Automatic FTP from Outdoor Ride

I’m curious if the auto FTP calculation uses a continuum of segment durations. Like a Critical Power curve. To figure out an estimated FTP from a ride. For example, my best 5 min sustained power might suggest an FTP of 275W. But my best 15 min might suggest 264W. The longer duration might be more reliable than a short duration.

And is fatigue considered? If you knock out a hard 26 min segment 3 hours into a ride that suggests 281W, is there a skew factor to bump up the FTP estimate above what that 20 min effort suggests (say to 290W)?

I know XERT attempts to apply sophisticated models to recalculate FTP including “near misses for breakthroughs” etc. That is overkill and too academic I think. But a little more insight into how/why TR came up with its auto-estimate would be really cool to see. Maybe generate a small popup report. Such as:

“We estimate your FTP from a best effort of XX minutes on your ride on XX date. This occurred 93 mins into that ride. Because this ride was 12 days ago, we’re adjusted our estimate down 5W.”

How does AI FTP Detection work?

“We’ve used Machine Learning - a subset of Artificial Intelligence - against millions of rides to understand FTP based on a multitude of factors. Our system looks at tons of dimensions of your recent and past training data to understand your most likely FTP if you were going to take an FTP test that day.”

How should I validate my AI Detected FTP?

I get FTP updates on other platforms. What makes this any different from what I see elsewhere?

Most other platforms use what we call capacitive efforts to estimate your FTP using an FTP estimation model 1. This means that they rely on you completing all-out efforts within your training data.

For example, if you had a very hard 20-minute climb during a ride (ridden at ~ 105% or greater of current FTP), they might use that effort to estimate a higher FTP. Our model does not rely on capacitive efforts. Instead, our model predicts your FTP based on your training history and your personal biometrics, along with many other features. Capacitive efforts and FTP tests rely on one performance. AI FTP Detection accounts for much more with a broad range of data points.

What does AI FTP Detection look at to detect my FTP? What types of efforts or workouts does AI FTP Detection consider?

“AI FTP Detection looks at your training history and your biometrics to detect your FTP. Because AI FTP Detection considers so much of your training data, it’s very unlikely a single type of effort or workout would substantially sway the model. You don’t have to do capacitive or all-out efforts for the feature to work. In fact, it works if you only do sweet spot or aerobic riding!”