I won’t dispute data. As a regular TTT racer in WTRL’s Lowest Mocha category, I would note that the power graph on my most successful races features blocks of green and yellow when I’m sitting in the draft, punctuated by spikes of red when I pull on the front. So. Is there a way to account for drafting in your analysis?
If you can win being on (or off) the front all the time and not drafting, surely you should move up a cat. Otherwise, since everyone else drafts, the blocks of green don’t distinguish one racer from another (i.e. you see blocks of green, but then so does everyone else), and we’re right back to having to broadly understand Andreas’ data showing winners/podium placers in below-A-cat somehow consistently working less hard than others.
That’s my point. Maybe a separate dataset for ITTs? That would eliminate the effects of drafting.
@Paul_Smith_1948_HERD, I’m about to set up a scraper against ZP. Ideally I would be able to scrape zwift.com too but that’s going to be far trickier. A longer-term project, but I think I can get somewhere already by looking at ZP’s avg HR recalculated using age (on average it should be reliable) and using a bigger data set. Anyway, once I’m done I can look at anything you ask. There’s plenty of fun ideas to dive into and analyze, not just cheating related stuff.
However, I forgot to mention that, but the WTRL’s were one of the race categories that I actually excluded but mainly because it’s a team race. Points races of other kinds were also excluded.
But yes, I should look at ITT’s exclusively. I have thought about it too and it’s not a bad idea. The cruisers thrive in mass starts though. Some of the cruiser dynamics require groups of other riders being present. Hard to explain in just a few words. You should try yourself and you’ll see. But there should be cruising in ITT’s too, I’m sure. And nobody could blame it on draft if so. Coming up next.
As for the data set I looked at there would be a whole lot of drafting going on in there too, but my reasoning was that everybody will try to do it. Some will be more successful drafting than others, yes, which would funnel them upwards in the placings. But let’s be real. There’s only so much to gain from drafting in Zwift. It does make a difference but not enough to throw people between HR zones. Because what I was looking at in the first study I detail in this thread was significant deviances in HR pattern. And successful drafting doesn’t really look that way. It doesn’t really look any particular way at all, all else equal.
You can take me as an example. I’m a complete #%&$ who never pulls (unless I have a veeery good reason to). When I race fair I don’t pull because I can’t. I’m already on the threshold, more or less the entire race, and if I accidentally end up at the front for a moment, then someone will pass me quickly and take over, on purpose or by accident, because my Watts will be too low to pull the group since I’m always racing with cruisers and/or stronger riders than me.
When I cruise I obviously never want to pull because I need to save some wiggle room in my average 20 min W/kg to be able to hammer my group into the tarmac later without going over limits. So I need to keep my Watts as low as possible. The lower, the more “fun” you can have later.
My two HR patterns are vastly different to me, even though I see far worse pattern deviances when I look at some other racers (which I am 100% certain are intentional cruisers). And the difference is not because of drafting since I always try to draft anyway. You can see an example of such a pair of HR patterns (my own) in a blog post that I wrote.
This might have been mentioned before, but just to raise an alternative concept: wouldn’t a ‘rolling’ category system be an easy fix? All this requires is auto-assigning participants and a way to split participants into more-or-less equal group sizes. This would translate into a group that can be devided in 4.0-3.2 wkg, but another time maybe 4.2-3.6 wkg or 3.7-3.0 wkg.
This allows every rider to occasionally race for a podium; it allows a more natural progression through categories by narrowing the gap between categories, and it makes sandbagging more difficult as the cat limit cannot be anticipated on prior to the race. This could also help to create data that can be used to ‘train’ on for a future points system.