Ranking based on finishing position using something like the chess ELO system, rather than using power numbers and weight to decide who our opponents are.
2 Likes
I appreciate there has been lots of tangents in this thread, but this has nothing to do with the calculation for the new system, and is a bit boring by nature of it’s repetition. We have something that is stopping sandbaggers and cruisers, that is huge. Yes there is a bigger picture and a Utopia out there, but let’s embrace this first rather than sit around for another 3 years moaning about why Utopia doesn’t exist yet.
By all means propose a ranking / results system in detail on the anti-sandbagging thread, so that we can expose lots of issues with that too. As @_JamesA_ZSUNR mentioned elsewhere, it presents a different set of problems, so different options is the ultimate approach. This model is the first one.
5 Likes
I appreciate that you find alternate suggestions annoying.
But given Zwift’s inability to deliver changes that require engineering, wasting time overengineering a new pseudo-power-based solution that extended how long we had to wait to get it and it really won’t solve much other than show that they can prevent people from joining a category, we are finding the entire discussion annoying.
We can only assume that Zwift really really believes they can predict how people should perform and that is the best solution.
Maybe we need a poll to see how many people in these many threads prefer power modelling to determine categories or a ranking system.
All of the people interesting in only a ranking solution are getting frustrated at the amount of time being wasted. We have been waiting for years. Promised solutions years ago. A test was rolled out years ago for anti-sandbagging and never deployed outside of test events. And now we see another test.
For something, we are not really interested in (other than preventing people from joining categories, which they could have done with the older category system.)
4 Likes
It’s seem odd to only be interested in a ranking system. But hey, if that’s your view fair enough. Opinions are like certain body parts after all.
My preference is a toolkit so that race organisers can split pens by different methods, and ranking would be a good one, but it also is not the Utopia you seem to think it is. Just look at any other online game with incredibly mature and well considered ranking systems - there is still a lot of user frustration. In Zwift specifically, you are likely to create race fields with much broader disparity in ability. OK, you may think this is better, because you are racing for rankings upgrades, and it is partially self-correcting, but just because that is your preference for how races should play out, it doesn’t mean it’s everyone’s cup of tea.
A ranking system that would be up to the task, and then also baking this in to core client, UI, dealing with new and returning riders, etc etc is a substantial piece of work. It should have resource and attention paid to it now for sure, but the system that is being implemented now is a very good one for solving two of zwift racing’s biggest issues - and we’ve got it. It exists. It’s a thing. And we wondered if that was ever going to be the case a few months ago.
I think a roadmap that goes something like:
- CP/MAP categories
- Race organiser boundary config (no more fixed categories)
- New global ranking system for progression
- Pens can be configured by ranking
Seems very reasonable, based on cost/benefit and within the remits of the challenging environment we know exists at Zwift.
I’m yet to actually see someone go to the effort of detailing exactly how a ranking or race results system would work. The principle is obvious, but one you get in to the weeds there are plenty of holes that can be picked, as many are actively trying to do with the biggest development in Zwift racing fairness there has ever been.
8 Likes
If we support such a toolbox , why does that involve parking one of the tools that make the most sense in a competitive sport context (results based classification) while we endlessly keep trying to do the complete opposite which is to not support the concept i.e racing leagues but instead a way to make absolutely sure every race experience is the same as the one the day before , and the day before and the day before because it is based on putting everyones numbers into a computer program and spitting out the same results each time on what riders you will be competitng against.
I DO support the concept of the toolbox and options , but we are a million miles from that here , we are endlessly testing to find the one true algorythm and way to calculate category break points. If we really beleive we should have a toolbox lets work towards that and that means an option for each of the parameters and how an event organiser could combine and use them. I could be wrong but I suspect that is not going to be on a roadmap to happen any time soon . With good reason too as I am sure everyone will reply and tell us , its “very complicated” . and I agree . its far far FAR simpler to just adopt what is proven to work as it is used by cycling affiliations and leagues across the world almost exclusively.
