If anyone with a way to measure power (not the Kurt computer though; it just uses their formula) feels like 'calibrating' their trainer for me and reporting back, that would be wonderful! I'm pretty sure my stats ability is up to calculating confidence intervals on this formula they provide. I'd need several participants, though... 3 at the very least? Pick a few wattages, and let me know what your speed is while holding each wattage... How about 150, 200, 250, 300, and 350 watts?

I'll report back with the results, which would be helpful to those of us without powermeters. Much like in the second reply to this thread, a confidence interval would allow you to say (with a certain high degree of confidence) that your power estimate is within a certain percentage or number of watts of your true power. Maybe Kurt trainers allow you to estimate your power to within a few watts, or maybe trainer-to-trainer there is enough variation that the estimate could be out by 100 watts, and is therefore basically useless. Right now, I have no idea.

Thanks