-
If I was only interested in a number for FTP, I'd probably just do a ramp test.
To be fair, even the 20 mins is arbitrary. You could just as well just get on the bike and rode as long and hard as you can, then infer an FTP from your power distribution curve for the ride.
As long as it's a max effort, it should be good. The pacing wouldn't be that important.
The problem is in how you infer an FTP (as opposed to a CP, which is relatively straight forward), as you end up multiplying by a somewhat arbitrary number to get a 20 minute or 60 minute FTP.
See my question above re 20 min tests in erg mode. Doesn’t really make sense to me but I suppose useful for verification.