It's looks like they making MMR like in WOW arena. Next step will be teams and rating boost(i hope they won't do team races). But looks better then current system.
I actually felt the group was really balanced (and this is a point at which the rating is probably least accurate as a lot of riders only have an estimation based on their recorded power - which may or may not be accurate). I have really high hopes for this system.
Virtually all competitive games use a variation of the Elo system, they often call it matchmaking rating or skill rating. It is much more complicated and abusable in team based games like League and WoW. For solo games like Chess, Starcraft 2 etc it's pretty straightforward and fool-proof. Whoever wins the race will still largely be a function of where the cutoff in rating falls, and whether the course suits the rider, but it's easy to see the huge advantage over a system where someone can win most of the races they enter for months and not get promoted.
This system is designed to be "more fair" but it needs a LOT of work to cover what the old A-D system provided. It's supposed to take more data points into account as well as how well you funish across multiple events. It's also supposed to take into account skills decay as well.
I was a several times a month Zwift rider and in Cat D. I did the Bologna TT route and it put my cat to cat C. I was always at the back of those races and it put me off entering any races as I just thought this category is way too hard for me
I get the idea, and it's a good step forward, but why are TT events not included 🤷♂ I train to mainly race TTs, and the odd event, so this will really mess up my category I feel 🤔
Where did you read that TT events are not included? I must have missed that. That being said, it is still in beta phase now, so you can join the Zwift Labs and give feedback. That said, I feel there should be some mechanic in place for time trialists.
TT courses should be based on all times on the course, not based on the field that happens to sign up in your pen. It will be better. I show up to a TT series with a few weak riders in the pen, but I can set a great time for the series or course, but because the pen was weak I get no points.
@@robertnobles8189 interesting thought! It kind of makes sense to me but the problem I see with that is the distribution of the riders amongst the categories will shift over time. For example, top of the B cat is somewhere around 600 rating now. I assume that when the system is released and a lot of people move to such system, the top of B cat will move to more like 700ish. In which case the result from today's TT wouldn't really be comparable to last months' etc.
@@RoadtoA TT is the race of truth. Your time is your time. It doesn’t matter what point system it’s under. The course is the same so is the movement of time. It stays constant.
@@robertnobles8189 what I meant is if you base your rating change off the all-time best, you are comparing it to rating that is no longer relevant. Example - you set a course record at a TT course with rating 700. After a year of using the system the distribution of riders in the categories shifted and now most people have higher rating. You have a rating of 900 now. Another rider beats your course record. What rating should be used for the calculation of their rating change? 700, 900, or other? To clarify, I like your idea, it makes sense, but I can't imagine it working technically.
You mentioned that you can also lose points, which afaik is not true. According to the info button on ZP, only your 5 best races in the past 90 days will count towards your rating. Hence I don't think it will have an effect on racing mentality (e.g. fighting for every place even when dropped as you mentioned). You can pretty much only gain points by racing actively and it should be fairly simple to rank up if you're allowed to join higher ranked races since it should be enough to beat better ranked dropped riders who stopped taking the race seriously. I am all for a new system, but without consequences for poor results, this system is heavily flawed. I don't even know how they came up with that weird idea. They could just copy any ELO system in the gaming industry.
You 100% can lose points on poor performance. Without it the system would be meaningless. You cannot lose points beneath your initial placement though. Quote from Zwift: "A higher finish against stronger competitors in races will increase your score more significantly. Conversely, lower placements or less strategic engagement in races will result in smaller gains or even reductions in your score. This method ensures that each racer’s score truly reflects their performance and growth in capability over time.”
@@Cid0484 They explicitly added the racing score and the info button yesterday or so, therefore it's not a matter of "up to date". The old race ranking is also still there. I did not read any FAQs. I just assumed that newly added information is correct.
There will be an ELO decay but I didn't find any details how it will work. As for the smurfing...you would still need to pay the subscription right? I can't see that being worth it for people.
Personally I think it's a step backwards and doing it on fitness/performance would be a better system if they improved on it. Rather than results based. Eg If I enter and place high in a few tiny races get bumped right up I'm never gonna want to try a long climb race as the disparity is gonna be massive between someone who is only racing and winning those types of races. Whereas before at least I knew people were somewhat similar fitness even though I'd be on the weaker end in some types of races.
