gottahaveone,
The more I think of it, the more it seems to me that the estimator is 45% of the weight-to-power ratio (45% of the lbs.curb per horsepower). I estimated and looked up a few of the more similar cars.
The BMW Z4 2.5i, 7.3 predicted, 7.1 manual/7.2 auto reported.
BMW Z4 3.0i, 6.1 predicted, 5.8 reported
MX-5 142hp 2447lbs, 7.8 predicted, 8.1 reported
MSM 178hp 2529lbs, 6.4 predicted, 6.7 reported
Honda S2000, 240hp 2835lbs, 5.3 predicted, 5.9 reported
Remarkably accurate considering it only uses two factors. That is interesting in itself, as that implies the most important thing for 0-60 time is weight to power ratio. The other factors that byebmw mentioned only have a few tenths of a second influence. I was wondering to myself if the difference in the Honda is a result of the relatively low peak torque and the small working range (peak torque is at a high rpm, and peak power is only a couple thousand RPM higher). If I were REALLY curious, I might try plotting or regressing peak torque, and RPM working range to see if it might be improved - but it seems like a lot of work just for that extra few tenths of predictability.
I recall at extremes (under 10lbs per hp or something like that) the equation becomes less and less accurate. I think that it is less predictive for things like corvettes and porshches.
So, for the
Solstice, 2860lbs, 177 hp, 7.3 predicted, 7.2 reported.
and for this rumored engine, and knowing that adding a turbo and intercooler is prolly 100 lbs, give or take:
HO Solstice, 2960, "253 hp" [from GMI speculation], 5.3 predicted.
Suspiciously close to the Clay Dean "good numbers" quote.
I still don't know from which orifice he's pulling a g+ lateral out of