RowToTri wrote:
So if you are testing something new like a helmet or shoes or whatever, how do you know the difference you are seeing is not dominated by minor positional changes from the last time you tested? In a tunnel there are ways to control it somewhat, but if you go a week or a month between tests, how do you know?
Well I don't know. My current methodology is based of RChung-style testing on a velodrome. I'll do it about monthly. I'll go out and do a bunch of laps, change something, bunch more laps. Then go back to the original setup as a sanity check, bunch more laps. Maybe a few rounds of that. If there was some combo that had a strong signal for being faster, that's the new baseline.
Then say, two months later, I'll start with that baseline as the starting point. Of course, you're right. I don't measure my body position to make sure I'm in that exact same position. Or maybe when I was doing maintenance I changed my bar position by 2mm. I don't know. It's not wind-tunnel grade.
So I guess the answer is there are ways to approach aero testing: 1) super precise measurement controlling everything possible, get a small of amount of high-quality data 2) decent measurement, shit-ton of data
It's like a classical FTP test vs. WKO mFTP. You can go through the FTP process every time you test, make sure the temperature is the same, calibrate all the instrumentation, do the same warmup, etc. Or you can just go collect a shit-ton of training data and estimate it and get pretty darn close.
With the new devices, I think it might be more like a gradient descent, to borrow engineering terminology. Am I more aero than last week? Last month? There's no need to rapidly converge. Maybe this week I'll experiment with overlapped hands. Maybe next week I'll borrow my buddy's Aerohead. How does my gravel bike compare to my road bike?
I don't think it's a replacement for wind tunnel testing. Or Jim @ ERO-type-expertise. But when you go arrive at the tunnel, for that final optimization, you can bring a lot to the table.