Are you able to confirm if categories are based purely on average estimated times or a blend of average time and power?
Why do I ask? If purely based on time, how will you:
A) categorise someone who has never ridden the course before
B) basically the same as above, categorise people when Zwift release a new world?
They said that they rank each course at a different difficulty ex. 1-6 so even if you have never ridden that course, they will have data on a course of the same difficulty.
The race in C5 (19:30) was not very good for me
3 reasons:
I was in the same race.
I think a lot of the C racers were caught out by the start. I was nearly dropped myself. 2 minutes of 4+w/kg was a bit spicy and that starting pace was worthy of a full B cat race. It dropped down to 3w/kg afterwards though which doesn’t happen in a B race.
Yes it felt like a group ride. Probably because for the first time ever there wasn’t a group of sandbaggers pulling the race along. In a race with even competition and 3 times up the reverse KOM I would expect people to take it easy and hammer it up the hill.
I can’t speak for anyone else but I just assumed you were a miscategorised rider and had I no intention of chasing. And if you weren’t a miscategorised rider and you want to wear yourself out on ill-timed attacks then I’m not going to try to stop you. There was no point in pulling at FTP on flat terrain in a hilly race. Whenever I rolled to the front I just soft pedaled because there was no point in trying harder. The climbs were always going to be the decider so I think most people were trying to save energy so they could attack the climbs.
Personally I lasted until I was dropped on the final KOM climb. The first two times I could match the jump in w/kg (from 3 to 5w/kg) but on the final lap it seems there was an extra 1 w/kg needed which was a step to far for me. A better result than if I was in a normal B cat race. As a low B I would have been dropped the first time up the KOM.
My question has been and remains: what is the purpose of adding additional factors that you have already published to race ranking, and how do you project these factors play out in the different race scenarios Zwift supports in contrast to pure ranking? I.e., what are the hypotheses you’re testing here? Why is this better? (And to be clear, I am not saying it cannot be, I want to understand why and how you think it will be.)
…Unless the only reason to add factors is specifically to obfuscate the TrueSkill curve and testing is trying to ensure that the modifications don’t cause too much drift, I guess.
And post scriptum, while I appreciate that WTRL often gets unnecessary flak and probably worse, I continue to be disappointed with the antagonistic tone that comes up again and again. I get that things might get frustrating, especially under pressure, but I think we all want this (and other WTRL endeavors) to succeed, and for better or worse one part of that is PR.
1 & 2: What I was getting at is that a result based finish doesn’t mean the top racers always win. That a mid pack person could win since the pace is pulled back besides the attack points and the sprint finish.
3: based on my intensity factor of 1.00 for the race, I was in the right category. I was hoping there would be more fireworks not just sparklers…
Did the west C3 race and found the race a good time for even being on planned easy week with legs not feeling the best in between base endurance blocks so I was not going all out the way through the race. First lap with the main group and a few PRs, second lap in the chase group, third lap hard solo effort. Was unsure about going against As in the group. Found it was not so bad. Nice not having rocketship blast off start as usual and more sensible. Not sure if it was the course or the field of racers for main reason or a combination of both. Thou points race changes that fast, from just a regular race the tests were. Good test for the 1min and 5min as the downhills messed up the 20min data readings. This will be fun starting next week. Hope the classics series goes well.
Raced Women’s C2-Wild West. There was only 4 us in the event, but we seemed fairly evenly matched (finished in a bunch sprint). We were able to exchange some banter during the race, so weren’t really going all out - probably just as well no-one attacked as I would have been out the back very quickly!
My usual timezone is APAC (DownUnder) and riders who normally beat me were classed as C3 in that zone, so not sure how the classification will pan out going forward for the Classics. Across all timezones there were very few women in C1 and C2, so maybe given the smaller pool of riders a smaller number of classes for women might make sense?
Thanks for all your effort WTRL - really appreciate all the work you do to create opportunities for fair and competitive racing for women.
To all the people asking for more explanation please, give them time on this.
We’ve been asking for them to do exactly what they’re doing now, so I’d like them to focus on the data they’re gathering and trying different things to make it work.
