If I understand it correctly in simplified terms, it is basically about two figures that ONLY affect the car and NOT the influence of the chassis dynamometer. These are the two values, firstly the power at the crankshaft, and secondly the power at the drive wheel on the tarmac.
On the chassis dynamometer, BOTH values must be calculated retrospectively, using the data from the drag load measurement.
In the case of measurement at the crankshaft, the drag of the drive train is also taken into account.
In the case of wanting the power number at the tyre, only the drag of the roller dynamometer is added back in.
And all those who write that you should not expect more than a "before and after" measurement are of course right when it comes to measuring only the relative increase in performance due to a change in the engine, best at the same day. Then all the differences mentioned above hardly play a role.
