1. Observe — where regularity costs you
Lap time is not lost on your best lap. It is lost in the gap between that lap and all the others.
At Monza, a driver with a 1:51.3 best runs a 1:52.2 typical lap. Of that second, about 0.45 s goes into just two corners: 0.25 s per lap at the first chicane and 0.20 s at the second. The other four corners are already consistent.
SimRace.app measures every pass of every corner and ranks the corners by the time they cost you per lap. You know exactly where to work, in tenths, not impressions.
2. Compare with yourself — your best pass is the reference
The most useful reference is not a stranger's lap. It is you, on your best pass of that corner.
On the slow passes of the first chicane, the brake is released later and the car coasts longer between brake and throttle than on the good passes. The minimum speed drops, and the tenths go with it.
Every piece of advice comes from your own comparison, good passes against slow passes, with the measured difference: braking point, brake release, coasting, throttle. Not a generic rule.
3. Compare cars and setups — with proof
Every driver has a style and a car that suits it. And a setup change is only worth something if it can be measured.
Same pace, different regularity: ±0.56 s lap to lap with one car, ±0.80 s with another. For a setup, a guided A / B / A / B test gives a verdict such as: B is faster by 0.15 s per lap (90% range 0.07 to 0.23 s), mostly at Ascari, or simply: no proven difference.
You choose your car and your setup on evidence, with your driving variation separated from the effect of the setup. You drive with confidence in what is under you.
4. Progress — session after session
Progress means shrinking the gap between your typical lap and your ideal lap, the sum of your best passes.
A regularity curve for every session, your corners over time, and a follow-up of every piece of advice: given on one session, applied or not on the next, and the time it actually gained.
You know what works, what does not, and what to work on in your next session.
Learning, not guessing
A gradient-boosted decision-tree model learns, from all your passes on a circuit, which inputs cost you time. It is validated on sessions it has never seen, and it stays silent until it is reliable: until then, the measured comparison is shown. The engineers receive these measured facts and comment on numbers, not on a single snapshot.
Where you see it in the pit wall
- CORNERS — corners ranked by time lost per lap, measured cause, your best pass against your slowest pass.
- PROGRESS — session by session, your corners over time, comparison by car, advice follow-up, learning status, import of your LMU sessions.
- SETUP — the A/B test: guided protocol and verdict with its margin of error.
- LAPS — real lap-to-lap spread, on representative laps only.
- Coach — at most one message every three laps, on the corner that cost the most, only when the cause is clear.