The statistics on how to win cat enforcement events

That was probably an oversight in setting up the Event. ZP use different boundaries than pen enforcement, so I would assume that DQ will be turned off. There is no need for a DQ if the system decide in what pen you start that is the aim of pen enforcement.

3 Likes

@Gerrie_Delport_ODZ , we are not done with bullet no 2.

I realize it is now “technically possible” to avoid the no 2 bullet style DQ’s. And that’s a good thing. What remains is washing away the sort of culture/philosophy where these DQ’s happened/can still happen and for everyone to realize why we can’t have any of that.

Why we can’t have after-the-fact DQ’s:
You must always be allowed and encouraged to do your best. If that means that one day you finally manage to outperform your previous performances, whether in results or in Newtonian measures, then that is always a good thing, something for you to strive for, something for others to celebrate and reward you for. You should never be punished for that (except when caught cheating). You may sometimes get a cat upgrade afterwards but you always keep your result/placing.

That’s the philosophy behind why these DQ’s are bad. The basic ethics of sports. Then there are additional upsides. Notice how banning these DQ’s suddenly makes it so that exploiters (through weight advantage or cruising) are never safe. Even though few can put up a fight against them (since the fight isn’t fair) it is still possible, “legally” possible, for someone else to beat them and take back that gold where it belongs. It will only happen rarely if there are racers present in a position to exploit the system, whether intentionally or not, but at least you are allowed to contest them. Today there are many situations where these racers shouldn’t be allowed to win for a sports ethical standpoint but still do and no one is allowed to do anything about it.

However, that is only one part of bullet no 2. Then there is the other part, the performance ceiling. It is very unnatural and unhealthy to have performance ceilings in sports. But this is the way our cats (or pens or whatever is the politically correct nomenclature at the moment) are divided from B to D. With performance ceilings.

Cat A doesn’t have a ceiling and is thus much healthier. For example, there is no heavy weight advantage in cat A. Weight isn’t the point in cat A, whereas in B-D it is. You also can’t successfully cruise a race in cat A, it’s impossible per definition.

These performance ceilings in B-D can only disappear once we reach the third bullet.

3 Likes

But they don’t applie to pen enforcement races, if Zwift put you in the correct pen then you can go as hard as possible without the fear of a DQ. That is the whole point of Pen enforcement.

then you need to make another bullet, because performance ceiling has nothing to do with DQ.

Won’t the performance ceiling be fixed by bullet #3

2 Likes

There most certainly is. To be in cat A you need high W/Kg, so any heavyweight rider with high enough w/Kg is immediately advantaged because they have high raw power to go along with it.

You made 3 points that need to be critically addressed, 2 of them have been by the recent changes, that sounds like good progress to me.

1 Like

i’ve ran the numbers through a supercomputer and the trick to winning pen enforcement events is to have an 8x body weight ATG backsquat and lungs the size of raisins

2 Likes

It has (had) everything to do with the DQ’s, the way I see it. Zwift racing has been permeated by this ridiculous idea that you are supposed to race within certain predetermined performance bands (with ceilings) and that those who find them selves at the upper end of those bands are “the winners”, as long as they respect the bands. If they don’t respect the bands, you punish them with a DQ.

I also understand that nobody really planned this. It just happened. Things unfolded, bit by bit, and there they were, the W/kg cats. You would have to be a complete idiot to plan racing like this. And now the situation calls for acceptance from everyone. Yes, this was the wrong to do it. Yes, both the platform of Zwift and its subscribers deserve a better fate than to keep it that way. Yes, change is sometimes necessary although scary. Yes, I am willing to stop hugging the past.

You are right of course in a sense. In cat A weight matters in pretty much the same way it does in RL cycling (good, no?). At the same time the avg weight in cat A and the variance of weight there is lower/smaller than in lower categories, but relative to competitors there are scenarios where high Watts pay off for sure.

I would never advocate a change of the in-game physics to make it more… egalitarian. The problem with the performance ceilings is they add a pronounced heavy weight advantage without changing a thing about the game physics. It’s that part that needs to go and what I referred to.

What I meant was that when I ran a similar statistical test of the importance of weight in the past like the one above, I found statistically significant average weight differences between the podium and the losers in cat B-D but NOT in cat A. There was, then, a small difference (1 kg) pointing in the opposite direction (better to be somewhat lighter) but this difference was not statistically significant and could be random.

