Originally Posted by
unterhausen
well, the assumption that a rider will ride at exactly the same power level regardless of light choice probably isn't true. But that assumption is a really good starting point. And if the watts that are wasted cause one not to ride with a faster group of riders, then the time penalty could be huge. This is why Steamer contacted you to for your opinion about testing, it's certainly an interesting question to me.
The next step up in more detailed integration is to model the ride in three parts. Uphills, downhills, and flat. So if you have a 200K with 6600 feet of climbing, how about breaking the calc into three parts - 66.6K at 3% grade, 66.6K at -3% grade, and 66.6K of flat. Use 85% of FTP for the uphills, 55% of FTP for the flats, and 25% of FTP for the downhills. Look at the impact of the dynohub at each condition, using a wattage loss that matches the approximate speed of each portion.
This is still not reality, but gotta be more accurate than doing it all as 200K of steady state.
Edited to add: I just realized that I don't totally understand the point you were making in your first sentence.
Originally Posted by
unterhausen
well, the assumption that a rider will ride at exactly the same power level regardless of light choice probably isn't true.
I thought you were commenting on the error introduced by doing the simple calc holding slope and power constant (cause that is obviously not real...). But I realize now you were saying something I bit different (I think).