In the realm of performance tuning, "the dyno pull" is nto to determine actul output, but rather provide a way to evaluate the results of changes made.
By observing the delta youknow wether you are on the right track or not.
As with any "measuremnet" taken against a complex system - there are obviously many variables able to impact absolute numbers, but you can bank on the changes in output being a good measure of success.
This is true whether dynoing a car or bike, or benchmarking a PC, or measuring top speed or 1/4 mile drag race times.
And yes - even Nurburgring lap times

-- Consistant dupliaction of results or measurable changes are what import