I can see the argument but I think in a way you can use it on the fitness based system too. You can get bumped into high category by your endurance but then you'll die in crit races and vice versa. I think that's inevitable, you will always be better at some style of riding. The problem with a fitness based system is you can really cheat it when you understand how it works. Noone can force you to test your max capacity in any shape or form which means that if you desire you can grief lower categories very easily (which happens now). Result based system will let you do that once or twice and then you're gone.
@@RoadtoA Yeah they can not do a maximal effort, but that also means they cannot do power over what you would expect in the cat. So even if they are sandbagging in theory they still are only doing what someone in that cat could do.(when working correctly) They could fix that if they looked at the course then your power they could put you a cat up in races that suit you type thing based on recorded fitness. (which they can do with results based as well) flat, hilly and mountain rankings.
What's crazy is the guy in 3rd place has a 5 minute power of 4.23 w/kg and averaged 3.8 w/kg for the race and has a much lower seed score than me who has nowhere even close to those numbers.
It will take some time for the score to settle down. The seed score is based off max efforts at 30 seconds and 10 minutes. If they do not have their max efforts recorded at these durations, they are seeded in much lower bracket than they should be. But that should solve itself fairly quickly.
@@RoadtoA Agreed, the seed score is problematic. Hopefully it works itself out and works as well as zwiftracingapp which has been results-based for a while. In my race (another 450-550), the zwift racingapp score ranged all the way from 1526 down to 942 - I think you can guess who the winner was ;)
I mean but what's the point? If you do not effort in the end, your sandbagging was irrelevant. If you do, you get promoted and you're done in this group.
@@E.Dulwich_Zwifter Basically, even if Tadej Pogacar joins 250-350 bracket and rides in a way so he doesn't win, it does not matter that he is there because he is not hurting the group in any way. :)
Well, i did 4 mini races today (I'm a low C) and got dropped on all of them pretty quickly...still racing was entertaining but not much different for me...
@@RoadtoA 420, which meant I was put in a 400-575 group. So to be fair me getting dropped is probably right as I'm at the lower end of the category. Also it looks like my racing score has dropped to 412 in the last race but my profile hasn't been updated (so this is also good). This is probably going to work out fine after a more races...but maybe that 400-575 group bucket should be split up a bit more.
@@neeekneek oh yes, that makes sense. You are probably not fighting for the podium in such a pen, but on the other hand, you have a potential to gain a lot of points if you do manage to show some result. On the other hand , if you are racing at the top of the pen (which is my case now) you need to perform or you will lose a lot.
the algorithm they are using is simply not fit. it is weighted far too much for 30second efforts, and very little weighted towards race performance. i'm a mid cat c racer, my zrs is in the 470's, not because of my race performance, i can just about hold the group for 1km, then i'm done for, but i have done some of the work outs that have 30 second efforts. if it was fit for purpose the algorithm would identify i'm not a 470 racer. it is gross incompetence.
@@percyveer2355 I can see your frustration but perhaps the problem is opposite - not that you don't belong in 470 group but that your opponents don't. Don't forget that it has been only beta with a low amount of races to really shift the numbers.
@@RoadtoA hi, interesting way to look at it. however, the fact that a 30 second effort, just one, has such an impact on zrs is simply wrong. they have hundreds of my previous races, they have my power profile, and yet, one, 30 second effort is valued so highly. the weighting is simply wrong, in races as i said, i get destroyed, yet the score barely drops. i will be doing some hard max efforts today, i expect my zrs to increase further, it has almost become a game in itself, how high can i get this score. the fact it is beta is not good enough.
@@percyveer2355 I agree and I actually gave them the same feedback during the testing phase. I think that the emphasis during seeding should be on 1-5 min power which is the area that, imo, has the biggest impact on Zwift races. As far as I know, they are looking at ways to implement more power/duration points into the seeding. That said, when the system is used by all the racers, the initial seeding won't matter that much - the riders will quickly shift to a rating level where they belong.
@@RoadtoA i think that is the issue. in races i have been in, many of those in far too high a category just mentioned about not bothering. also, in the races i took part in, the numbers seem lower. this can lead to almost self selecting data that would support the initial, bad algorithm. imo there needs to be greater impact of race performance, this should be higher weighting than a 30sec effort, the power and heart rates should also be measured through out the entire race, this will identify those that are sand bagging/cheating. zwift has everyones power profile, it should not be too hard to identify anomalies. i just don think they want to.