2 Likes
Ranking in one form or another is widely used for IRL bike racing. Effectively only the details of how to achieve rank change from place to place.
Locally our provincial cycling association (UCI affiliate responsible for this area) policy is riders must have two wins or three top-five placings. If less than five riders in a category only two will receive upgrade credit.
The point is that it is simple, easy to define, easy to implement, easy for people to understand.
I’ve been trying all day to think of an IRL sport that uses physical characteristics (other than gender) to define categories. The best I could think of was boxing and wrestling which (if I understand) use weight limits to define categories, but then within those categories use results to rank competitors.
That would be like Zwift races for under 65kg, 65-80kg and over 80kg, but then still with A/B/C/D based on a results-based rank, which actually might make some people happy. And of course implementable by Zwift with information already available (weight.)
1 Like
Well we got Pen Enforcement extended to category from previously only been applicable to gender , which cost nothing and was sitting waiting to just be implemente but it did take how many years exactly … and instead of just adding that we are I remind you getting another week of test events only , so there is no commitment or roadmap published that says that at the end of this week the fact that pen enforcement will work (which it does , it doesnt need any testing to proove that we can only enter events we are allowed ) . it will be rolled out immediately afterwards .
Further the simple roll out of this feature has got conflated with the perceived requirement to tinker with the method of calculation of a users category . No matter how postive or negative this might turn out to be it will by default result in a further delay to acheiving the first step of simply supporting enforcing of entry requirements ( whatever they may be ) . IMO by coupling these two things together our dream of options looks as far away as it has ever been , maybe further because it is being so tightly coupled .
I am not sure I am maybe thinking about ranking in the same way as others , and how you have positioned it ( Global) . What I see is ranking (Points) is something that can be set at many levels and even by event series. Which actually is also already available , certainly as available as the w/kg categorisation that we are “enhancing” now in the great algorithm project… It just needs the ability to pen enforce it .
Here , in Zwiftpower 2 values . Points and Rank .
1 Like
In a race ranking algo, how many races does it take to get someone to where they “should” be in the rank? I presume they would need to slot people first into a category based on their prior riding until they have enough races to have a rank that would boost them out of that category correct? ZP uses a default score of 600, and then works with your best 5 races in the last 90 days (correct?).
For people who don’t race that often… would rank appropriately bump them up quickly enough they wouldn’t be messing with the front group on a category they should not be in? ie. Don’t they need a better initial categorization based on the power curve of that rider in either case? If they defaulted to today’s 95% of 20 min effort would we not still see a group of people coast then win the D/C/B category before they have 5 races in the last 90 days to push them up?
The ZHQ Anti-Sandbagging races were “the biggest development in Zwift racing fairness” when they were brought in, and we all know how that went.
Ridiculous that Zwift didn’t go for the low-hanging fruit of simply switching on enforcement of the existing categories for pen assignment, which would have been a nailed-on, guaranteed step forward.
Separate point - shouldn’t this thread be in the Racing forum? Or just merged into the massive anti-sandbagging thread since that’s where a large number of posts on a ranking system are located?
1 Like
It’s not IRL racing though, so there are a far wider set of options available. IRL racing couldn’t sort by physical metrics even if it wanted to.
Also just because real life cycling federations use results to promote through categories, doesn’t make that system good. I can’t attest to the rest of the world, but in the UK the system is pretty terrible. 4th cat is full of Zwift cat A racers. Probably more than 50% of the field. In Zwift that would be a terrible experience for most riders (it’s a pretty terrible experience for most riders in real life too).
There are many differences - racing frequency being a big one, not just more frequent, but a huge disparity in frequency from one rider to the next.
Again, the principle is of course great, but if it is so simple why can’t someone show me how it can be applied in Zwift, so that I can demonstrate the large number of issues that will arise from it, and why it is probably significantly more complex than you seem to think.
I appreciate you would rather they don’t spend time on a system you don’t like and instead spend time on a system you do, but surely you can see that the end of sandbagging and cruising now is a great step, especially if we wouldn’t get a ranking / results based system for at least another year?
