Sometimes I will do the “so, what’s your excuse?” survey after a workout and think “because it was 32 degrees Celsius,” but there isn’t a Temperature option (or a Hydration option, either…). My Garmin records temperature (and I am pretty sure smart trainers have temp sensors…), so this might be worth feeding into your ML algorithm.
1 Like
- They do for the purposes of adjusting power data as the trainer changes temp thru a workout. But I’m not aware of any trainers that export that temperature data for use as you suggest. It’s purely an internal metric we can’t access.
That aside, I do think it would be good for TR to include and “Environment” option to cover Temperature and Humidity at least.
4 Likes
I know this is an older thread, but I would be curious to know how much impact ambient temperature has on the newly released AI model (and if it would be worth recording workouts on my Garmin instead of via the laptop to capture this data)
It’s currently the middle of Summer where I live, and if I do a workout in the morning its 22 degrees Celsius in my training shed and that’s totally fine, but on Tuesdays I have a short 30min endurance ride which I tend to do in my lunch break, problem is, during my lunchbreak it can be 38-45 degrees in the shed at that time (no amount of fans help with this, even though I have 3).
Obviously, the new AI model is seeing this endurance ride, looking at my heart rate going through the roof and thinks ‘quick, wind back his FTP prediction because this endurance ride is killing him!’ and really, it’s just hot.
Given it’s only 30mins I kind of don’t mind riding in this heat as a way to grab some easy heat adaptation. But it would be good to have some way of inputting the temperature data with the workout (other than notes) or confirming if the AI is even using it, so I can look at including my Garmin for recording, or maybe adding a Bluetooth temp sensor to the laptop?
How much does your HR vary by temperature? I actually had a similar thought today as I was out riding. And it made me wonder how much the AI relies or utilizes HR data to make changes.
So I was outside for the first time in probably 3 months. We’re just starting to warm up here. And it’s still a little chilly. But I was halfway through my ride and noticed my HR was lower than what’ it’s been while doing the same power indoors. So when I finished I took a deeper dive. This ride was around 90 min, avg power 191W, avg HR 111 BPM. Rated easy. Yesterday I did a little over an hour inside, avg power 199W, avg HR 129 BPM. Again rated easy. Then, I went back through other indoor Z2 rides and found my indoor HR is anywhere from 15 BPM up to 25 BPM higher for the same power and RPE. Meaning an easy ride outdoors I’m chilling around 115 BPM, maybe 120 BPM. Inside it’s fairly normal to be above 130BPM or even 140 BPM for the same power. Definitely due to heat. But it got me wondering if it’s messing with the AI system. Because I’ve seen a couple intensity days get changed to easier workouts after a seemingly chill endurance ride inside. I rate them easy because that’s what they are, but my HR is 20 BPM higher. So is the AI seeing this HR and thinking I’m going harder than usual or think I’m fatigued?
+1 on the potential value of this feature. I do a mix of indoor and outdoor training. The indoor stuff is a tightly controlled environment, pretty much always the same. But outside can be anywhere from freezing to extremely hot/humid in central Texas once we get into May/June and beyond. Huge swings in HR based on those conditions. To the point that I stopped wearing a HR strap for years (but have started wearing it again this year).
And another good datapoint for me would be caffeine intake. I use caffeine as a tool for training/racing, but only in select situations. Just like heat, I’ll see a noticeable jump in HR when I use caffeine.
Without knowing the environmental conditions or my caffeine intake, it’s really hard for me to look at HR data from different efforts and make a comparison. I assume AI would have similar struggles without having the context for the HR. The temperature/humidity would be tough to nail down though since it can vary so dramatically over the course of a ride.
Yes, my HR can vary quite a lot depending on the temperature. As one example, I could do a 1.5-2hr ride at 200W outside in the early mornings, and my HR will tend to stay around 125-130 in cool conditions. But a 30min spin at 190W in the Shed at 35 degrees, and I’m sitting on 155bpm.
The ride still feels ‘easy’, and I mark it as such, but obviously, my body is working a lot harder.
Doesn’t seem right for it to affect my FTP calculation so much. One good test might be to not wear my HR sensor for the 30min rides and see what happens
1 Like
I was waiting for a topic like this. Semi-failed (skipped minimal last bits) two workouts of late, last one yesterday. Both days were over 37C. Ask Alex Wilde what riding Epic in Paarl is like in March - imagine it (INDOORS!) in Jan and Feb and doing threshold workouts!
I also got “penalised” a couple of weeks ago for a higher HR when I rated an easy endurance ride as easy, but my HR was elevated because it was something like 39 at the time… it felt easy, but not everyone lives in the northern hemisphere.