@@brulsmurf Just the initial score though. I know that currently you cannot drop below that number but I am fairly sure most people will go up in their rating when the system releases (now you have low A-high B at around ~600) so there's about 400 points left for things to setyle down. Also, from the discussions on Zwift Labs I know the team is looking into the methodology of the initial placement.
@@RoadtoA Also the assumption that a bicycle race is a zero-sum game, and thus is suitable for an elo-based ranking system isn't correct. Sprinters are depended on their teammates for bringing them to the finish line. The sprinter wins, but only because of his teams effort.
@@brulsmurf Fair point. Still though it's IMO better than a system based on numbers that you can EASILY cheat. This might not be perfect but at least will keep sandbaggers away.
It will take some time before the distribution of riders settles. Now most people have a rating somewhat close to their initial one. After some racing, the better riders will shift to the top brackets and vice versa. :)
A is full of cheaters, other categories aren't competitive at all, and exist only to boost the ego of beginners, making them think they're racers. So more (delusional) satisfied costumers. Zwift being fair would be zwift whitout cheaters. I love how non competitive turists play dumb and act like As would be supra humans starting with an unfair davantage. Or like a place in a race shows your merits, whatever the level of the field. Find enough grandmas and race against if you want to feel as a winner. Don't need to train, just find weak people enough 😂
@@dupondavignon869 Are you okay, dude? Do you wanna talk? :-) Cheaters are problems for sure, but one problem at a time. In some way, I can't see cheating being removed from Zwift completely, because a lot of it depends on user's input. When there is a way to cheat, people will cheat I'm afraid.
its easy to game this anyway. or as a guy in this guys elo is using a trainer(cough) not really a trainer but will connect to zwift, and delivers really bad power. EZ wins for him at 2.1 in this cat.
@@RoadtoA I'm ok. But what's the real point of wanting to be vs "fair" opponents? What does this "fair" word mean? You didn't answer. Are they supra-humans? Are they guys racing with a road bike vs bmx? When you see a guy outdoor, do you compare to him, but if you see he's way too strong, you don't say to yourself "I'm too weak", but, "it's unfair, your FTP is like 4.2w/kg. It would be fair if you were 3.9, and as such, in my category"? (I suppose you're B, I'll don't check, to be honest, I don't conceive being lower than A if following a decent training for a couple years ; just riding z2 being enough.) That's funny how guys like to "win a race" in lower category, and think being, like, a mid A for example has no value compared to a "B winner", even if the A is way stronger than the B 🤭 Ego. Ego. Ego. Make people being (playing) dumbs. And THEY have the same problem than the cheaters : their ego, which make them flee reality :)
Thx for the shout out! :D You are so right. I am really looking forward to this system in action in full force later this year!
Thought this was a change tailored exactly for you, having seen your videos lately. :)
@@RoadtoA Yes! Hate it when the A pens are empty. This change really looks great!
OPTIMAL!!
It's looks like they making MMR like in WOW arena. Next step will be teams and rating boost(i hope they won't do team races). But looks better then current system.
I actually felt the group was really balanced (and this is a point at which the rating is probably least accurate as a lot of riders only have an estimation based on their recorded power - which may or may not be accurate).
I have really high hopes for this system.
Virtually all competitive games use a variation of the Elo system, they often call it matchmaking rating or skill rating. It is much more complicated and abusable in team based games like League and WoW. For solo games like Chess, Starcraft 2 etc it's pretty straightforward and fool-proof. Whoever wins the race will still largely be a function of where the cutoff in rating falls, and whether the course suits the rider, but it's easy to see the huge advantage over a system where someone can win most of the races they enter for months and not get promoted.
@@AndrewisTri-tn1uo 100% agree.
its not
I have not done any Zwift racing in a while but. Could not tell the difference between a B race and a C race.
Prime time C racing is, in my experience, very close to off-prime B races. But this change should change that.
This system is designed to be "more fair" but it needs a LOT of work to cover what the old A-D system provided. It's supposed to take more data points into account as well as how well you funish across multiple events. It's also supposed to take into account skills decay as well.
@@jamesmckenzie3532 the decay is included. But I agree it will take some time for things to settle down.
