What I don't want is an engine that makes 650 on the dyno and 600 in the car.
An engine will almost always show HP and torque numbers on the dyno that are higher than the ones it achieves in the car. This is because the dyno numbers are corrected to a standard set of weather conditions. Standard dyno correction is to assume sea level atmospheric pressure (29.92" barometer) and 60 degrees Fahrenheit inlet temperature. This is done so that comparisons between engines, or parts on a particular engine, are standardized. If you are any higher than sea level, or any higher in inlet temperature, the actual torque and HP numbers will go down.
On the other hand, if you are running at sea level, and with a high pressure system overhead, and the temperature is cooler than 60F, you will be making more power than what is shown on the dyno.
The dyno will show you the raw, uncorrected data as well as the corrected data, and the raw data is the power that the engine is making with the weather conditions at the dyno.
There's a bunch of information in my book about how weather conditions affect an engine's performance. It gives you the formulas to calculate correction factors so you will know how much power you are losing if you are running at, say, 1200 feet of altitude and 75 degrees F, on a low pressure day, compared to standard dyno conditions. On my dyno, at an altitude of 940 feet, I will typically see summertime correction factors in the 3% to 7% range.
Bottom line is that standard dyno numbers (assuming they are accurate) are very useful for comparisons between engines, but you aren't likely to see numbers that high during normal operation.