Disturbing tendency comparing cat A race data to cat B-D

It is no secret that I am no fan of the W/kg category system and that I firmly believe it has to be replaced by a results-based system, like in real-life sports. But let’s leave opinions aside for a bit and look at raw data instead. I did.

I started comparing the level of effort in races between top contenders in various ways and have so far always come to the same conclusions, regardless of angle. Most recent was a comparison between the winner and the two other riders on the podium.

Level of effort can be defined as where in your HR distribution graph you spend the bulk of your time in a race. A race spent mostly in upper Zone 4 is harder work than one spent in lower Zone 4 or even Zone 3.

What we expect to see here in a healthy category system is either of these two:

  1. The winner works the hardest. Among the top contenders the fitness is about the same, so on average the one willing to do the hardest work will win.

  2. Differences in effort are random or negligible. Those at the top of a category are equally fit and work equally hard, more or less. While effort matters, there are other (random) factors affecting outcome, such as splits, powerups etc.

What we don’t expect to see is for the winner to work less hard than the others, at least not in cat B-D.

In cat A both A and A+ riders compete and obviously an A+ has an advantage. Since A and particularly A+ riders are relatively scarce, it is not uncommon with a podium with an A+ on top followed by two A’s. And it is conceivable that the A+ rider worked somewhat less hard than the others, as you only have to go hard enough to win even if you have some over-capacity as an A+. So maybe any tendency for winners to work the hardest is mitigated by this fact. Perhaps winners in cat A even work slightly less hard than the rest of the podium on average.

But what would we think if we found that winners in cat C worked less hard than the others? It would be hard to explain and would go against our intuition and our ideas of sports ideals. It may also suggest that there is some cruising going on in cat C.

A sandbagger is someone signing up to a lower category race, someone who doesn’t care about W/kg limits and who will go flying off the front at race start. We have all met them. They are easily spotted.

A cruiser is more devious and is a rider who would belong to a higher category, according to the W/kg system, but who stays behind in a lower category and makes sure not to exceed the W/kg limit. With a decimal or two to spare on the running W/kg average, easily monitored with your smart-trainer app, he can drop you in climbs, sprints etc by putting you at VO2Max, all while he himself is merely… cruising. A creature of my imagination? Look for yourself.

As opposed to the sandbagger, the cruiser will not be detected by any anti-sandbagging measures and will not be DQ’d or upgraded by ZP if he is careful. The obvious way to spot a cruiser is only after the race, by studying HR distribution graphs. Someone that gets the podium often while still showing a lower effort in the HR graphs than other contenders can be defined as a cruiser, whether he cruises consciously or not. At any rate, he has a clear advantage over a rider who can barely reach the category W/kg ceiling.

Anyway, I went through 100 races in cat A and cat C respectively and compared the two with regards to how the level of effort differed between the winner to the other two on the podium. I then ran a statistical test to find if there was indeed a significant difference, this in order to exclude the possibility of any such difference just being chance, a not too improbable chance.

Well, there is always the chance, You can in theory flip a coin a 100 times and have it come up heads every time, but if the probability for that turns out to be very low (it is), then you would rather look for other explanations than dismiss it as just a random event. Maybe have a closer look at that coin you used.

It turns out that effort in races is not at all what we might expect (see above regarding expectations). Even though A+ riders might have an advantage, winners in cat A, where there is no upper performance ceiling that you must avoid, actually work harder than winners in cat C. Or to be more exact, it is far less likely that a cat A winner will have worked less hard than the rest of the podium than it is for a cat C winner.

This difference between cat A and cat C is statistically significant, i.e. you can’t easily explain it away or argue yourself out of it. Rather, the results suggest that the unnatural performance ceiling in cat B-D does strange things to race results, or even that there is a statistically visible element of cruising to cat C.

I assume that the results can be replicated and that you would get the same results in a comparison between A and B or A and D (when I looked at those from different angles it did indeed seem so).

If you want more details on method then you’ll have to find my blog. Google for zwiftcruiser and blogspot and you will find it.

I would do more tests, but they are awfully time consuming since I don’t have access to raw data and had to copy numbers off ZP and Zwift by hand.

