Yeah, I've been a fan of the Autocar tests and have read the way they go about testing. However, I believe that reaction time shouldn't be included in the calculation (or it should be minimised if it is included). Since reaction time is only dependent on the driver, not the car, I don't think it's really fair to keep it in. For example, in the September 2002 issue of Autocar, the reaction time of the driver of the SR3 is 0.63 seconds, whilst the reaction times of the driver for the two opponents are both less than half of the Radical's. This adds another 3 tenths or so on to the Radical's time which isn't really fair to the car. It's worth noting that the reaction times for the drivers of the Jaguar R3 and Gould hillclimb car are 0.20 seconds and 0.16 seconds respectively, proving that racing drivers can minimise reaction times.
In the game, after you've done a few runs you can get to judge when the car's going to hit 100mph exactly and prepare for it. In the runs that I've done, I've not waited to see 100mph on the display (as I believe they do in Autocar), rather I've predicted when I'll hit 100mph and got ready to brake. I always make sure I've hit at least 100mph though, so the car has accelerated to at least 100mph and is braking from at least 100mph. This way, I'm basically factoring reaction times out of the equation (which is how I believe it should be), showing the true potential of the car.
I judge reaction time in the game as the time between the display showing 100mph and the braking bar showing full activation.