@@RoadtoA That is true. That's why it's being tested. Some adjustments are small but some may require a lot of work.
I was a several times a month Zwift rider and in Cat D. I did the Bologna TT route and it put my cat to cat C. I was always at the back of those races and it put me off entering any races as I just thought this category is way too hard for me
This should really improve your experience then!
I get the idea, and it's a good step forward, but why are TT events not included 🤷♂ I train to mainly race TTs, and the odd event, so this will really mess up my category I feel 🤔
Where did you read that TT events are not included? I must have missed that. That being said, it is still in beta phase now, so you can join the Zwift Labs and give feedback.
That said, I feel there should be some mechanic in place for time trialists.
TT courses should be based on all times on the course, not based on the field that happens to sign up in your pen. It will be better. I show up to a TT series with a few weak riders in the pen, but I can set a great time for the series or course, but because the pen was weak I get no points.
@@robertnobles8189 interesting thought! It kind of makes sense to me but the problem I see with that is the distribution of the riders amongst the categories will shift over time. For example, top of the B cat is somewhere around 600 rating now. I assume that when the system is released and a lot of people move to such system, the top of B cat will move to more like 700ish. In which case the result from today's TT wouldn't really be comparable to last months' etc.
@@RoadtoA TT is the race of truth. Your time is your time. It doesn’t matter what point system it’s under. The course is the same so is the movement of time. It stays constant.
@@robertnobles8189 what I meant is if you base your rating change off the all-time best, you are comparing it to rating that is no longer relevant.
Example - you set a course record at a TT course with rating 700. After a year of using the system the distribution of riders in the categories shifted and now most people have higher rating. You have a rating of 900 now. Another rider beats your course record. What rating should be used for the calculation of their rating change? 700, 900, or other?
To clarify, I like your idea, it makes sense, but I can't imagine it working technically.
You mentioned that you can also lose points, which afaik is not true. According to the info button on ZP, only your 5 best races in the past 90 days will count towards your rating. Hence I don't think it will have an effect on racing mentality (e.g. fighting for every place even when dropped as you mentioned).
You can pretty much only gain points by racing actively and it should be fairly simple to rank up if you're allowed to join higher ranked races since it should be enough to beat better ranked dropped riders who stopped taking the race seriously.
I am all for a new system, but without consequences for poor results, this system is heavily flawed. I don't even know how they came up with that weird idea. They could just copy any ELO system in the gaming industry.
I don't think ZP is up to date yet. The FAQ on the Zwift page regarding the new system clearly states that you CAN lose points.
You 100% can lose points on poor performance. Without it the system would be meaningless. You cannot lose points beneath your initial placement though.
Quote from Zwift:
"A higher finish against stronger competitors in races will increase your score more significantly. Conversely, lower placements or less strategic engagement in races will result in smaller gains or even reductions in your score. This method ensures that each racer’s score truly reflects their performance and growth in capability over time.”
Oh, I can see the confusion. This system isn't based on the ZP system. It is based on the score of Zwiftracing.app
@@Cid0484 ZP uses different system. This one is based on Zwiftracing.app and you are correct you can lose points.
@@Cid0484 They explicitly added the racing score and the info button yesterday or so, therefore it's not a matter of "up to date". The old race ranking is also still there. I did not read any FAQs. I just assumed that newly added information is correct.
Will the ELO decay when not active? Do you think smurfing will be a problem? (i.e. people starting a new account in a month or two)
There will be an ELO decay but I didn't find any details how it will work. As for the smurfing...you would still need to pay the subscription right? I can't see that being worth it for people.
I can now confirm myself that there is a decay. In two days I dropped from 551 to 550.
oh come on. smurfing isn't a thing.
Personally I think it's a step backwards and doing it on fitness/performance would be a better system if they improved on it. Rather than results based. Eg If I enter and place high in a few tiny races get bumped right up I'm never gonna want to try a long climb race as the disparity is gonna be massive between someone who is only racing and winning those types of races. Whereas before at least I knew people were somewhat similar fitness even though I'd be on the weaker end in some types of races.
I can see the argument but I think in a way you can use it on the fitness based system too. You can get bumped into high category by your endurance but then you'll die in crit races and vice versa. I think that's inevitable, you will always be better at some style of riding.
