Same air: racing two runs from different days

Same-air correction takes runs recorded in different weather and expresses them all as if they had happened on one standard day. Without it, comparing a cold-night pass to a summer-afternoon pass mostly measures the weather. With it, the comparison is about the cars — which is the only way asynchronous racing can be fair.

The problem it solves

Asynchronous racing means nobody meets. Your run happened on a Tuesday evening in October; the car you are racing ran on a Sunday afternoon in July, four hundred miles away. If those two raw times are simply compared, a good chunk of the result is just whose weather was better.

That is not a small effect. Between a cold, dense winter night and a hot, humid afternoon, the same unchanged car can differ by several tenths in the quarter mile — routinely more than the gap between two genuinely well-matched cars.

How it works, conceptually

Every run is recorded along with the conditions it happened in: temperature, pressure, humidity, and the elevation of the spot. Those collapse intodensity altitude — a single measure of how thin the air was. From there the run is expressed in terms of a shared reference day, so every run in the system is quoted on the same footing.

The reference is the standard motorsports day the sport has used for decades: 60 °F, 29.92 inHg, dry, sea level. There is nothing special about that particular day except that everyone agrees on it — and agreement is the whole value. Because both sides of every comparison are handled identically, the choice of reference cancels out of the result.

What correction is honest about

A corrected number is an estimate, and pretending otherwise would be the easiest way to lose people's trust. It is built from relationships that hold well on average across many cars, not from a measurement of your specific engine's response to air.

• It does not know your car. Two cars with identical corrected times did not necessarily lose the same amount to the air.
• Boost changes it. A forced-induction engine recovers much of what thin air takes, so standard corrections overstate its loss. Applying an unmodified naturally aspirated correction to a turbo car is a known limitation, not a hidden one.
• It cannot fix a bad run. A missed shift or a spun launch is still in there. Correction adjusts the air, nothing else.
• It has uncertainty. Which is why a corrected figure deserves a range rather than false precision to three decimal places.

Why we show a ± band

A corrected time is really a distribution, not a point. The measurement has a floor, the correction has assumptions, and both contribute spread. Presenting a single confident number hides that. Showing the band is less flattering and considerably more honest — and when two cars' bands overlap heavily, the right answer is that the race was too close to call, not that one of them "won" by a thousandth.

Measured, not claimed

Correction only matters because the underlying number is real. Everything in AsyncRacing starts from a recorded acceleration curve — what the car actually did — rather than a horsepower figure someone typed in. People misremember, dynos disagree between shops, and spec sheets describe a car that left a factory years ago. A recorded pass does not have opinions.

Correction is what makes those recordings comparable. Together, that is the whole idea: race the number, and make sure the number means the same thing for everyone.