It looks like there were some workout failures in the last week. What is the plan over the next 28 for your prediction?
A lot of “failed” workouts as in workouts not completed with them rated as fatigue.
seems like you’re not recovered enough to complete your workouts at stated AIFTP and the system is picking that up and adjusting the intensity down. This dropping intensity is to help you get back on track and complete workouts then build up from there.
Being pedantic, he did say apparently
I bet there’s too much work and not enough rest/recovery ….
In my peak last year I had an LT2 of 400W (I think approx FTP). I’ve clearly detrained over Christmas, but now AI pegs me at 317, with a forecast increase to 333 in a month. I don’t really mind which number it is, but I’m curious how it’ll develop over time, especially when my season starts in April.
My FTP was 211 for some time. I have been taking it easy due to an injury the last few weeks. I kept getting app notification that AIFTP needed to adjust. So it recommended that I switch from 211 to 190. I accepted it. A few days later, this new AI switched me from 190 to 159 lol. So I “lost” 50 W in a span of a few days lol. I don’t see how this new system is accurate. Oh and AI FTP prediction says I’ll be at 138 in 28 days because I have nothing planned lol
I’m not sure why this is controversial. What would you expect to happen to your fitness if you do nothing for a month?
so someone who doesn’t do anything for a month is expected to lose 80W or 50% of FTP? Does that make sense? Because 211 to 130 in a month is a big jump.
What date was your 211 FTP detection, and what have you done since then?
If you don’t train you lose fitness. FTP is not some sort of immutable trait.
Post your training calendar over the last month and your TSS before and since the 211 if you want help.
I’d expect the system to drop your FTP. You’ve done nothing to show the model any evidence to support a higher number. It’s doing what you’d want it to do.
This makes sense, then. I realise you’ve been injured, and that’s unfortunate, but as it’s meant no intensity for over two months (or three and a half, if you’re looking at the drop to the future prediction) it’s no surprise you’re losing threshold fitness. You need to spend some time at or close to that zone to maintain it - a low level of easy endurance won’t do it. If nothing else, just look at your TSS graph, since that 211 FTP detection you’re doing barely a quarter of what you were before.
Yes, there’s been a change in the measurement basis of AIFTP, which dropped you from 190 to 159, but I’d see that as the system putting you at a level where it thinks it can give you the best workouts to get you back on track.
I’d suggest adopting a training plan for the next few months - assuming you’re now able to train again - and see how the workouts feel, and how your predictions look a) after putting a plan in your calendar and b) after completing the first month’s work.
Good luck.
Trainer Road’s explanation makes no sense. It’s not like gettIng a new bathroom scale. They sold us the old scale and the new scale. They are now saying that their old scale was defective, so why should we trust the new scale?
if my bathroom scale was defective, then I would get a new one from a different manufacturer. Is this what we should be doing?
Curious what Cellphones do you use? Have they improved it over the years?
But they’re not saying it was defective, they’re saying that the new one is more precise.
Going with the scales as a proxy it used to measure my weight to one decimal place so I weigh 70kilos, the new one is now able to do 3 places so I’m actually 69.67kg. Both work, but one is more accurate.
Always get them the wrong way round!
For predicting workout difficulty from past workouts, I’m thinking repeatability(precision) matters more. As you want the number to be consistent from test to test
We still don’t really have data to make any claims about accuracy, as it appears that some people went up and others down. With large data sets there will be outliers just like with the Ramp test.