The problem with a fitness based system is you can really cheat it when you understand how it works. Noone can force you to test your max capacity in any shape or form which means that if you desire you can grief lower categories very easily (which happens now). Result based system will let you do that once or twice and then you're gone.
@@RoadtoA Yeah they can not do a maximal effort, but that also means they cannot do power over what you would expect in the cat. So even if they are sandbagging in theory they still are only doing what someone in that cat could do.(when working correctly) They could fix that if they looked at the course then your power they could put you a cat up in races that suit you type thing based on recorded fitness. (which they can do with results based as well) flat, hilly and mountain rankings.
What's crazy is the guy in 3rd place has a 5 minute power of 4.23 w/kg and averaged 3.8 w/kg for the race and has a much lower seed score than me who has nowhere even close to those numbers.
It will take some time for the score to settle down. The seed score is based off max efforts at 30 seconds and 10 minutes. If they do not have their max efforts recorded at these durations, they are seeded in much lower bracket than they should be. But that should solve itself fairly quickly.
@@RoadtoA Agreed, the seed score is problematic. Hopefully it works itself out and works as well as zwiftracingapp which has been results-based for a while. In my race (another 450-550), the zwift racingapp score ranged all the way from 1526 down to 942 - I think you can guess who the winner was ;)
Surely you could still sandbag by not going flat out?
I mean but what's the point? If you do not effort in the end, your sandbagging was irrelevant. If you do, you get promoted and you're done in this group.
@@RoadtoA I get it's pointless unless you want an easy race...which isn't much fun :)
@@E.Dulwich_Zwifter Basically, even if Tadej Pogacar joins 250-350 bracket and rides in a way so he doesn't win, it does not matter that he is there because he is not hurting the group in any way. :)
Well, i did 4 mini races today (I'm a low C) and got dropped on all of them pretty quickly...still racing was entertaining but not much different for me...
And what is your rating and what was the group you were riding in, out of interest?
@@RoadtoA 420, which meant I was put in a 400-575 group. So to be fair me getting dropped is probably right as I'm at the lower end of the category. Also it looks like my racing score has dropped to 412 in the last race but my profile hasn't been updated (so this is also good).
This is probably going to work out fine after a more races...but maybe that 400-575 group bucket should be split up a bit more.
@@neeekneek oh yes, that makes sense. You are probably not fighting for the podium in such a pen, but on the other hand, you have a potential to gain a lot of points if you do manage to show some result. On the other hand , if you are racing at the top of the pen (which is my case now) you need to perform or you will lose a lot.
the algorithm they are using is simply not fit. it is weighted far too much for 30second efforts, and very little weighted towards race performance. i'm a mid cat c racer, my zrs is in the 470's, not because of my race performance, i can just about hold the group for 1km, then i'm done for, but i have done some of the work outs that have 30 second efforts. if it was fit for purpose the algorithm would identify i'm not a 470 racer. it is gross incompetence.
@@percyveer2355 I can see your frustration but perhaps the problem is opposite - not that you don't belong in 470 group but that your opponents don't.
Don't forget that it has been only beta with a low amount of races to really shift the numbers.
@@RoadtoA hi, interesting way to look at it. however, the fact that a 30 second effort, just one, has such an impact on zrs is simply wrong. they have hundreds of my previous races, they have my power profile, and yet, one, 30 second effort is valued so highly. the weighting is simply wrong, in races as i said, i get destroyed, yet the score barely drops. i will be doing some hard max efforts today, i expect my zrs to increase further, it has almost become a game in itself, how high can i get this score. the fact it is beta is not good enough.
@@percyveer2355 I agree and I actually gave them the same feedback during the testing phase. I think that the emphasis during seeding should be on 1-5 min power which is the area that, imo, has the biggest impact on Zwift races.
As far as I know, they are looking at ways to implement more power/duration points into the seeding.
That said, when the system is used by all the racers, the initial seeding won't matter that much - the riders will quickly shift to a rating level where they belong.
@@RoadtoA i think that is the issue. in races i have been in, many of those in far too high a category just mentioned about not bothering. also, in the races i took part in, the numbers seem lower. this can lead to almost self selecting data that would support the initial, bad algorithm. imo there needs to be greater impact of race performance, this should be higher weighting than a 30sec effort, the power and heart rates should also be measured through out the entire race, this will identify those that are sand bagging/cheating. zwift has everyones power profile, it should not be too hard to identify anomalies. i just don think they want to.
