Zwift probably don’t know what are the problems with ZRS, so lets do a townhall meeting.
Problem: Users don’t understand how the ZRS algorithm works or what factors affect their score.
Why it matters: Without clarity, racers can’t trust the system or optimize their racing experience.
Why do you gain less score beating a bigger field?
Problem: It’s been months since the last update or public communication. Why it matters: A lack of iteration and feedback loop slows down trust and momentum. If it is abandon just abandon it and go back to CE.
Problem: ZRS isn’t integrated with race organization, sign-ups, or results in a meaningful way. You can’t even track score change history.
Problem: ZRS seems to favor certain race formats, and doesn’t always adapt well to varied styles like TTTs or endurance events.
Why it matters: Racers who enjoy different event types are left with a skewed score. Who would have thought Tiny Race vs 100km Epic races use different powers?
Full confession, I like what they’ve done with ZRS. I hated it at first, but the changes that have been made since, though months ago, were an improvement. As a strong C/lower B rider with the CE system (which I think we should absolutely not go back to), I was competing regularly for podiums, but it kept me from being a stronger cyclist in general. With the mix of riders from categories above, I’ve only gotten better as a rider, both inside and outside, in the time since the change.
That is not to say that it shouldn’t be improved upon.
Points 1 and 4 above are probably the biggest flaws. With what seems essentially as a top 1/2 - 2/3 of the field getting a score increase and the rest a decrease, without much taken into account of the scores of your competitors (as the vELO system at zracing.app does), it’s hard to get a sense of the logic of exactly how things are affected from race to race. And yes, the reasoning behind larger fields getting smaller score increases is also a mystery.
The most meaningful ZRS related number is probably the delta between your seed score and your ZRS. This is a reflection of how often you have finished well.
It is however not as meaningful as your ZRapp score, which is better at reflecting the same thing by taking into account strength of field + terrain… and not giving random outcomes based on size of field.
I have a seed score of 617 and can win Advanced races. Seems untenable at scale to start people like me in C then graduate me up very very slowly. I ran a test on a second account and won 5 straight C races before being upgraded to B. I just rode off the front and kept a pace. (My zrapp velo is around 1950 but I don’t really check it)
So either… see that I’ve done this once and whack me into Advanced OR implement a much bigger podium bonus.
If you race an event that uses ZRS for entry but doesn’t score ZRS then it displays your seed instead - maybe a bug (odd that this exists as an option).
To see your 100% seed score, or estimate that of another rider from data you can see…
At 261W for 5mins and 96Kg, my 100% seed score is 209.
Which having not looked at this garbage system for a while, because I’ve not raced in about two months and have been riding outside when not full of lurgy, makes it odd at first glance why my Racing Score is 170.
85% of 209 is ~177.
So many riders are close to 15% below their seed score on 90 day data and it’s allowing a number of them to race in a lower pen.
100% seed score being the floor still has to be a priority bandaid for this diabolical race system.
That velo system app is really great. It has a lot of cool metrics to look at and is well organized. Thanks for sharing the site, I hadn’t used it before.
This calculator is only an estimate. I ran it on my own numbers twice when I was at my ZRS floor, and both times the calculator over-estimated my ZRS seed by about 10 points. I believe this is because the calculator doesn’t consider rider height and the actual Zwift algorithm gives us tall riders a break (which is appreciated, since the Zwift physics engine has a grudge against us).
The only really big problem [sic!] with ZRS is that it isn’t a relative score (relative to other racers). It’s what kills ZRS.
The score calculation must take other participants’ starting score, result and score change into account as well. Yes, it’s going to put a heavier load on the servers running relative calculations, but it needs to be done. Any other problems with ZRS are just minor tweak issues by comparison.
It would have worked if we had distinct racing seasons and a definite number of entrants each season. But we don’t. Racing in Zwift is non-stop, drop-in.
When ZRS was first introduced, I was suddenly able to keep pace with the peloton in the vast majority of the races. I wasn’t winning but I wasn’t dropped at the first 2 minute burst. V2 changed the power values for about a week and then V3 went to 10 minute only for power. IMHO, it makes no sense to ignore the contribution of short power, 1-2 minutes, in race performance. I’ve been complaining about this for a while to no avail and a few of the posters here on the forum say it doesn’t matter.
