You’re right that it should be Actual vs Digital Twin for Half-Marathon on the Past Data tab. Did you maybe use “Custom Configuration” event type and Marathon was one of the disciplines you selected? If not please update one more time but if the problem persists please email me with your account email at markus@aiendurance.com and I’ll have a look, thanks!
Hi Markus, your thought’s on the MoveSense heart strap ?
Bruce Rogers seems to be very happy with it’s performance.
[Movesense HR+ Sensor]
(Movesense HR+ Sensor — Movesense)
Thanks.
Hi Markus, given that the Venu 3 does not log HRV correctly, is there any value in daily entering the HRV using the check-in function?
Thanks!
Hi Markus – not to gutter your work (as it were) but I think your algo needs a bit of tuning. At least for me. The VO2 and AC workouts it laid out for me were way low. For ex:
-90sec VO2 intervals that land in my sweet spot/sub-threshold zone.
-20-sec anaerobic intervals that are 300 watts below where I do actually them.
Note: the algo had access to data from some of these recent workouts (where I put up true VO2 and AC power) so it should have been closer.
I hope I don’t come off overly critical (entirely possible I missed something). But I did want to pass along my out-of-the-box experience with the app.
Regards,
Brenon
Hi Roberto,
Agree with Bruce. Movesense and Suunto HRM are great for HRV as well
Hi Duppie,
Actually while the Venu doesn’t log in-activity HRV I think it still may log overnight HR/HRV. For the latter you need to connect Garmin Health and enable the “Daily Health Stats” checkbox. If you don’t want to use that data and are using another app we’re not (indirectly) integrated with you can enter the values manually.
The recovery model works totally fine without DFA alpha 1 but ‘only’ overnight HR/HRV
Hi Brenon,
No problem and thank you for the feedback.
For the zone detection: it may well be that we’re detecting your zones too low. We use a combo of threshold measurements and power distribution over your recent activities to estimate them but depending on your recent activities it looks like that didn’t work very well. You can also manually adjust the zones in the Account settings.
For the AC: we prescribe them rather vaguely above upper limit of VO2Max zone but feel free to do them as high as you can.
I know the readiness index because I have been gathering data from quite a few athletes, and honestly, in my humble opinion, it is not very useful. In fact, several tests on the same day have yielded totally different results. It is a value that can easily be falsified by changing the warm-up. Not all training sessions require the same warm-up. As for alpha-1 at 0.75, since the pandemic, I have been saying that it is not very useful, and I have completely verified that. In fact, the comparisons I have with my artificial intelligence trained to detect thresholds and verified with different laboratory tests are infinitely more accurate than using 0.75 or even 1.0. I started testing AI Endurance when it came out (I test all platforms), and although the idea is good, I think the approach needs to change. I don’t know what has changed since I tested it because in the end, if I try everything, I have to allocate a large part of my salary to applications. I would be happy to know what changes it has undergone.
Hi Luisma,
As we describe here it’s a useful comparison to previous activity’s warmups if the current activity falls within the ‘support’ of the previous data. If not it’s hard to compare indeed.
When you say you have completely verified that: can you point me to this work? Couldn’t find anything in quick search online.
We’re actually looking into respiration rate as well and agree it’s very promising in finding the thresholds. We should have a blog post out about this some time next week.
The “update one more time” worked, now everything is on Half-Marathon. I didn’t do a “Custom Configuration”, but maybe something got wrong when I tested the difference between pace and power. Looking forward to testing the training plan now and figuring out how it works for me.
Thanks!
Fortunately, I have access to a lot of data from athletes and I was able to conduct tests with them, of course not in a laboratory, but in controlled situations by me. It has no scientific validity, but it helps me know that I was not mistaken. By the way, I am Iñigo’s partner in this alpha-1 topic. I believe that we should clearly define what standard warm-up is; in my humble opinion, we cannot talk about a standard warm-up since each training session requires a specific and different warm-up depending on what we are going to demand from the athlete. I am a strong advocate of alpha-1; I have been researching this with my resources for a long time. In fact, before AI Endurance was released, I was using alpha with a script in Python that I developed. I am going to review AI Endurance again to see what has changed.
Thanks Markus,
I’ve got the Heart Sentenial version with the Ritmia app to check my arrhythmias and it’s working well for me.
Cheers
Thanks Marcus!
It now imports overnight HRV from my Venu 3 as well in-activity-HRV using the Edge 1030 Plus.
Hey @Markus_Rummel_AIE,
How is AI Endurance going to dovetail with the fact that I am an endurance athlete that rides/races exclusively on single speed mountain bikes. Events are typically 6-12 hours or multi-day stage races. I know this is a bit of a niche and one of the limiting factors to other AI offering I have tried is lack of support for targeting stage races.
I have a long history of training and racing, some seasons leveraging a coach and others self-coached. This gives me a fairly good idea of what I need to do but I am also a bit lazy. I hoping to find an AI platform that can do 90% of the work for me instead of 40-50%. Let me know what you think.
Thanks,
Dan
Hi Dan,
Thanks for the question. First thing would be if you have power data on your single speed MTB rides? Our models currently only work with power (if you have some rides with only heart rate that’s ok though).
We also currently don’t support stage races as a goal event.
If you still wanna give it a try, there’s a 14 day free trial with no payment info required so you can see for yourself if it’s useful to you or not.
Thanks @Markus_Rummel_AIE. I’ve got over 10 years of power data so that is not an issue. I’ll take a look come the start of 2025 and see how well it works with my event plan for the year.
I notice your co-founder Chris is in Colorado Springs and so am I so that’s a check in the plus column
Hello @Markus_Rummel_AIE,
I am interested in trying out an AI plan. I am wondering : is the algorithm able to take previous training seasons into account (activities and hr/power long term response and changes) to optimise and induvidualize the plan ? Does it adapt according to short term hr / pw response to training ?
Thanks !
Yes, we take your past data into account. Mostly power but we use HR/HRV as well. We usually go back as far as we can, for example Garmin allows up to two years of past data.
@Markus_Rummel_AIE The functionality looks very complete but the UI is unfortunately a bit buggy.
- Generic error on password reset
- Full screen error after changing availability and then logged-out
- Deleting a workout takes a very long time, or it shows an error
- Almost all workouts are imported twice from Garmin
- Have a look at the console, it shows 1500+ errors
- Can’t find an option to import a workout and other basic functionality
If you need a tester, let me know!
Thanks for the updates @Markus_Rummel_AIE I use Apple Watch Hrv and imports to intervals with elitehrv. Can you leverage this?