7 Likes

Thanks for sharing this Andreas. Can I challenge you on a few of your points:

  1. You seem to make the assumption that someone who chooses to sit in the draft and rides tactically is a less worthy winner than someone who rides flat out. I disagree. Sport, and cycling in particular has always been about tactics. You aren’t suggesting are you that Mark Cavendish sitting in the draft for 99% of a Tour stage and then being led out over the last 1k “goes against our intuition and sport ideals”? Because on your definition he appears to be a “cruiser” doesnt he?

  2. I think you might be guilty of conflating your “cruisers” with sandbaggers. A sandbagger deliberately rides in races below their category. IN my race tonight there were 17 B riders signed up to a C race. ZP deals with them - they are DQ-ed (I know there are other important arguments about spoiling races etc, but let’s not repeat that here).

  3. As I read your argument, you seem to be suggesting that some people keep deliberately below the threshold to avoid being DQ-ed on the ZP system. Two problems with that argument as I see it:

(a) Riders are either categorised as, say a C rider on ZP, or they’re not (see 2 above)

(b) If you are suggesting people who could ride harder and be recategorised as a B are “cruising”, where is your evidence for that? How can we know without asking them, whether their intention was to stay below the limits? How can we know they are not moving up on purpose, to stay at the front of their category?

Many of us may be able to think of one or two examples for whom that appears to be the case, but it’s a hunch; an unsupported assertion.

  1. While “Level of effort can be defined as where in your HR distribution graph you spend the bulk of your time in a race”, it’s not the only metric is it. I might spend most of my time in upper Z4/Z5 in a Crit but in Z2/3 on a long mountain race over several hours. Intervals are often harder than paced rides but bulk of time is in lower HR zones.

Anyway, just my two penneth. Thanks for posting this - rarely any discussion on here these days - it’s more of a tech help forum!

3 Likes

I don’t recall, but does Zwift store our max HR or does it just assign a generic HR zone chart to us based on age? My max HR will be different than others in my age group, so unless this is specific to each individual the HR zone graph is not a great basis of the argument. Not saying that you are wrong, but the data could be off.

I agree we need a different categorization method.

1 Like

This is one of the arts of winning Zwift races, it can be done various ways, by reverse weight doping, where a rider add weight to stay with cat limits, and then there is the obvious control your w/kg during the race and make sure not to go over.

This is easy to spot look for riders that have top 3 places for the last 20 races, then look at their racing performance vs training.

w/kg categories is just weight classes like boxing. Cycling should use points system like IRL.

3 Likes

You enter it yourself in settings I think, Mike

Weight doping is a different issue.

Andreas is suggesting deliberate coasting by riders who should move up a category. I havent seen any evidence of that, largely because it’s based on a rider’s intention, which you cannot establish without asking them, can you?

What do you mean by “look at their racing performance vs training”?

Ultimately, I agree points over W/KG all day long.

1 Like

O you just have to follow Zwift Riders on FB and you will soon realize that it is a art to stay in the correct Cat. I can even mention names of those the do it openly. ( no I won’t). There is no rule stopping you form doing it.

You will see they Race cat C but then you look at there training rides you will see they go over the 95% of best 20min limit regularly. Some produce B numbers in training or free rides but ZP races they are top C.

2 Likes

Well that’s depressing Gerrie. Must be my naivety.

2 Likes

And that is why we have been asking Zwift to for racing ranking based on results.

4 Likes

I wondered about this last week looking at a D race, where a rider well out ahead averaged 113 bpm for the event.

1 Like

If ranking should be based on results, then how many top results should you have to level up?
I just started racing in D, and in some races, there are just 5 to 10 participants, and if I get a 2. or 3. place, should I then level up?
If I get placed in C, then I would be of the back before I am out of the starting pens, as I can not produce the W/kg the C-racers can.

Zwift is (probably) unique in MMOs in that it doesn’t have a performance based ranking system. Like real life there a number of options for how you rank, and how you deal with new riders, but all of them are better than the specific power based categories.

When I raced on Zwift (ages ago) then I discovered there was another problem with the categories. I’m (to put it politely) a heavy guy, and my FTP hung about 2.4W/kg but 300W. This mean I was stuck in category D. On hilly races I was vaguely competitive but on the flat I would usually crush it because I could easily keep up with guys that were 2.4W/kg but 200W. Cruising, but in this case legitimate cruising, was what I did because the object is the “with the minimum effort.” The only thing I had to really watch out for was working out if the people ahead were sandbaggers.

