New Ahfs Posted

[Lynn A McCarty]

Originally Posted by Just Observing

I am not claiming I got it figured out, but here is my understanding of the sliding thresh hold.


Racer #1 GT/CA 350 295/322 chevy combo ( A very common combo)

Racer #2 GT/CA 400 Pontiac ( A very rare combo)

These two guys duke it out at both the US nats and Sports nats at Columbus. Both racers go -1.150 under in the class finals at both races. Thus triggering a review under the "new" AHFS.

Since racer #1 combo is very common and there are several runs in the database, say 25, if this combo average is only -.800 under then it would get HP on the sliding scale AHFS.

While Racer #2 with a rare combo and less than 5 runs in the database has to be -1.00 under to get HP! Racer #2 could go to a third race and slow down his car to qualify at only -.69 under to lower his average to below the -1.00 threshold.

In my opinion it should be the exact opposite of what is proposed. Low # of runs in Database should have a lower threshold to recieve HP compared to High # of runs in the database should be a higher threshold to recieve HP.

I dont understand what you mean by higher or lower "threshold". However, to fully understand this is to understand the law of averages. Statistically, an average means that there are as many above the average as below the average.

So.......if you have a good number of cars running a certain engine family, then do you realize how bogus it means that an engine "average" that 1.15 under number? This means that to overcome the cars that go 6 under it takes an equal number of cars to go 1.7 under.

So......why dont we see that much? Sandbagging! They all crop around 1 to 1.15 under cause they know where the upper limit us!!!!!!!!!!!!!!!!!!!!!!!!!! If at the bottom of the curve they are all going a second under, you need to know to what you are looking. "Cramming" of events tells a story. If there isnt a standard bell curve of events, something is adversely affecting the combo.

If you run the statistical curve properly, it becomes very apparent what is happening. A combo with low numbers cannot be "averaged" out. Who is to say where he falls on the curve? He might be average, he might be slow, he might be very talented. There is no way to know, so these cases need to be weighted separately as they dont fit statistically unless there are enough numbers.

AHFS' biggest mistake is they statistically assign the Upper Limit to be the AVERAGE. An upper limit can never be the average. That is statistically and scientifically invalid.