IT should really think about welcoming Older SM's....... Without a new class..

[Ron Earp]

The Miata example isn't a good one as that car just got a weight adjustment using the 'what we know' theory. The most fair way to do it would be to determine a % gain factor...say 1 or 2% (based on actual research, could be more, could be non-existent based on results) and simply add it to all cars with power steering, and then allow them to remove it.

The problem is that PS loss research is very hard to do. Some of it is methodology, but it's also that the loss is probably about the same absolute value as the noise of the measurement. First we'd need to establish a Limit of Detection for the dyno, LOD, which is the lowest value that can be distinguished from noise. But statistically we can't quantitate values at the LOD, we'd need to work at the Limit of Quantitation, the LOQ, which is about three times higher.

http://en.wikipedia.org/wiki/Detection_limit

Where I'm going with this is if we have two measurements, A and B, and we determine that the LOD of our measurement is 2hp, then we can't statistically say with greater than 95% confidence that A is different from B unless the difference is around 6hp (I'm estimating based on LOD = 3*stdev blank / m, compared to LOQ = 10*Stdev blank / m).

This PS stuff is going to be lost in the noise, and probably the reason why manufacturers do not consider it when stating hp numbers.

Still, I'd be happy to try and measure it and I have the skills to get it done.