I tend to do maf logs and boost logs when running cars on a dyno to back up the results, 99% of the time they are bang on.
90% of stock TFSI I have ran have made around 8-10bhp more.. I used to worry a little until I started to do VCDS logs with every power run and was seeing the air flow work out to the same as the dyno reading and the actual load a little higher than the requested load.
IMO the issues with the DD rolling road is that the fly wheel hp reading is only an "estimated " reading based the wheel hp readings with a "fixed" % for transmission lossed added to make the fly wheel power.... How can this be right? Where do DD get this % from? and I dont think I have ever ran two Mk5 TFSIs which have had the exact same transmission losses? A ruel of thumb is around 12% but this can vary to 17-18%
The DD is one of the best dynos I have ever used for mapping and working from wheel hp, but power runs and fly wheel power its just a lottery like every dyno.
Some rolling roads worth out the actual losses during the power runs coast down.