Tactics (such as drafting in the pack) don’t work well as the draft advantage is pretty small and every race is a time trial. Even when people latched onto a C group at the start of races then, provided the race lasted 30 mins or more, I would probably catch them anyway (or, more commonly, they were just fliers)

Didn’t get long to get bored with this and stop racing.

I’m happy to have a discussion in here too, And I agree, there is too little discussion and also, if I may, critical thinking in this and other Zwift forums. Often people strike me as overly conservative and complacent. You get used to something and you stop questioning and reflecting.

Someone had a great idea on the fence yesterday (don’t make the fence drop people, just increase the air resistance in front of it instead) but it was somewhat shot down by others saying it would be better if the fence as is was fixed instead. I don’t like that approach. Just because I’m used to something being this or that way doesn’t mean it’s the best way. If something isn’t working well, then maybe it can’t realistically be fixed or it just isn’t worth the hassle it would mean. Maybe take a fundamentally different approach instead. With the fence. Or the W/kg categories.

I’ll reply to your points:

  1. You seem to make the assumption that someone who chooses to sit in the draft and rides tactically is a less worthy winner than someone who rides flat out. I disagree. Sport, and cycling in particular has always been about tactics. You aren’t suggesting are you that Mark Cavendish sitting in the draft for 99% of a Tour stage and then being led out over the last 1k “goes against our intuition and sport ideals”? Because on your definition he appears to be a “cruiser” doesnt he?

Hmm… how long should I make the counter-argument? I’ll do the full monty.

Regarding “cruising” in RL events, I was thinking the same when watching Tour of Flanders yesterday. Well, Cavendish didn’t do too well, but there were other sprinters who were carried all the way to the finish. And no, their HR profiles will not look the same as the chain gang’s or the breakaways’. But two things are missing here.

First, IRL racing is not like Zwift racing, something I discuss at some point in the blog regarding VTdF. Zwift races are very short compared to outdoor continental events. They should tend to lead to increasingly hard efforts. Because you can. And because it will pay off (in cat A). And I do believe we have seen just that in Zwift events with RL pros lately. In Tour for All I got the impression many riders brought the RL pace into Zwift and didn’t quite “get it”. In VTdF, the racing was mostly much harder. They finally got it. There was little saving energy mid-stage. And you don’t have to save energy mid-stage. You can go flat out from start to finish. I’d even like to argue that given equal fitness, someone willing to go flat out is at an advantage. Not so much so, though, if you go flat out up the leg snapper in Innsbruckring in cat B-D, and someone is glued to your wheel, still at tempo pace… Doesn’t that happen IRL too? Yes, it does. But nobody will get a DQ for going too hard there, that’s the difference.

Secondly, as for Zwift and my little study specifically, my assumption is that I know nothing of the 100 x 2 races. Couldn’t, like you say, someone with a lower HR just have been smarter and drafted more efficiently? Someone getting caught in splits time and again and who has to chase back will neither be riding very efficiently or be saving energy. But since we know nothing of each of the particular races, couldn’t also the opposite be true? That one guy in high Zone 4 is actually very efficient at drafting, but he still has to work harder than the next guy to hang in there? My point is that any such differences in efficient tactics will likely even out on average between riders at different effort levels over a large number of races. So a rider showing high effort isn’t necessarily a good and efficient rider, but neither can we assume that he is bad. All we can say is he does work harder. And all we can say is that working hard does seem to pay off in cat A but not in cat C, and then we have to find an explanation. And the explanation can’t be drafting, because, if any, the cat A riders, riding in a category where effort pays off better, would surely know how to draft efficiently compared to the C riders.

  1. I think you might be guilty of conflating your “cruisers” with sandbaggers. A sandbagger deliberately rides in races below their category. IN my race tonight there were 17 B riders signed up to a C race. ZP deals with them - they are DQ-ed (I know there are other important arguments about spoiling races etc, but let’s not repeat that here).

