I think its relatively well known that the dynos as mechancial instruments are open to mis-use etc..
I was fortunate to have a long chat with the chaps and they went through the ways they knew that the system could be unintentionally or deliberately abused..
One of the reasons I used these guys was that being a sad bunny, I've downloaded dozens of their published runs and correlated all their N20 runs with other sources..
I can say that I've not spotted any inconsistences in their numbers viz others viz OEM.
More interesting to me is why there are minor variances against the fact that with today's manufacturing techniques and robotic assembly etc why little variances do crop up.
I think using the same guys time n time again at least you get within their domain and apple for apple comparison if they are professional which I would say they are..when they do a 'dyno run'..they spend 20-30 minutes on what they call a run, running it up and down, part throttle, full throttle assessing the various parameters so I'm content that they are doing an adequate job..
As an aside I have worked with guys who have 5 yes 5 full test cells where they run engines non stop for 100+ hours..so yes you can do it better but it costs..
You could argue that the inertia type systems are more reflective of real world motoring..you are not going to hold one of these engines at 100% power for more than about 15 seconds when you will be past 100mph
