Personally I think AI has limited uses but what its good at is handling loads of data in the background 24/7 with minimal human supervision.
With so many Zwift users it time to improve things using AI to build a full HR profile for each rider.
AI could do the analysis factoring in age, sex, height, weight and build a profile that compares you to others in the same set. This would also allow Zwift to pull outliers with unrealistic results because like it or not statistical analysis becomes possible.
The only thing that users may complain about is having to use a HR monitor for every single ride, that includes training, racing and pretty much every single time you are on the bike to build and maintain the full profile accuracy. This would be a small price to pay in getting more verifiable results.
Zwift already has all the data it needs being fed to it, its just a matter of implementation.
I’ll be honest, I’m struggling to see what purpose this would offer.
I don’t need to compare myself to others and we are all so genetically different that you could have two people with the same age, height, weight etc but with complety different output.
The outliers as you call them are detectable by far easier methods
One of the worlds best has a heart rate average of 150 even when others are at 190 just in her draft.
I tested a HR monitor that converted your heart rate into watts. It was a bit weird. I would go very slow to start and then pick up speed once my heart rate was up to race average. Sprinting was not fun. Holding your breath worked better than going very hard.
A person I have ridden with the last few summers is over 70 years old, and I’ve seen his heart rate above 190bpm multiple times, not sure what his max is. He said he used to go well above 200bpm years ago. My max HR is about 183bpm and I am over 20 years younger.
So.. Even “outliers” are sometimes hard to determine when there’s so much variability person-to-person.
Your HR is not always the same for the same power output. The HR just shows how your body reacts to the training today, which can be affected by many things: overreaching in training, too much/not enough coffee, sleep deprivation, mental stress, etc.
I am on the extreme slow end of HR. During a hard interval session, I will average around 103 BPM. I have visited a cardiologist and she told me I was just fine and had an outliner heart rate, with nothing to worry about. I rode the Seattle to Portland, a 205-mile single day event and averaged 108 BPM and 187 watts. I would guess I would be flagged if I were to race.
I’m the exact opposite, maybe an outlier in the other direction (I have no idea and don’t really care). And I rarely use a HRM unless I’m racing.
It’s hard for me to understand why Zwift would do anything like Carl is suggesting as long as they want to let people on $250 spin bikes join races. If they did adopt this proposal, what would they do with the information? They are not generally interested in verifiable results, so… where is this going?