A sandbagger is a different beast. A much more stupid or careless cheater than a cruiser. Are there cruisers? Yes, there are. If you start looking at HR graphs in races, you will find them. And I’m pretty convinced some of them know full well what they are up to. To prove this point and draw attention to it, I have started to cruise myself. I hate it, it’s dirty work, but someone’s gotta do it. As I haven’t had a lot of time with Zwift this summer, being “stuck” outdoors, I haven’t perfected it yet. But I aim to downgrade legitimately to cat D during the upcoming fall and start cruising seriously and give the full inside view on the blog. Just to show it can be done, just like the cruisers I come across time and again.

Then there is a huge grey area where people’s efforts differ but it likely is not a question of deliberate cruising. More a question of preferences or random events in races. It just so happens that people show different levels of effort when racing and that the race results reflect that, only not in the way you would think. It comes with the system.

And really, it’s pointless going after individual subscribers, like ZP does with the sandbaggers, even when discussing deliberate cruisers. I’d like to argue that it is the W/kg system that creates these evil miscreants, sandbaggers included. Like some famous sociologist once said (I remember this vaguely from Uni), something along the lines of “Every society gets the kind of criminals it deserves”. So we should go after the system instead, not the criminals here. With a results-based categorization, both sandbaggers and deliberate (and even unknowing) cruisers will cease to exist by definition. They will no longer be an issue. It is really that simple.

  1. As I read your argument, you seem to be suggesting that some people keep deliberately below the threshold to avoid being DQ-ed on the ZP system. Two problems with that argument as I see it:

(a) Riders are either categorised as, say a C rider on ZP, or they’re not (see 2 above)

It’s easy to downgrade. It just takes time. I will do this next. For others, with some precautions you can avoid getting the “rightful” (according to ZP) upgrade by making sure you don’t go over limits, even though you could if you raced harder. So shouldn’t you be allowed to choose your level of efforts in races just like in workouts? Of course! Not everyone enjoys a near-death experience. I just don’t think you should be consistently rewarded with wins for going less hard than others, not all the time. But are you? Yes, there is this counter-intuitive tendency, like I have just shown in data. It can’t be denied, it’s there, in our faces, once we actually start looking.

What separates the lower categories is not how well you have done in the past, like in RL sports. It’s not your perceived effort either. It’s a raw power number and a raw weight number. Which is completely absurd, when you think about it. If you’re in cat B and you really want to win a race, then what your category says is “Go hard! But not too hard! If you don’t go hard you will lose. But if you go too hard, I will DQ you!”

You absolutely cannot put a perfomance ceiling in a sport. Imagine the qualifiers in the Olympics. To make it to the finals in the high jump you need to go over certain heights in the trial heats. Now imagine you could get a DQ if you went too high in the qualifiers. It’s absurd. Imagine some division 2 football team getting DQ’d for scoring too many goals during a game (“Sorry, you were too good, you don’t belong in this division!”). You can’t govern categorization during a game or a race. You have to do that between races. And you should do it based on results, not on some physics metrics.

(b) If you are suggesting people who could ride harder and be recategorised as a B are “cruising”, where is your evidence for that? How can we know without asking them, whether their intention was to stay below the limits? How can we know they are not moving up on purpose, to stay at the front of their category?

I could show you cases of obvious cruising. And they are not just of the sandbagger type, guys with initials and a number for in-game name, or named Spacefrog or whatever. There are even rather high profile riders cruising. And then they brag about their results. I just don’t want to point fingers. Or you can just watch me. I will be the spokesperson for the cruisers. And I will start with 3 underperformance races in C in late August. Then we just wait for the cooldown and recategorization.

But it doesn’t really matter whether someone holds back on purpose or not. We should go after the system, not the individuals. Put a people into poverty and starvation and you will see a lot of new petty thieves, it goes without saying. Forbid locks and you will see a surge in burglary. Put a performance ceiling in a sport and you will have sandbaggers and cruisers. Introduce a results-based categorization and we no longer have sandbaggers and cruisers. We will still have dopers and hackers, but not the cheaters unique to Zwift. And we will never have to ask ourselves “did this person do this or that on purpose or not”. The question becomes irrelevant.

  1. While “Level of effort can be defined as where in your HR distribution graph you spend the bulk of your time in a race”, it’s not the only metric is it. I might spend most of my time in upper Z4/Z5 in a Crit but in Z2/3 on a long mountain race over several hours. Intervals are often harder than paced rides but bulk of time is in lower HR zones.

