The problem, or should I say key to testing is maintaining consistency. This includes measurement tools, Tyre pressures, room temperature, cooling etc.
We encourage all our guys to come for their pre-season testing, and subsequent follow ups through the year. We can guarantee the same testing conditions, motivational abuse/encouragement etc etc. All this will provide an accurate measurement of 20min threshold in those specific consitions.
As mentioned, the difference between on road and indoor results differ between different people, however we work hard to minimise any thermal effect of testing indoors.
Our protocol is outlined here http://www.koolstofcoaching.com/articles/training/how-fitness-tests-improve-race-results
As for Tacx trainers and wattage. we like the Tacx machines, but the wattage they show is only accurate to themselves.
And a final point, when measuring power with a powertap, you will always get skewed results compared to those given by a Computrainer. This apparently has to do with the hub being clasped, and it somehow effecting torque readings. Way above my technical knowledge. We have done some tests and the difference seems to be greatest (roughly 20-30 watts discrepency) at mid range watts (220-280watts) and then the discrepency lessons as wattages go over 300.
- where cycling dreams come true... sometimes!