Can TR AI give us an estimated 5 minute power

Hi guys, this has probably been asked already, but could AI software also figure out our max 5 min power? That would be handy, and would save doing 5 min all out efforts to test it, which I find even worse than 20 minute ones :joy:

It should be around 118% of FTP, at least it is according to the model used on intervals.icu. Could be a bit higher or lower depending on how big your anaerobic capacity is.

2 Likes

Yes, Mine is 120% based on the intervals model

1 Like

And I just checked wko5 too, it’s 121% there

There is so much variation person to person you’re better off just doing a 5 minute effort. There is some pacing required and it may take you a few tries to get it right, but there is growth in that experience and ultimately it’s only 5 minutes.

3 Likes

There’s quite a range among different people when it comes to the ratio of 5 min power to FTP. And quite a range within the same athlete, depending on what s/he has been focusing on in training. E.g. my ratio was 125% at one point when I was doing lots of Sufferfest that focused on VO2max, whereas atm it’s more like 117%.

I think it’s quite hard to test as pacing such an effort is considerably harder than pacing an effort at threshold. And it’s extremely unpleasant! Both of these are good reasons to have it done by AI. I personally am paying for TR literally just for the AIFTP atm because I find this feature so useful. So it would be really helpful if they added a 5 min power AI feature.

1 Like

It’s going to depend on the person. I just looked up some past tests and I was 125% ish. That could have been tail end of a VO2 block though, didn’t look at the timing when I did it.

You can use these ranges as a pacing estimate, maybe go out at the low end, but bust your ass and go all out that last 60 seconds…

Yes, of course.
If TR does that behind the scenes, they are not saying anything.

At least in my experience that’s a bit low. I had a 7ish-minute Strava segment where I managed around 117–118 % for that duration — after around 40 km in the legs and knowing I’d have to ride back another 40–70 k.

I was too lazy to go down to my laptop, in wko5 it’s 121% for me (time trialer phenotype) but 120% is probably a good ballpark

1 Like

I do 4-5 min efforts most weeks. My current ratio is 123%.

The model can only do so much and relies on the quality of the data provided to it. The best way to get true 5m max numbers, FTP or anything else is always going to be testing it.

1 Like

Or race up a couple of categories on Zwift! :laughing:

1 Like

If you use Strava and a power meter, you can just look at your power curve for the year and see what your best five minutes was. Or if you use TrainingPeaks, you can look at your peak performances there as well. Or like everyone said, just do it…it’s only five minutes. And only the last couple will really really hurt. :slight_smile:

Another advantage of the AI approach is picking up the relatively small changes that accrue over a month.

TR has that functionality as well.

In my experience, the best way to test 5 minute power is to find a roughly five minute hill that you are familiar with/ride frequently. You can then pace based on landmarks rather than time. I know much better what “hammer from here to the rock wall up ahead” feels like than “hammer for 1:45”. Admittedly, this may came from experience as a runner and everything because focused on distance.

Are such small/incremental/frequent ranges actually relevant? How would it affect your training/racing?

Can’t you get this from the pdc? No need for any “ai”

5 minute hill, I doubt I can find anything longer than 1 minute where I live :rofl:

2 Likes

You have a 1-minute hill?!! I’m jealous! We only have Dutch mountains (aka wind) where I currently live.

1 Like

I sure have

https://www.strava.com/segments/944171

:rofl:

1 Like