I am a triathlete using TR for just bike training. I plan my own workouts for the swim and run. I had a 90 minute threshold bike planned today, but after a 70 minute easy long run yesterday, AI suggested a 30 minute recovery ride. How does the new AI factor in swims and runs? Should I put my future planned swim/run workouts in the calendar?
I would put your planned workouts in. Have you historically been importing runs in? I’ve heard it can take a while to know how unique a new stimulus is for you.
I’ve imported runs and swims ever since it was available. You got thinking - I’ve never put in the effort score to get TSS though. That’s probably what’s giving me trouble.
I’m pretty sure that you now can go back in time to add RPE, where previously it was locked after like 48 hours. Not sure if it’s worth the effort but it might help.
I guess it has to do with just taking the duration of the run into account and not the easy intensity.
Have a look here, I think these bugs could be related. Unfortunately, I’ve never heard back on those and these are still reproducible:
https://www.trainerroad.com/forum/t/trainerroad-ai-beta-testing/107104/3679
https://www.trainerroad.com/forum/t/trainerroad-ai-beta-testing/107104/2880
(@Caro.Gomez-Villafane Would be nice if you could acknowledge that these bugs are accepted and tracked? I know it might not be top priority for the AI beta release (even though fatigue detection plays a big part of FTP prediction) but at least if I knew this is tracked internally I wouldn’t need to test and mess around with runs on my calendar wondering. Unfortunately Sean never reacted to this request)
So if you feel good because the run was easy, I think it’s ok to ignore the yellow warning.
@coop I’ve been playing with the triathlon plans against bike plans with runs and swims added.
I don’t think you need to enter tss or rpe on runs for it to affect fatigue management. I’m only running once a week, biking three times a week now.
A bike only plan shows yellow days after the hard bikes. Adding the runs and swims lights me up with red days and yellow days for a few week, then those go back to more manageable level of fatigue. This happens whether it’s a tri plan which includes intensity, or manual with none.
Either way ftp prediction drops significantly except on MV Olympic, for me. And that chimes with the ramp up rate plans require.
As FTP Build and MV Olympic have the same ftp prediction I’m sticking with FTP Build and adding runs, swims and lifting manually.
There’s no real need to put all of your planned activities in your calendar that I’m aware of, although doing so will give you a better idea of what your upcoming training will look like.
If you’re syncing completed runs and swims over, we’ll adapt your TR plan based on your fatigue levels, as you’ve mentioned.
Just know that those guidelines are there to prevent you from overtraining, and we’ll give you a longer leash over time as you adjust to your training load and/or we learn that what you’re doing is sustainable long-term. There is the option to turn off Fatigue Detection in your account settings if that’s what you’re after, though.
Sorry! Looking into this! ![]()
Any news on this?
Hey @Rizzi, sorry for the late reply here!
I was able to confirm that this is not a bug but rather how runs have worked since the start of our Fatigue Detection system. However, this is now causing an “issue” with the ability to see what impact they have on the future. All of this to say, this is not how we want it to work permanently, but I don’t have an ETA of when we’ll change it
I’m really glad to know about this. @Rizzi good job picking up on this bug. TR team, I feel like this is something that should be made more clear. The fact that you can input the intensity of your run and get a TSS score for it to me has always clearly implied that that information is being taken into account and used for the various TR tools, predictions, warnings, etc. I’m quite surprised to learn that the impact of runs is only calculated based on duration.
Perhaps you can work on a user guide for what data is/is not currently being utilized for AI FTP Detection, Fatigue/RLGL, etc. Especially for the non-cycling activities. I’d like to allow the TR models to use as much info as possible, but I also want to understand how it’s being used so I don’t misinterpret what results I’m getting or have my cycling workouts scaled back because TR is interpreting my runs or other non-cycling activities as being more stressful or fatiguing than they actually are.
Also, does the same currently hold true with predicted fatigue out forward? If I have a planned 90 minute run at tempo or threshold intensity, does TR give that the same predicted fatigue as if I instead put a 90 minute run at recovery pace?
If predicted fatigue is based on the same mechanism as red light green light then it’s just duration that seems to be accounted for. So unfortunately setting intensity to threshold or recovery makes no difference. Hope TR advances this in the next months.
As I said here:
Run duration is the only factor the AI uses to adapt the plan, correct? Is swim duration included the same way?