I think now is the time for data gathering and for Zwifters to feedback on feel; later will be the time for us to ask how a new system works and what didn’t work.
With that in mind, when this process is complete I’d like to see Zwift provide some kind of detailed Q&A/interviews. But until then I’m happy for them to just give simple details.
Reading between the lines, I think it is less an attempt to obfuscate the TrueSkill curve and more that a ranking system is much more complex to implement, and it makes sense to apply some shorter term improvements first - especially seeing as they want to use it for the Zwift Classics series which is starting imminently.
TrueSkill also needs to be licensed for commercial use, so I would imagine that would could some delay too.
I was in that race too, nice to race with you @Mike_Naessens!
I have pretty much the same comments on the race. I was shocked that the start wasn’t balls-to-the-wall for half a lap and that the entire group stayed together! The first split happened first time up the KOM, I was anticipating a split there and made that front group of 10. 2nd lap was pretty chill but I didn’t quite have the legs to stay with the first 5 riders the last time up the KOM and finished in the next group of 4.
I felt the racing was hard but fair, and IMO this is a good first step in the right direction. Hats off to WTRL for getting the ball rolling on this - I look forward to racing the Classics in the upcoming weeks if this autocat test is anything to go by!
EDITED for typos
I couldn’t agree more. Complex metrics based on weight and power numbers only contribute to an already flawed system. I am a fairly light weight male athlete and I’ve been “that rider” who was quickly promoted to B in the past, as I was having to consistently bury myself just so I wouldn’t get lapped during the race. I stopped racing on Zwift long enough to get downgraded and now I’m a C level racer. I also race IRL (USA Cycling Cat 4) and my most recent C level results are equivalent to my IRL results. Based on my Autocat placement, I was mixed in with a great deal of B level racers yesterday. Bottom line, I was several minutes off the leading pace during yesterday’s C4 Middle Earth race. Right from the start, I was producing on average, 3.8 + wkg (beyond my limits) and I was “yo-yoing” just to stay on the back of the group before we even hit the first KOM. At which point, I blew up and got dropped for good at the base of the hill. In the end, the B racers controlled the pace and I found my result very typical of those when I raced as a categorized B racer due to the ZP methods of categorization. I would like ZP to heavily consider placing more weight on individual race result points as their primary method of automatic upgrading and/or downgrading requests of the racers.
Category upgrade/downgrade based on race results seems to be in the current design.
I saw in my race yesterday that the top finishers were the same that has been in the top of B cat for a long time. These are sprinters with a W/kg well within the upper B limit. With the avg speed based system they will be able to “under perform” and get course speeds that will not trigger an upgrade. Moving these racers up based on results will solve this.
How do we decipher the start times for the various Zwift Classic events? 1626228000 means absolutely nothing to me. I really like to ride fairly early in the morning and hope one of the start times is accommodating…
I assume that will be fixed soon WTRL are aware of it. It is a formatting issue.
They need to get their web dev to sort that out but i believe it should be as below if you translate them into readable times,
in UTC
NZ 7am
AUS 9am
Far East 10am
West Aus 11am
East Euro 4.30pm
Cent euro 5.45pm
West euro 6.45pm
Amer east 11.00pm
Amer Cent 00.00am
Amer West 2.00am
Thanks Gerrie! Looks like I’m going to be riding the West Ozzie event as that translates to 4am here in California. A noontime UTC start would be perfect but I’ll make the effort to make that start. I truly look forward to the experience.
It is all fixed now.
What will the protocol be for racing up one classification? Will this option be available prior to Stage 1?
I can resonate with this as I haven’t raced in a long time for the same reason. I plan on getting some racing under my belt and then join the next round of autocat events to test it out. I suggest it’s all about trying and giving feedback and hopefully the ‘time’ element trumps the ‘w/kg’ element of the autocat system.
As a lighter rider in the situation mentioned, I totally agree here. The system needs massive tweaking to account for us.
I raced in the C1 cat (mostly A+s) as i was assigned (im currently a strong B), but was easily dropped at first KOM and ended up doing a TT to come in last 5 mins behind.
I would have been much more suited to C2 or C3