Anti sandbagging and other areas that need development and communication

The suggested promotion/relegation system is only relevant if categories are going to be based on performance metrics. In that case, there has to be a mechanism that corrects the limitations in the system that don’t take consistent winning and losing into consideration.

I don’t think it should be used if categories are split by race ranking. The race ranking system has this mechanism built-in, and categories can be split at some more or less arbitrary ranking numbers. There has to be some rules preventing people on the borderline from constantly jumping between categories though.

@OleKristian You could theoretically use performance metrics to detect repeated intentional downgrades in a ranked, results-based system. The only time performance metrics should be used other than when seeding newcomers into racing, if you ask me. It’s hard to actually prove such ranking exploits though. Maybe they had an infection or some other valid reason for suddenly performing worse. But if there is a repeated pattern of below-power-curve-performance as soon as a certain guy gets upgraded, then sure… (As a comparison, I have seen examples of cruisers going on a 3 month cooldown from monitored activities, more than once, as soon as they accidentally get upgraded - there will always be some people who will stop at nothing to be able to keep exploiting).

However, I think it’s better to just let them go. This is as far as I am willing to go when it comes to the subscriber freedom Min (I can only guess it’s him, his wish) so stubbornly defends at any cost (us). It stops there in my book.

So if someone wants to yoyo up and down, let him. He can’t keep winning forever in the lower cat in a results-based system (unlike today’s exploiters) and while underperforming in the next cat he is not really hurting anyone.

Or at least try that approach and keep a backup plan. If it then turns out these yoyo’s become an all too common problem, that basically every race has yoyo’s in them in the same way every race today has sandbaggers and almost every race has cruisers, then there is a clear need for some kind of enforcement, then you need to act somehow.

But like Ian suggests, please do visit the other thread. Don’t let your ideas get buried in this mass grave.

1 Like

Whoa! Only five paragraphs and four of them are short! The Apocalypse is upon us!

3 Likes

It’s all about posting frequency. Now that I start to slowly approach the frequency of the rest of you, the number of paragraphs dwindle. The more often you post, the less you have in store to say each time.

2 Likes

If you have a good ranking system, then really the only reason to split people into different categories really is to keep the overall size of each start to a reasonable size.

If you have a race that will attract hundreds of people, really just want to split it roughly 35-30-25-20 to get four reasonable starts. You can incentivize (or disincentivize) people to race in their “proper” cat by making it more likely that you will be successful (because you won’t get dropped at the gun) thereby likely improving your rank and still allow people to race up a cat if they want to race with their buddies etc.

If anyone wishes to see a race ranking results based system introduced and would like to contribute to how it might work please post here Race ranking discussion

1 Like

Regardless of how Zwift judges our race efforts, whether that is by pure performance metrics or by finishing position, allowing racers who aren’t registered on Zwiftpower to compete in races and allowing racers to quit before the race end with no consequences makes a complete mockery of the racing. Even moreso in league events held over a number of races.

5 Likes

Have a missed a post?
I thought even those who are not signed upto Zwiftpower will be impacted by cat enforcement?

If people don’t want to be on zwiftpower that’s upto them (I’m of the opinion zwift should auto enrol everyone) but I expect them to have to race in the correct category.

On the non ZP users not showing in zwift power, I’m not sure I agree with the reasons given even for that… Dont want to publish their data without consent, all their data is published in the companion app and anyone can view it there? It’s the same people viewing it via the web interface so what’s the difference. The companion app is not locked down to active users only.

1 Like

You are 100% correct.

The dilemma is that if people make activities/profile private in Zwift/CA they cant make them private in ZP.

Cant block paying customers from racing but Zwift have no choice re privacy rights. Gets challenging if Zwift were going to use ZP for displaying new CP/MP and Ranking.

They can show in the results, just blank out their data. You can opt out showing some data now in ZP such as HR.

That user will show in the companion app race results so it’s not completely private even when applying privacy settings.

Good points. Think its only HR the user can select to opt-out of currently but yes, they can remove all user data easy enough. Have to wonder why if its so easy, they havent got it on the roadmap yet, or even done it already.

Yeap, really hope that people can´t change weight and height with the ease that we see today.

I wonder how Zwift races would go down if we had category enforecement via weight groups, instead of performance data.

Would four (or maybe split further to five) categories something along the lines of…
-60Kg
60-75Kg
75-90Kg
90Kg+

typically give a decent field entry split?

Or how about you simply sign up for a race and ~1min or so before start signups stop and the entries are split by weight into four equally sized groups as possible?

Weight doping would be off the charts, i could easily change my weight to under 59 kgs and destroy the field.

1 Like

Possibly not across the whole field but what about category pens ( provided they were large enough) being split by weight?

Surely even Zwifts computer programmes could spot that an ‘out of game’ or ‘not active in an event’ change of weight of 38% might warrant further investigation.:balance_scale::microscope::lying_face:

No, weight is an important variable but it’s not a good category. Very good races are had across these weight boundaries in every single race of the day.

If you split by weight the ability variation in the field would be huge and blobs very small, more like a TT.

As an example in the Critty Critty Bang Bang Race A race on Saturday, less than 3 seconds covered people from all 4 of your categories (from 60.0 to 91.0). The 91kg guy would smash his weight peers if he was made to race them.

2 Likes

I don’t agree with the second sentence, not if we are discussing anything but cat A. And I can prove it with data.

I do, however, agree very much with the first sentence. So let’s make an IMPORTANT POINT about the the idea of weight as categorization, a clarification of why weight would be such bad basis for category definitions:

Weight is a variable that will definitely impact race results, THEREFORE it should NOT be used as categorization.

The same goes for ANY OTHER variable that has an impact on race results. They should NOT be used as categorization.

While it may seem intuitive to use a variable that has an impact on race results for categorization it is a fallacy. It’s when you try to predict future results based on some external metric and use that to put people into pens that you go wrong and all sorts of bad stuff happens.

20 min W/kg is a horrible basis for categories, not because it is W/kg but because it is intended to act like a proxy for future results. Using such measures to defines categories will only lead to unfair races and exploits (not necessarily related to each other, i.e. they can be unfair even without exploits).

There are several reasons why it is such a bad idea but one of them is that they will not be perfect predictors and this makes them fertile ground for exploits and unfairness. And even if they did perfectly predict future results, then it would still be a bad idea because racing would then be pointless.

Thus neither weight nor height nor Watts nor 1 or 5 or 20 min W/kg or any other such measure should be used for categorization. Ever.

A categorization should NEVER try to predict future results. It should be built around a completely different principle. You should never be in a category because someone predicted it will match your future results (I can make an exception for first-time racers). It should always be because you earned that categorization.

And btw there is really only one variable suited for categorization, one that has nothing to do with physics. You all know what I’m referring to.

1 Like

In Zwift, heavy weights with a W/Kg near the top of the cat limit win most stages regardless of profile.

Very light weight riders below FTP threshold stand a chance on very hilly course as they can get a higher W/Kg limit.

Those in the middle will never win. e.g. There are no top 40 Cat C racers between 70-75Kg.

The big guys get to match them in W/Kg on the climbs and destroy them with absolute watts on the flats.

Any middle weight Cat C rider that put out the watts to match the heavies would be bumped to Cat B as soon as they tried.

8 Likes