Can you provide the correlation between TR AI FTP and max 60 min power for TR users? I realize they are different metrics with different purposes. However, I would find the correlation interesting and helpful for race pace planning. Ideally, I would like to see an equation with a R2, but if that is proprietary, maybe just a rough estimate or rule of thumb. Thank you
Edit - Sorry for the unclear post. For the “you” above, I was referring to TR (maybe I should have messaged them directly). To clarify, I am not looking for individual inputs. As others have mentioned, there are many significant variables. Comparing a few individuals or even a few thousand individuals is probably meaningless. However, with a very large dataset, ML can identify correlations (hence the magic of TR). So, my request is specifically referring to the correlation in the TR dataset between TR AI FTP and max 60 min power. I realized the correlation may be low, but I am still curious. Thanks.
Past me, when I started structured training again in December after a few months of off season, coffee rides and some injury rehab - AI FTP was a fair bit higher than what I could have done for 60 minutes but fine for easing me back into some low level interval workouts
Current me, nearing the end of my second block of Base training having done lots of Sweetspot and Threshold workouts recently and built my levels pretty high - pretty confident I could hold my AI FTP (which got a significant bump a few weeks ago when AI went to general release) for close to an hour
Future me, in 2 weeks time when my AI FTP is predicted to jump >6% and I move into my first Build block - AI FTP will likely exceed what I can do for 60 minutes again (unless I have a barnstormer of a recovery week), but interval sessions are shifting over to VO2 which I normally find relatively easier than threshold workouts so should be just about right for training level purposes. (Which now makes me wonder whether AI FTP is at all forward looking I.e. It looks at the upcoming workout types when deciding how high to set it)
In other words, the relationship between the two numbers is going to change not just for different individuals but for the same individual at different stages of their training plan
Interesting to see some of the AI selected numbers.
AI FTP (current): 258W - I adjusted manually to 245W yesterday.
20 minute: 249W - Zwift race so filled with a lot of suprathreshold efforts
30 minute: 244W - same race
60 minute: 224W (NP was 225W for this effort)
None of these efforts were paced - I was mostly just trying to keep up with strong people on Zwift. Still curious how the TR AI chose such a high number initially, though, as there is no supporting evidence for it in my workout history.
One way that I like to look at this relationship is to pull up time trial type workouts like Lola, Reynolds, Dardanelles +1 and Dunderberg +1. The AI will estimate the expected probability of failure, and I figure that whatever workout gives me about 50% chance of failure has to be a pretty good estimate of my TTE at the current FTP. That puts me at about 40 minutes at my current setting.