That is within a league. Are you suggesting we could use that exact approach, but on a global scale?
Just checking the suggestion, if that is the case I can demonstrate what sort of race experience that will give you.
Except that those riders only last a couple of races and they will get upgraded.
James, why the sudden anti-ranking sentiment, since you were previously so in favour of it?
1 Like
I think he’s pro-ranking but anti-points.
I do agree with his point that a points-based experience could lead to a wide range in abilities in races. Wider than a good ranking system. I don’t see this as hugely problematic, a group ride and bunch sprint time after time isn’t that exciting anyway. A lot of this debate depends on what you think the point of restricted pens is anyway. Is it to give people a group to ride with, or to give a range of experiences (ie, hammer and nail)? I focus on the latter.
I can see Zwift HQ’s logic here.
Implementing a ranking system immediately would be a free-for-all until enough results made it stable, putting off many riders who won’t wait until then.
With the current enforcement, race results will be more realistic and encourage more (myself included) to start or return to racing. As these results feed through, a calculation of ranking can be made in the background; when ready, Zwift can allow pen enforcement by ranking.
As for enforcing the current ZP categories, I think people are forgetting that they are already an evolution of the original w/kg boundaries. They were already due to be tweaked on the basis of a rider’s power curve, all Zwift HQ has done is take this step along with pen enforcement. Organisers already used the power profile in ZP to identify cruisers, so why not take this opportunity to include it in this latest development?
Next steps would be what I, James, and others have suggested: giving organisers the tools and flexibility to organise and enforce pens by power profile, ranking, weight, age, height, whatever measure takes your fancy
8 Likes
This sounds good but I can see it being problematic. In fact I’m now confident that minority cases will see themselves frozen out of decent racing opportunities. We have already seen this in the test events. When the pen arrangements were harsh on heavyweights, there was an outcry and the system was changed in a day. Now the pen arrangements are harsh on lightweights, and … crickets. “Oh, we’ll just test this and maybe think about doing something later”. Later: “well 90% of people are happy, so we’ll keep things as they are”.
WTRL in particular have a well documented bias against lightweights, which quite likely fed into their decision to ban children from their events. With the core customer base being heavyweight cruisers, it will be only too easy to pander to their wishes.
3 Likes
I don’t know why you think I’m anti ranking. In every statement I’ve made I’ve said I am very pro ranking, especially a good global ranking system as a measurement of progression instead of cats. And most of all, I am pro giving race organisers options. If there was a ranking system, and race organisers could choose to split fields by rankings or by the new power metrics, both would give quite different experiences, and I would want to race both options.
In my hill climb series I could split fields up based on their 2m-6m power. If I could only split the fields based on ranking it would be terrible.
Similarly, if a race organiser put a race on for road to sky, rankings would also make less sense than 20m w/kg.
The system implemented has killed (IMO) 2 of the big issues we see today, sandbaggers and cruisers. That is great. Once this is in, I will rejoin the fight for a proper ranking system and a toolkit for race organisers, but I can still see how good this system is, and how valuable it would still be even once a good ranking system is in place.
3 Likes
Spot on, thank you
1 Like
You (or I, or anyone else) could set up a race series where lightweights are better looked after. If we get it right, I am sure it would do well and be well attended. This is what is great about flexible pens for organisers, the cream will rise to the top. At the moment innovation is constrained by the fixed boundaries. Either you use the Zwift system, or you come up with your own but have no enforcement.
2 Likes
It would be great if it gets to the point where organizers can customize race pen criteria to such an extent. Even including perhaps a race event example of <68kg riders only. Further customize so that the 4 pens could be a subdivision of just 1 or 2 pens.
So, an A/B level race only, split in 4 performance criteria pens (instead of just the 2), and only for lightweight riders. You begin to get to a place where a race event actually could be constructed where almost anyone eligible to enter such a race, could have a chance of winning.