Yes, obviously this must be so. You can’t go high Z4 for several hours unless your name is Sepp Kuss. But it’s not level of effort per se that is of interest. It is relative level of effort. A long distance race will be mainly tempo pace, but if the level of effort differs between riders in that race, and if they are all on the upper category W/kg limit, then isn’t that peculiar?

Still, it is a very crude method, agreed. I do think HR is the single best simple measure available, but I don’t even have the raw data to analyze in better ways. So yes, the analysis is very crude. But I hereby formally challenge the guys at ZP and Zwift to do better analysis and come up with a different conclusion. I’m 100% sure they won’t be able to. It’s there, in the data, what W/kg cats do to racing in the lower categories.

And there is a very simple way out of it, one that Eric Min mentioned himself in the Zwift Podcast of 4 Feb. I just want to see it happen. Until then you will find me nagging about it, and you will find me cruising just to show it can be done.

1 Like

I’ve never understood why someone who enters too low a category and then flies off the front is called a sandbagger.

That’s not the dictionary definition of sandbagging, which is deliberately masking your true performance level. Like literally carrying sandbags to slow you down.

The “cruiser” being referred to here is the sandbagger.

2 Likes

By “win” the race, are we talking about the one who crosses the line first, or the one who is left over in first place once all the over-category racers have been DQed? And by “over-category,” are we talking about over-category by any amount (3.3 instead of 3.2… or 4.1 instead of 4.0… etc)?

lol Andreas I guess I mean to ask, When you get (back) to your first hand cruising experimentation (attempting to win or at least podium regularly below category), what criteria will you follow? Are you going to attempt to win/podium overall (while remaining “in category”), or rather only taking into consideration those who aren’t DQed? I suppose it would be a tall task to do the former.

By “win” the race, are we talking about the one who crosses the line first, or the one who is left over in first place once all the over-category racers have been DQed? And by “over-category,” are we talking about over-category by any amount (3.3 instead of 3.2… or 4.1 instead of 4.0… etc)?

We are talking about the glorious winners according to ZP, so an unknown number of racers will have been DQ’d already, yes.

There is too little data easily available (without coding some web scraping tool) to look into avg W/kg differences, so I skipped that. They’re all in one bunch, regardless of how much they could overshoot the ceiling if they wanted to, if that’s what you meant. That perspective, all of them grouped, is both a weakness and a strength, I think, when studying the data.

And I agree with what @Steve_Hammatt says above, that ‘sandbagger’ is a bad term. The only actual sandbagging going on with them is entering a lower category. Everything else they do has very little to do with sandbagging. It’s just that when people talk about sandbaggers, they tend to refer to the flyers, because they are the highly visible cheaters. And all the anti-sandbagging measures target the frequent flyers and only them. So I wanted a term to capture the others, the devious ones that hide their true strength throughout the race (the real sandbaggers) and opted for ‘cruisers’. Someone else in here (I don’t remember who sadly so I can’t give credit right now) coined the term earlier in a sarcastic remark in some thread and I liked it.

lol Andreas I guess I mean to ask, When you get (back) to your first hand cruising experimentation (attempting to win or at least podium regularly below category), what criteria will you follow? Are you going to attempt to win/podium overall (while remaining “in category”), or rather only taking into consideration those who aren’t DQed? I suppose it would be a tall task to do the former.

I will of course try to consistently get as many high placings - podiums or even wins preferably - on ZP. You can’t go for Zwift wins because then ZP will DQ you since you will be riding with the frequent flyers. You have to pick groups carefully after start and be ready to drop if you notice you’re starting to go over limits mid-race. You need a little wiggle room in your W/kg average because it’s that wiggle room that is going to get you to the podium.

At the same time, there is a challenge to cruising. Often you are not competing against people racing fair. You are instead competing against other cruisers in the front group, but you still have to watch your effort. It’s not necessarily super easy to win as a cruiser. It’s just so much easier than a flat out effort.