I guess the reasonable interpretation of that result would be that although we can all picture scenarios where higher Watts (and therefore somewhat higher weight) gives an advantage even in cat A (like a crit on Downtown Dolphin) but that this hypothetical advantage is offset by X number of other possible advantages you could have, ones that don’t involve above average weight. In other words, a healthier and more interesting racing scene.

I think I said this before. But with Pen enforcement you don’t need DQ’s anymore. If Zwift determine that you are a B racer but you perform better than you did ever you get the win but next race you will be starting in A.

2 Likes

Wonderful analysis, but you have answered the wrong question.

The Category Enforcement algorithm is not intended to pick the winners. It is intended to create groups where everyone (ideally) can feel competitive.

The question needs to be re-framed.

What are the differences between the riders at the front and those at the tail of the race? The answer might even help Zwift improve the category definitions.

Defining categories using 15 second power will not result in competitive races. The riders with higher 20 minute power will ride off the front and the winners will have the highest 15 second powers from within that group. Perform your analysis on an open category race and show me the p value for 15 sec W/kg - Podium vs All Losers.

2 Likes

Yes and no. I think the new system (and the old system for those that respected the boundaries) is still heavily favouring certain types of riders. Some riders put in a strong performance (that may not even mean a podium) and get instantly upgraded, others can sit in a non A category and win multiple times and seemingly never get upgraded.

If you have an FTP close to the cat ceiling and a strong sprint you are going to win a lot and never get upgraded. If you have an FTP slightly above the ceiling and a weak sprint you have a very low chance of winning, with a solo breakaway being your only chance, but there is a strong possibility of being caught before the line by the blob and out sprinted and if you go for it you will likely be upgraded win or not.

2 Likes

I understand what you are saying @David_Stark : But that has nothing to do with being DQ’ed after the race like we had in the old system.

1 Like

No, but I’m trying (badly) to make the point that if you start as B, there are riders that can win and keep winning and Zwift still decides they are B, there are others that can put in a strong performance once and lose (not due to bad race craft, just being lighter) and Zwift can decide they are now A.

1 Like

But I was never arguing that point. We are discussing that the new system of pen enforcement does not need to DQ any rider anymore. (except if they were cheating like changing weight Hight or something like that)

2 Likes

Set aside any sort of category system at all for a moment.

The winners of the vast majority of Zwift races are the ones with the biggest combination of efficiency/sprint.

When you start to apply any sort of rationalised field splits (w/kg, rankings, performance, wins, pure watts, whatever) this fact doesn’t change. If you promoted all of the sprinters, you would end up with fields of sprinters with hugely different power profiles and races would be a mess. But still - the riders with this biggest combination would win.

That’s because Zwift racing, with its current courses and physics, is set up to favour this type of racer.

Is this a problem? Not really, in real life certain race types are won by a certain type of rider too. However there are things that could be done to spread the balance a bit:

  • More course variety designed specifically for race dynamics
  • Fix pack speed - goes too fast when riders are not working
  • Fix pack descending - particularly speed around corners
  • Create dynamics to favour or encourage breakaways
8 Likes

Yes, ranks/points categories will also have winners and losers and it won’t even be close between them. But the huge difference is the lack of sensible, healthy promotion in our current system, which you would have given ranks, points or a combo of those. There wouldn’t be any heavy cruisers winning and winning. Win they would at first, surely, and you could even be happy for them and impressed, but then they’d get promoted. And then you have people like me who couldn’t beat these guys, even the ones in the cat below, but I’m still in a cat above and not even a hundred failed races getting spat out the back after a couple of clicks is going to demote me, not without me faking. This fossilized system is unsound and not like a sport.

I just finished a ride while listening to the latest Zwiftcast like two weeks too late, where @xflintx is interviewed. You’ve probably heard it long ago. And I don’t know what to do now. Should I cry? Bang my head against the wall? Shoot myself? I’ll try to be stoic but…

He R E A L L Y D O E S N ’ T understand.

He’s saying we can’t move to ranked racing before we get the power based divisions between categories right (and acknowledges that there’s a long way to go still) because then results based racing will make it even worse.

What the… NooOOOooOOOooo!

