Which spot
Every place you marked, read at the same hour.
One reading, run at each spot on the plan, so the only thing that changes between them is where you are standing. It compares what you have marked and nothing else — a corner the model has never been told about has no geometry and cannot be read.
Sit down, eat, and still want to be there ninety minutes later.
Starts at the hour this use normally happens. Move it to the hour that actually annoys you.
The best spot in July is regularly not the best spot in October, and that is the whole reason to check.
Which spot
Dinner outside at 7:00 PM
The herb planter is the best spot here for dinner outside at this hour, 5 points clear of The dining table. Both are losing to what it feels like; The herb planter simply gets less of it.
- The herb planterNo
What it feels like is what costs it — about 46°F against at least 59°F.
- The dining tableNo5 behind on the 0–100 scale
What it feels like is what costs it — about 48°F against at least 59°F.
| Spot | Verdict | Standing | What is costing it |
|---|---|---|---|
| The herb planter | No | leader | What it feels like |
| The dining table | No | 5 behind on the 0–100 scale | What it feels like |
Every spot is read at the same hour on the same day, through the same geometry, so the only thing that differs between them is where they are. The bars are the 0–100 reporting scale, mapped from those readings against what this use needs: they rank the spots, they do not measure them. A gap under five on it is not a decision this engine can defend, and it says so rather than picking one.
The same answer, on the plan
Shadows at 7:00 PM in October. The best spot is not always the one that looks best on a drawing, which is why both are here.
To scale · to the measurements given: Drawn to scale from the measurements given. Not a survey. It is not a survey.
Once you know where
A ranking tells you where to sit today. It does not tell you what to change — for that, the winner and the loser both have their own reading, and a change tested at one is not tested at the other.