A few questions on CP and Power Training

Orange Fish said:

When calculating CP from the equation (Work=CP*t+AWC), are there specific durations that are recommended? I've read that the tests should last between 3 and 30 minutes, and have seen one example that uses 1 min, 6 min, and 12 min. How does one decide on the appropriate duration? If AWC is closely correlated with total work performed during a 30-s Wingate, can data from that test be used to calculate CP? Would this CP value only be useful in training for shorter (around 30-s) efforts then? Or would an overly short test such as the 30-s Wingate overestimate CP?

1 minute is too short to exhaust AWC, at least for me. To produce my guesstimate for CP and AWC, I fished maximal values for 3, 5, 12, and 20 minutes out of the previous month's mean maximal power data, fit a line to it, and was done. This is not fully kosher CP practice; better estimates can be obtained by doing focused efforts of those lengths. But, I selected those lengths because 1) they were between 3 and 30 minutes, and 2) I knew I'd done near-maximal efforts on group rides and in other testing/training at those durations.

My intended application for CP isn't to direct training, so much as to predict performance, especially hypothetically. E.g., if I care about a 20-minute time trial, I can see where my CP is (pretty good relative to my size) and where my AWC is (pretty darned lousy), and use the CP model to figure out whether getting an extra 30% out of AWC (conceivable) or another 10% out of CP (possible, but definitely a stretch this season) gets me more time.

Would anybody else like to see Dr. Coggan's power profile with just two columns: AWC and CP? I guess we've already got CP, which is spitting distance from FT. But 1 minute and 5 minutes are funny durations from which to try to compute AWC.

Quoted post said:

Regarding Functional Threshold Power, I read that it is closely correlated with CP. If this is the case, then can one choose either a CP test or a 40k TT to base training levels on? Since some people may not have completed a 40k TT before or aren't familiar with pacing themselves in such a ride, would a shorter test between 3-30 minutes in duration be a better idea?

I don't personally do any formal testing of FTP, and I don't think I'm alone. A 1hr time trial sounds like a terribly unpleasant way to get in 100 TSS points to me. Instead, I use a combination of CP, 95% of 20min max power, the "drop-off" in the mean max power chart for the last month, etc., to guesstimate FT. Since it's not that important to have a laser-accurate number, and since those three methods always seem to agree within a few watts for me, I just train and periodically review those three estimation techniques. I'm not really sure it's worth rejiggering your zones, etc., unless you've made a 10W or so improvement, and that big an improvement is subjectively noticeable (i.e., L4 feels a little too easy).