That’s the whole point! Performance based categories will always have these exploitable artifacts. That’s why we’re miles off target. And only if we stop hugging this idea of their superiority, exorcise it, have it drawn and quartered, buried 18’ deep with a stake through the heart and then incinerated just for good measure, and then install a results based system, only then will we rid ourselves of these artifacts. They will g o a w a y.

How can it be so hard to understand that the flaws I talk about are caused by performance based categorization, unique to performance based categorization? And what to do when the link between the community and the company development really can’t be helped to understand this?

We’re in for a looooong haul towards fair racing.

Stiff upper lip. Here we go…

4 Likes

I guess it depends on how ranking will work. If an A class racer is allowed to enter races and beat up on lower ranked/lower power racers then how is that a good idea/experience? Shouldn’t you first be confident that the people you are racing against are in the same league as you, then compete to earn points?

3 Likes

A good ranking system can learn and adjust very quickly to rank a rider with some confidence, but the average race experience would be less enjoyable for the average racer if the only option was to split fields using rankings.

It makes more sense to offer both as an option for race organisers, but removal of power based pens is just cutting off your nose to spite your face.

Surely the ideal system is a myriad of different ways to split people up within a race, and a ranking system that promotes progression and gives a rider something to fight for in every race, whether they are the hammer or the nail.

4 Likes

This as you said is the open and shut evidence on the situation as it is unfolding right now. To think that results ranking is a progression or extension to power based divsions speaks volumes to me aswell . Not only are they are not . They are completely different , nearly opposites , in concept , and not linked in any way , there is no dependent features and there is no reason to be thinking you need to “complete” one before starting the other unless as you say there is a misunderstanding on how results ranking could be consumed. I dont think that to be the case , I think its more an indication of the level of commitment we can really expect . There isnt the capacity or resourcing in Zwift to progress this so its excused with some fake reason to not do it.

If its not and we beleive this to be the case (results systems require power categorisation as a pre-requisite) , prepare to be underwhelmed or completely dumbfounded with any results based system that is going to arrive when the current obsession with power based categorisation finally eats its own tail in what could be , based on how long it has taken to get this far , some very distant future date.

3 Likes

With a few million racers how will ranking function with out a good baseline where a rider should start racing. Do you throw every new rider in the top or bottom, it won’t be fun for them to be at the top and unfair to te other if they are at the bottom.

There is new riders every day.

We need a ranking system yes but it will take time.

4 Likes

Like I said already before Christmas, AutoKitten was good enough for that purpose. It sucked as categorization, just as much as the the test model now gone live sucks for the same purpose. But for initial seeding (and that alone) it’s good enough.

Sure, it will misfire for a certain percentage and some will get squeezed (like me) but if so, then there is an end in sight (unlike now). Just race a few races, do your best, and get demoted by simply sucking. Or, if you are undervalued by the stupid power based system, just race a little, have fun winning while you do it, and then you’ll get your promotion to where you belong whether you want it or not. Promoted from doing well, not from pushing X W/kg, mind you.

With half a brain, this is how you would do it.

How about this litmus test? Present both alternatives to the UCI (current system vs results-based with performance based initial seeding) and get a second opinion. Or better yet, propose to them that the next elite Zwift event be divided in W/kg cats and see if they’re still interested in the partnership. Hahahaha!

Zwift aren’t that stupid. They probably know on some level that they’re selling us poop in racing while we finance the clean elite show that in turn is intended to draw more people to the platform. We are the serfs that make it all happen.

Explain to me this:
Why shouldn’t UCI introduce W/kg cats themselves as the divider between World, Pro and Continental? Clearly it must be superior by Zwift’s reckoning.

3 Likes

in a ranking system i am going to glass crank every race until i end up at a level where the flat pace is z1 or z2 and then i’m going to do course specific intervals in those before sitting up at the line to maintain that rank, having completely destroyed a field i have no business whatsoever being in, and a bunch of guys who would normally be protected from sharing a pen with me are going to come on these forums and mald at zwift about their terrible race experience.

why would i do something like that? it doesn’t matter… it might be something as trivial as making sure my activity is recorded on the zwiftpower website because i’m using it as a training diary since my trainingpeaks subscription runs out soon and i don’t want to pay the annual. what matters is that you’re going have to come up with some way of stopping me. personally i would suggest some kind of power based category system

4 Likes