I first did bear creek 10 years ago and have done it over 30 times and never failed it . I use it as a fitness test as I have so much historical data to compare it with.
So you can imagine my surprise when the all new AI which supposedly has analysed my historical data suggests not doing bear creek as I have a high chance of failing.
Why would it do this ? I’m going to do it any way
It’s possible your FTP has recently increased, which would mean the workout is harder now than when your FTP was lower.
“Data”. FTP changes, fatigue levels, RPE feedback for recent workouts, short-term overreaching potential that it thinks may impact future workouts in a week or two? Various possibilities.
The new AI gave an increase in FTP but I was expecting this as I have just done 3 months of base training and i was going to do bear creek to verify this but I was very surprised when TR suggested not to do it especially as I’ve never failed this work out
I don’t think the AI goes back far enough or detailed enough in our personal training history to make these assessments but is instead doing the assessments based on other athletes failure rate.
Your true failure rate could be as high as 10%, rule of 3 says upper confidence interval is 10%. I think TR doesn’t recommend if greater than 3%. What was it saying the odds were?
Screenshot above says 5.2%… so a notch above the probability levels where TR suggests “not recommended”, since it’s trying to keep you in that goldilocks zone where workouts feel just right.
ie. there’s every likelihood that @john1967 will successfully complete Bear Creek, but he may have to dig a touch deeper than TR’s model suggests is optimal for normal progression throughout a plan where it’s trying to limit overreach and fatigue that might impair progress further down the line.
a.k.a. “Nothing to see here”. There is no issue; we now just have a tool providing insight into anticipated workout difficulty that’s beyond what we’ve had before, and there’s a learning curve involved in understanding how to interpret this info, especially if we wish to make manual choices that override the model, so that we understand the shades of grey vs. a black/white interpretation.
Fair enough. I was just surprised to see the system not recommending a workout I’ve never failed.
Has trainer road not taken into account my previous 3 months of base training ?
It has, that’s why it’s telling you that you have a 94.8% chance of completing it, it’s just that the 5.2% chance of failure is high enough that they don’t recommend it. Maybe you’ve always had a ~5% chance of failure on previous attempts and never knew about it.
Have you been doing over unders or above threshold work?
If you haven’t that is why it thinks most likely you are going to find it hard but there is a chance of failure. I thought you said yourself it’s a benchmark workout for you which I would think would imply hard or very hard.
test the system. Try it, rate it then come back and tell us how it went.
Edit: The great thing about the new system is your ability to test it out and see how well stated probabilities match up with your experience. I’ve done just that with the system and have arrived at the conclusion that it’s fairly accurate on how it perceives difficulty.
Yep I’ll give it a try and see how it goes.
The problem I’ve got is after a 3 month base period done on another platform,i need to start a build block and my TR plan has given me 3 weeks of base workouts that I don’t need so I’m having to alter my workouts from endurance to threshold workouts and TR is resisting.
Are you on a plan? If so, then don’t forget the warning is in the context of that plan and the future work it sees you have lined up - the probabilities (and, indeed, the warning text box) tell you it thinks you’re very likely to achieve it, but given the %fail odds it’s not recommending it in order to protect the quality of future work.
For the previous 30 times, by comparison, it’s just been ‘you want do Bear Creek? yeah, knock yourself out…..’ 
When you set up your TR plan you should be able to back date the start so that the base phases coincide with the base already done, so the first part of the plan done with TR starts at build.
Thankyou. This has solved the problem and I now have a much more suitable plan which I’m going to stick too and see how things develop.
Thanks for that
I’ve solved my problem by back dating the start of my plan to last November. This has put into a build phase starting today which is exactly what I had planned so I’ll be sticking to my plan for this 28 day period and see how it goes.