Also, it’s more fun to win on ZP than on Zwift anway, because of all those stupid threads here or on FB where some poor newbie first discovers the massive cheating in Zwift, asks a concerned question about it, not really understanding what is going on, and the entire bandwagon goes: "Duuuh, dude, you need to look on ZP! ZP is the OFFICIAL results! It’s da TRUUUUF!’ Yeah right…

There’s a full race report on my blog describing my very first attempt as a cruiser, an unexpected Zwift win BTW (I thought it would be harder…). It was a Beta Crit City race that would have been a ZP win too if I had been declassed already - I only overdid it a little, so it would have been fine with ZP. At the time I had the idea that it would be more challenging not to monitor your average W/kg (“Feel the Force, Luke!”), but it’s too hard, too easy to get carried away, so I won’t do that in the future.

The Betas are extra fun since all the anti-sandbagging measures are in place. People were literally getting green coned left and right, but as a low power, low bandwidth rider I couldn’t get green coned if I tried, so I’m completely safe. You do slip under the radar as a cruiser. All the efforts of Zwift and ZP combined to stop the cheating are completely in vain with regards to cruisers. There are absolutely zero preventive measures against cruising, but to try come up with some would be a bad idea (don’t try please). Just as anti-sandbagging measures was a bad idea. You’re just trying to fix something that was built flawed, like plastering cracks on a house built on loose sand, and will just make things worse one way or the other.

With regards to your first question again, I did that race not long after a month of corona infection and my fitness was in terrible shape. So it turned into mainly a lowish Zone 4, which by my standards, the way I have been forced(?) to race in the past, I still perceived as overly easy. But I do see cruisers who consistently race from Zone 3, with only brief excursions into Zone 4 and 5 in order to drop people (and they do…). It must be so boring!

Note: I can’t believe I’m having this conversation… I HATE cheaters and it stings my heart every time I see one of those newbie threads. But there you go, the inside view, all for free. That’s how you do it.

If ranking should be based on results, then how many top results should you have to level up?
I just started racing in D, and in some races, there are just 5 to 10 participants, and if I get a 2. or 3. place, should I then level up?
If I get placed in C, then I would be of the back before I am out of the starting pens, as I can not produce the W/kg the C-racers can.

@Ole_Jensen, you have some fair questions about how exactly it would work with the “better alternative” to W/kg categories (according to me and many others) i.e. results-based categories. And we can’t just criticize what we don’t like, we also have to present at least some details on an alternative, since Zwift seems to be in no rush to do so themselves, whether they are working on it secretly or not.

There are ways to handle all of the things you address in different results-based systems. What would work the best for Zwift has to be tweaked out, but even starting fresh I see no problem with starting up a results-based system in parallel with the old system during a transition period. Actually, the old categorizations could serve as initial seeding, or we’d have to pass all riders through the lowest category at start, which would be somewhat chaotic.

Official races, ones that adhere to certain standards, would affect your ranking/category. Then organizers could also choose to host inofficial but nevertheless fun (or even quirky) races that don’t affect ranking but use the riders’ current categories so that you still can’t enter a cat you don’t belong to.

There are many approaches to systems, but let’s mention 2 of the most well-known ones. First, sounds by the name like you’re a Scandinavian like me, and over here we’re used to the age categories, which are almost as dumb as the W/kg system, at least if you want to engage as many as possible in the fun of racing. But over the pond the US Cycling Federation has a better system.

If you’d rather read for yourself, you can download the rule book at the USCF website and on this page the US racing categories are detailed:

Basically, in road racing they have 5 categories. One “Novice” (formerly cat 5) which you will graduate from after having participated in a set number of races. This is the Beginner’s Races you see sometimes in Zwift and that many ask for. Not too serious but you get a taste of racing.

Then, you are moved to cat 4 (and then 3 and so on). By participating in a minimum number of races in a season and by scoring points in those according to a points table in the link above, once you reach a certain points threshold you get an upgrade to a higher cat. You can either apply voluntarily for a move, with a lower points threshold, or you get a mandatory upgrade if surpassing a somewhat higher points threshold.

How do you get the points? The winner will score the most points in a race, the runner-up a little less and down through the results table. A big event with many participants has a deeper points scheme than a small event, and thus you get around the problem of podiums becoming too heavily weighted points-wise in events with low turnout. If you show “bad” results over a season in your category, you can either apply for or get a mandatory downgrade.

Some variation of the USCF categories would surely work for Zwift too with some modifications. But there is also another appealing system, the FIS system of the International Skiing Federation. It’s not really a category system but produces ranking scores that is used to rank skiers in the World Cup and determine an overall winner.

