| There are ways to compensate for things you can't control. A Box-Behnken design of experiment (which I used to do all the time for the pharmaceutical companies) would work like this: 3 cars (they can be different set ups but must be in good working order), 2 days, 3 CPUs, multiple runs on each CPU each day. If someone wants to use more than one program, supply another CPU. If Ash (I'm sorry, we've never been introduced) wants to submit more than one CPU, fine. Then, when the results are known, submit them to each of the CPU companies/people and allow them to respond before the results are published. Further, I would do dyno runs and "user runs" where a person tests a car without knowing what CPU is currently in the car. Numbers don't lie. Benchmark methods lie. I can do the math and supply a stock 94 TT.
|
