come back, and you drop it again for the Year sheets. Studies saved before 2026-09-27 kept no place, so they come back onto whatever site is loaded, and say so.
  • …and, since 2026-10-01, everything else about it: every viewport's display (layers, weights, blends, colours, the section, the air's drawing, the render's sky), the 1 · 3 · 4 layout, the camera with its projection and lens, the date and hour, the ground's whole drawing (surface, colour, map layers and their options, extent, seat) and the wind terrain class. Loading the study puts all of it back. Studies saved before then change the design and leave the display and camera as they are.
  • Save (top right) is the same as Save this study: one click, no account and no Work to pick. To hand the whole project to someone else, use Export → Project file or the Share link.
  • Share this study (beside Prev / Next) sends the study on the bench — its model (an imported one too), its place and ground, its display, camera and date — as a link (signed in) or a file (no sign-in). Whoever opens it gets it in their own studies and on their bench, ready to work on; a file opens under Share link → Open a project file. Opening the same link twice replaces the copy rather than adding a second.

Something here names a button you cannot find? The tool moved on and this page did not — tell John (jsclark2@illinois.edu). The general manual for every tool is How to use.