This morning I completed Diester on Zwift. During the 5min rest intervals between sets I occasionally spun backwards, or took a full rest without any spinning. My completed workout says: “3:42 of pauses lowered your score.” It downgraded the workout from a 2.3 to 1.9.
Why would pauses during rest intervals lower the score? Especially when the workout instructions specifically advise to backpedal?
When you stop pedaling for more than a couple of seconds, your workout pauses, and those pauses will affect the score of your workout.
Backpedaling is a way to give your legs a break without pausing the workout, but you’re still moving forward in the workout with effectively zero power.
We’d only recommend backpedaling if you absolutely need to. I think we should take a closer look at the workout instructions to make sure that’s clear, especially since we calculate power and paused time a little differently now.
It’s also worth noting that if you’re backpedaling while doing a TR workout in Zwift, the workout will automatically pause, unlike in the TR app..
Thanks for the explanation, Eddie! This highlights a key difference between doing a workout on the TrainerRoad app (or Garmin) and Zwift.
On TrainerRoad, the “Pedal to Resume/Pause” setting allows the timer to keep running even if you stop pedaling. Zwift, however, automatically pauses the workout the moment your avatar stops moving. To keep a Zwift workout active while resting, you either have to keep the legs turning or time your breaks for long descents.
Zwift would definitely benefit from adding a “manual pause” toggle similar to TR’s. In lieu of that, it’d be helpful if TR could discern whether the pause was during a rest interval or a work interval, and adjust accordingly.
This is good feedback! I’ll share it with the team. 
Work and difficulty builds during a workout to achieve the desired training stress. Taking a break or resting defeats that. Don’t take breaks, or if you do, expect the consequences.
Taking a break during work portions sure, but far far less of an impact (if any) from going a bit easier during the recovery portions of an interval in terms of physiological impact.
Strongly agree. Having your score lowered because of a “break” during a recovery interval doesn’t make sense. It’s the point of a recovery interval.
If anyone else runs into this issue:
Shortening rest intervals is one way to make a workout harder, e. g. 30-30s vs. 30-20s vs. 30-15s are all using the same format, but those are getting progressively harder.
Hence, it does make sense to lower the workout score. Of course, the impact on the score should (and I am sure does) depend on the type of workout and the duration of the rest intervals. Extending a 5-minute break to 5:30 minutes should have less impact than extending a 2-minute break to 2:30 minutes.
It’s a feature, not a bug. If you remove breaks, you’ll worsen the quality of the input data and TR AI’s adaptations won’t be as good.
Can we simply turn off Autopause?
Disagree with this aslong as the time doesnt exceed the prescribed rest period.
Doing an extremely low wattage for a set period of time, back pedalling or stopping pedalling entirely will all result in the same thing.
That few mins of 90W when your FTP is 300W isnt going to make a noticable difference to your training stress accross the workout.
alternatively i think you can also change the tr workout to outside, then reupload the fit file (w breaks, doesnt matter) and then match it to the “outdoor” workout. Since outdoor workouts just check if the workout has been completed yes/no if my memory serves me right
Let’s clear up the debate: I’m not talking about how long the break is. Whether it’s 30 seconds or 30 minutes isn’t the issue. The issue is the type of rest: full stop vs. soft/back pedaling for that same window of time. Because you can make the distinction from the data in the FIT file, TR shouldn’t “downgrade” your workout just because you full stopped during a rest interval. That feels like a bug, not a feature.
And many of us disagree with you. Fulling stopping and continuing to pedal are not the same and should not be scored the same. I’m thankful TR does it this way. It is not a bug.
Then why is the workout level not downgraded if you have auto-pause disabled in TR and you full-stop during a rest interval?
That sounds like going for an outdoor ride, excluding zeros, and then claiming your average speed was the number without the zeros. If you want to work around the system, it sounds like you know how.
Nope. That’s different.
Are you sure this is the case? I haven’t seen this happen yet. 
If that’s happening, I’d say that’s probably a bug.
Also, there’s something to be said for active recovery. Building the ability to process lactate while doing some amount of work is a really important function to develop.
While the differences may be small, I’d imagine that if you had two athletes, and one pedaled through all of their recovery intervals and the other rested during them, they would develop different results over time.
Yup, it’s the case. I’m happy to send you two different FIT files that exhibit the behavior:
File A: during a rest interval the user stops pedaling for 2 minutes. The platform stops writing record messages during this break, resulting in a 2-minute “jump” in timestamps.
File B: the user again stops pedaling for 2 minutes during a rest interval, but in this scenario the platform continues writing record messages during the break. The record messages increment by 1-second, but power/cadence/speed are all 0.
In File A, TR will declare the rest periods as “breaks”, and downgrade the PL of the workout.
In File B, TR will not downgrade the PL of the workout.
Same user/rider behavior, different results in TR.
Maybe I’m missing something here.
Why do we think that active recovery at like 80w vs back-pedalling vs passive recovery on the toilet or in the kitchen… has any meaningful impact on training outcomes?
Or why taking an extra few minutes during the rest period would have any meaningful impact on outcomes?
To the best of my knowledge there is almost no science on this in general, and what does exist, suggests that longer rests between intervals promotes more training adaptations, and a mix of passive and active recovery between intervals may be superior.
That would be changing the interval itself, not the rest between intervals..
Lowering the workout score because of a difference in rest between intervals does not make much sense from a physiological standpoint.
Like if my workout is 3x20 at 95% doing that with 10m rests where you coast down a hill or 10m rests at like 30-40% makes little to no difference in what adaptations you will gain from the workout itself.