You can read about it here:

Each time you participate in a race you get a rank adjustment according to this formula:

P = F * (Tx / To) - F

P: Points
F: A factor that varies with race type. A sprint race might be 1400 and a longer one 1800 etc
Tx: The finish time of the rider/skier, in seconds
To: The finish time of the winner in the same race

So let’s say there’s a 29 km race in Zwift and that this type of distance has an F-factor of 2000. And let’s say your finish time is 51:22 (3082 sec) and the winner’s is 47:35 (2855 sec). It doesn’t matter whether you get on the podium or not, it’s your time in relation to the winner that matters, so in this way the system balances events with both few and a lot of participants. So your points contribution of this race would be:

P = 2000 * (3082 / 2855) - 2000 = 159

And if you were the winner in this race:

P = 2000 * (2855 / 2855) - 2000 = 0

So a low point is better. And then FIS takes your current 5 best races in the season (which could be a rolling season, like the standard ZP last 90 days, and it could be best 3 races instead) and your ranking is the average of those 5.

Since the FIS system is not a category system exactly (just one category), you’d have to define a cutoff or other trigger mechanism for when and where you are eligible for a category upgrade (or downgrade). Note that this doen’t have to be a limit carved in stone but could be adjusted as you go to get good sizes of all the categories. And either you need to keep a separate ranking for each category or you can use a universal one but modify the formula slightly, maybe like this (I have NOT thought this through properly):

P = 2^C * (F * (Tx / To) - F), where C is your category with 1 being the highest category. Then you’d get ranking score bumps between categories for easier separation, but you’d still need cutoffs. Once you run your first race in a higher category, you would be starting from a high (=bad) ranking but the kind of points you get from then on will be quite lower on average.

Or maybe you should not only take into account your own vs the winner’s finish times but also your respective rankings? This is common in online computer games. If you beat a high ranked rider, then that is worth more than beating someone equal or inferior to you.

You’d have to address the need for some kind of seasonality and points degradation for lack of activity without disincentivizing people from racing at a level and frequency they feel comfortable with, and to not punish people who prefer outdoors during the warmer months or who get injured or busy in life too harshly. There are ways.

Or Zwift could do what every multiplayer online game does - reward high activity and make Zwift junkies of us all: “Sorry hun, could we postpone our 10 year anniversary a day or two? See, there’s this important race tonight and my ranking would drop if…” Rather not, right?

There are many possible approaches to designing a results-based category system. But they all have the merit of insta-vaporizing certain forms of cheating and they don’t impose the suffocating performance ceilings on the riders in the lower categories - you just go as hard as you can (or are willing to) and you try to race as well as you can, and you get rewarded accordingly. And you end up in the category where you belong, you cannot hide.

Also, if Zwift will continue to add strategic and tactical elements to the racing like not only the good old powerups but also steering and whatever the future brings, then a results-based ranking better captures the overall prowess of a rider. It’s the results that matter, not the Watts pushed, and getting results is then not just about raw and unrefined fitness but other skills too, and a results-based ranking can capture that. The W/kg system can’t.

3 Likes

When it is the rating,I have not been able to follow a 1st group in the hill,so I’m fine with Category B?
(My FTP became 236W(4.3PWR) from 221W(4.0PWR) recently)

You mean you got a 4.3 in the race (and not when doing the test)?

Well, if you keep going that hard, doing 4.3 or similar, then as soon as the average of your best 3 races the last 90 days goes over 4.0, you will get kicked out of B. And a 4.3 is over limits so that race will get a DQ. (You don’t have to sign up for A races until ZwiftPower says you are A though, it takes more than one race.)

And then people will tell you: “You got dropped in the hills? Then you need to work more on your VO2Max!” But it won’t help you if you race against riders who are strong enough to race in A but sign up for B races and stay below 4.0. They will keep dropping you in hills anyway. That’s how you win as a cruiser. That IS the tactics.

2 Likes

This is the way I think Zwift should go. If each rider has a ranking score we could have battle royal type races. Zwift can auto group us together and we can race. The there can be more than 4 race cats, you select a race got to the start pen and with 30sec to go you get assigned a group. If 1000 people enter then there will be 10 groups of 100 and if 10 people enter than one group, organizers of stage races or racing leagues will set point ranges as groups so that you always race in the same group.

8 Likes