Just wondering if it would be possible to get feedback from TrainerRoad at the end of a training block?
I think a lot of people are seeing predictions at the start of a plan, say a 10 watt gain, follow the plan (or not) and then are frustrated when they get to end of the plan and don’t get the predicted result.
Perhaps some feedback saying that I followed 70% of the plan but didn’t complete two of the VO2 sessions to a high enough standard would explain why the 10 watts wasn’t achieved.
And of course, it’s very difficult to know if you’ve followed the ‘10 watt plan’ if the plan is adapting all the time. Which is great but also makes it impossible to know if the original plan was followed.
Building a system that can explain the “why” can be a bigger task than building the system to drive results to begin with. The “why” in an AI-driven system is often “because that’s what the data/patterns show”, so it doesn’t always provide much actional guidance at the individual level. A good analogy is weather prediction where AI is very sophisticated. It’s easy to expose a simple forecast to a user, not so simple to explain why it predicted rain or why the forecast didn’t match the actual weather. And not to say any of that is impossible, just that it’s potentially a big ask. I do think a report card could be useful, just summarizing compliance against a plan even if it can’t be directly tied to performance outcome.
That’s a great explanation which makes a lot of sense.
Maybe it would be good to see where the AI started adapting the plan, like showing what was intended and what actually happened. So I could see the divergence.
Perhaps I ought to keep a closer eye on what ride is in the plan in 3 weeks time and what levels it’s set at.
Echoing this point. I wouldn’t be surprised if the TR team is already developing something like this - in part to save their team from having to answer one-off queries on the message boards (like they’ve done for me too… thanks!)
I think the TR product would be even more “Sticky” if the athletes could chat with it - to interpret things like why the FTP prediction went up/down, whether it’s feasible to reach ftp of X by end of season, what the difference is between different block plans on specific metrics, etc. etc. I’d gladly volunteer to beta test anything along these lines!