Turns out one of the loudest voices against including short game is @Peter_Twigt who won the Tiny Races in Z2 last week and the May 3rd event because he has significant short power:
In any Zwift race with one of the short climbs, who do you suppose will prevail? In one of the very few races with a long climb I MIGHT have a chance but we’re pretty close in 20 minute w/kg and I’m probably 15 years older.
I imagine most of the other voices against including short power have similar profiles. I’ve heard all of the arguments including it was arrived at scientifically or it doesn’t make any difference but for many of us, it does.
I can say for myself, racing is less fun with V3 because the categories don’t feel level. It’s like before CE when A and B riders would infiltrate the C races and blow things up from the gun.
Your Racing Score is 469, so you have tended to finish more often in the top half against your opposition (not that it tells us how strong your opposition was, as this isn’t a vELO system).
Peter’s Racing Score is 514, so he has tended to finish more often in the bottom half of his race pens.
If Racing Score was limiting the score floor to 100% of seed score, over the past almost 3 months until sometime in the next couple of days, Peter would have been in a higher Tiny Races pen than you (B vs C). But his 90-day 5min best from 4th March is about to expire.
I’ve not looked at all the races you have both participated in, too see if either of you have been selective in the course profiles of races you have started and finished, or whether the courses were in the lap of the gods through for example Eric’s selections for Tiny Races and whether they were a nice even spread.
There’s probably other stuff that could be commented on with the comparison with Peter, but I’m too ill/fatigued to spot it.
This system is trash, end of, as far as I’m concerned.
For some miraculess reason and still unknown to me I dropped about 50 points in my ZRS a few weeks ago. Was building my way up and suddenly dropped 50 points to around 421. I can look it up when it was but I did. I started the Tiny Races this weekend at 453, so had to be in the cat I raced in (390-510). And yes, I did win 3 of the 4 races. But as someone with a good sprint, the downhill finishes could not have suited me better. In the one race that had more climbs (The Richmond one) I finished 7th.
So yes, I did win, but with such finishes I have a good shot at a good result. Combine that with some time spend looking at the routes, where they finished and acted accordignly, surprising a lot of riders who clearly had no idea where the finish was.
But I can only do that when the finish is flat or downhill as in this case. If you look at the tinies from May 17th you see I hardly finish in the top 30 since they were all climbing routes.
So yes, sometimes you win and sometimes you wont. It is not like the CE system anymore where I just would win almost everything.
and my racing score went from 453 to 514 to am in the next cat now, so that is kinda working also. Of course the first race that doesnt suit me I will probably drop back, but just racing on routes that suit you is boring so I also do races on terrain that is hard for me, and I will never win.
And you’ve managed to validate the case for short game to be included in ZRS. I’m at 469 and we would’ve been racing the Tinies last week, and the week before and the… At any rate, why would I be penned with racers with nearly twice the short power but very similar FTP power? Makes no sense.
The scoring for wins, places and getting your ass handed to you make no sense either. Win a TT with one other guy who’s 60 points lower in ZRS and your score jumps 20 points. Get your ass handed to you in a race with 25 people and you lose 9 points. Come on Zwift, we can do better than this.
Fix this part and I think including the short term power isn’t necessary.
The biggest problem with including sprint and long-term power is that many people haven’t recorded anything near a maximum effort over those intervals, but you have to assume that whatever has been recorded is the maximum value they can achieve. Trying to balance it makes for an overly complex formula.
You can see issues from both ends - anyone with a good sprint was given a V1 seed score that made them unable to compete on anything but the flattest courses. You had A+ TT riders with scores significantly lower than C grade sprinters.
And the other extreme of CE where people stagnated at the top of a category because there was no need to push hard enough to breach the limit.
The 5min interval is a good enough balance between requiring some endurance and being likely that people have gone close to their best over the interval. The seed score calculation from it makes it fair over a wide range of weights, though it would be better if it included height too.
If the scoring worked it’d take care of the rest, nudging people higher if they are able to perform better than their seed score suggests.
True and agree. I think they should include all datanumbers and not just one. Wether that is 30s, 5m, 10m or 20m We need a combination of them all. I do think the 5m we have now is the best of these we have, but we kinda need for them all to count.
Yes, but if we are completly honest. Without a good sprint it will be very hard to do well in the tiny races. They are way too short for endurance to make a difference and to drop the sprinters.
And we just do not have the numbers for everyone to be in a race where there are only people in it with the same power profile.