The new system is still based on your power-numbers. It gives a bottom to your racing score. sprinters are punished in this system.
@@brulsmurf Just the initial score though. I know that currently you cannot drop below that number but I am fairly sure most people will go up in their rating when the system releases (now you have low A-high B at around ~600) so there's about 400 points left for things to setyle down.
Also, from the discussions on Zwift Labs I know the team is looking into the methodology of the initial placement.
@@RoadtoA Also the assumption that a bicycle race is a zero-sum game, and thus is suitable for an elo-based ranking system isn't correct. Sprinters are depended on their teammates for bringing them to the finish line. The sprinter wins, but only because of his teams effort.
@@brulsmurf Fair point. Still though it's IMO better than a system based on numbers that you can EASILY cheat.
This might not be perfect but at least will keep sandbaggers away.
@@RoadtoAIf you know beforehand in which pen you will be, you can cheat. The pens should be finalized 1 minute before the race. Like Indievelo does.
@@brulsmurf That is one I definitely agree with. Yes.
absolutely NOT. this needs way more work, and in any case, you can game it so EZ
How do you game it?
how tall are you?
@@Zhrine1337 i'm short, 170 cms.
@@RoadtoA i see. Then you are pretty strong for your height. How much do you weigh?
@@Zhrine1337 I currently weigh 73 kg (down from 82 in January). But I still should lose more to be on an optimal cycling weight.
@@RoadtoAimpressive weightloss!
@@Zhrine1337 thank you!
Sounds like the rating system in chess
Yup! But it's not open ended, 1000 is the max.
Yeah…it’s great….
Now literally no one is in A grade 😂
It will take some time before the distribution of riders settles. Now most people have a rating somewhat close to their initial one. After some racing, the better riders will shift to the top brackets and vice versa. :)
Why zwift in the summer
Still fun in summer, riding in the evening, when it's raining, many other reasons. :)
But I ride outside a lot as well.
Maaan, it was 12 celsius and pouring rain yesterday, give us a break lol
It's not summer everywhere.
@@Cableguy72 like here. I can be 12 c or it can be 27 c. You never know. 🤷
They should concentrate on getting rid of the makita dewault coffee dopers cheats instead of more crap thanks
What do you mean? You can't use coffee break during the race. Or did I misunderstand your point?
@@RoadtoAI just meant cheats in general the 7ft tall 10kg riders it terrible nowadays really puts me off zwift even more
@@kennymcallister2523 I can see that.
A is full of cheaters, other categories aren't competitive at all, and exist only to boost the ego of beginners, making them think they're racers. So more (delusional) satisfied costumers. Zwift being fair would be zwift whitout cheaters.
I love how non competitive turists play dumb and act like As would be supra humans starting with an unfair davantage.
Or like a place in a race shows your merits, whatever the level of the field. Find enough grandmas and race against if you want to feel as a winner. Don't need to train, just find weak people enough 😂
@@dupondavignon869 Are you okay, dude? Do you wanna talk? :-)
Cheaters are problems for sure, but one problem at a time. In some way, I can't see cheating being removed from Zwift completely, because a lot of it depends on user's input. When there is a way to cheat, people will cheat I'm afraid.
its easy to game this anyway. or as a guy in this guys elo is using a trainer(cough) not really a trainer but will connect to zwift, and delivers really bad power. EZ wins for him at 2.1 in this cat.
@@RoadtoA I'm ok. But what's the real point of wanting to be vs "fair" opponents? What does this "fair" word mean? You didn't answer. Are they supra-humans? Are they guys racing with a road bike vs bmx? When you see a guy outdoor, do you compare to him, but if you see he's way too strong, you don't say to yourself "I'm too weak", but, "it's unfair, your FTP is like 4.2w/kg. It would be fair if you were 3.9, and as such, in my category"? (I suppose you're B, I'll don't check, to be honest, I don't conceive being lower than A if following a decent training for a couple years ; just riding z2 being enough.)
That's funny how guys like to "win a race" in lower category, and think being, like, a mid A for example has no value compared to a "B winner", even if the A is way stronger than the B 🤭
Ego. Ego. Ego. Make people being (playing) dumbs. And THEY have the same problem than the cheaters : their ego, which make them flee reality :)