The mostly agreed upon formula is to take the rear wheel h/p and rear wheel torque that is produced by the dyno test and divide those numbers by .85 to derive crank h/p and torque fr a manual transmission car.
Also, it is important that whatever numbers that are generated by the dyno are "corrected" from the "raw" numbers according to SAE standards to account for atmospheric conditions present during your test. For example, a car tested on a hot humid day would make less power than one tested on a cool, low humidity day. The "correction" algorithm is supposed to account for that, but it is not perfect, and the further you get away from "perfect" SAE conditions, the more error is introduced. At least this is how it was explained to me by the Dynojet people.
There are different manufacturers of dynos out there (e.g. Mustang, Dynojet) and there are differences between individual machines produced by the manufacturer - I have experienced this personally. The dynojet located closes to me happens to measure a bit lower than some others i have seen, based upon mine and some other people's experiences under nearly identical atmospheric conditions. Things such as how much air pressure are in your tires at the time of the test may unduly influence the results.
So, probably the best advice is to use dynos for measurement of RELATIVE gains before and after making a particular mod on your car. Use the same dyno if you can. Unfortunately there are a number of folks who will for whatever reason, broadcast pretty incredible dyno results.
These cars DO make at least the power they are rated at and usually a bit more...
As always, it is "caveat emptor"
