Indeed. Training for races is now one of the less smart things to do on Zwift.
2 Likes
This is not consistent with experience. It was promised that the full equation would be released when ZRS was finished. It seems we’re at that point. What’s the equation?
2 Likes
This is the fundamental flaw. If you are a sprinter, your 30s and 10m max efforts are mutually exclusive. The drop off of your 30s power during a 10m effort is enormous, so combining these to seed you has no basis in reality.
Versus a non-sprinter who may have little drop-off, and therefore they are being seeded on numbers which have a basis in reality.
Image shows drop-off whilst doing a 9m maximal tiny race. No sprint to be seen. That is the reality.
3 Likes
How you get your score so high?
This whole system is flawed. How can we put everyone under 1 score?
We want to rank sprinters and climbers under the same system and surprised when one gets demolished in a race that favors the other?
Cat C pure sprinters won’t make it up to the Country Sprint with cat A ftp riders. They would be maxed out holding 4.5 w/kg for 5 mins, meanwhile the cat A riders are pushing 6 w/kg. And you have no draft to work with.
And that is an easy climb. Try doing Everything is Bagel and you will see 10 mins time difference,
1 Like
I still question why only 2 data point CP30 and CP600 are used? Why not 20 data points or 500 different data points?
This “reset” does not make sense to me.
My score went from 369 to 435 after the reset.
I have not cycled in 2 weeks and also have never had a score in the 400s+ which indicates that this is not just a “reset” but you have also changed the seed score calculations itself - otherwise 435 would have been my “seed score” anyway and I should never have been able to get below it?
So tell me; what has actually changed besides a reset - you just have changed the calculations as well?
The notion that most people complain about their score being too high, shows a fundamental problem.
Collectively if everyone don’t like their score being high they are going to figure ways to lower it.
I’ve had a mix of results from Racing Score events since trials began in July, vast majority being Tiny Races.
I had one good weekend in July, when the quartet was pretty flat, I got three second places from very close lead group sprints besides the winner of all four in pen E. The E score limit was also relatively low at 225.
Since then, pen E limit has been ~300 in Tiny Races, perhaps even 330 some weekends. I’ve sometimes had one good result in a weekend, along with a number of bad results because pen E is often full of CE Cs with better aerobic ability than my zmap.
Overnight, my score has reset to ~266 from being ~247 yesterday iirc.
My memory isn’t great, but I don’t recall seeing my score that high during the trial period.
I have a sprint and am heavy.
Curious to see the finishing times in the World Series Event Race that is now happening.
In the cat I would be in the 20 minute values differ from 4.5 w/kg (The A’s) to 1.8 w/kg (The D). And that in a 47km race.
Who came up with that idea that it was a good plan to put them into the same race should really be ashamed he even presented the idea.
All these years there were complaints that the differences in Cats were to big. And what does Zwift do … make them even bigger … 
Make it mass-start with everyone in the event going against each other. Works particularly well for longer races like this so you’re less likely to end up riding solo.
I’m thinking I might be the odd one out, but I think mine is probably a bit too low, but after doing a race I’m not sure it’s ever going to get to where it probably should be.
So I just did the World Race Series Stage 2 … which is not flat, so there was a lot of separation.
I started very bottom of the cat, at 358. I finished with folks all in the 400-500 range and my score only went up +4 to 362. At this rate there’s no chance I’ll get to the 400-500 range this season even if I finish around folks in this range consistently. On the other hand ZRA gave me a “Major zrCS rating adjustment from bottom of one category to middle of the next category for the same position.
This one might be wonky because it’s also a
points race with a sprint and KOM points maybe?
Here’s the list of folks and their ZRS who I finished around. My ZRS went up 4 points.
Edit: not that it matters much in my case as I’m not strong enough to be in the next cat anyhow, I am probably middle of the pack in this cat.
1 Like
That’s probably around what I would expect if it’s a score based on race results. I don’t know how many entered the race but I don’t think it would make sense to get a big jump in points for a 25th place finish, though beating some guys with higher scores still gave you a few points.
I guess what I would be more interested in is how some of the riders behind you with 500+ scores still got an increase, finishing in those positions whilst being beaten by plenty of riders with lower scores.
Both are interesting to me. In this case there were 64 riders who finished, only 5 of those 64 have a ZRS lower than mine after this race, and they were all behind me along with 35 others behind me who all have higher ZRS (in many cases much higher).
So… The average ZRS for those behind me is well above my ZRS, and the ones around me all increased ZRS even though they are in the high 400s or low 500s, again way above my ZRS.
Everybody in the top 38 gained ZRS (even if they finished behind a lot of folks with lower rankings), and almost everybody 39 and below dropped ZRS.
I will not be competitive in this category, but I’m probably middle of the pack.
Edit: Also note the person with the lowest ZRS who beat my time was 408. My score went up +4 from 358 to 362, at this rate I would need to do 12 races to get to 408, and that won’t happen this season probably since I tend to do a race every one or two weeks - and in this category I’ll probably mostly finish middle of the pack.
Then again, if everyone’s score is too high, then everyone is net-neutral on their placements.
Do people look at ZRS score changes on a number basis, or a percentage basis. If I look at my score changes since this started, seeded at a score of 2.
Then…
+4900%
-32%
+100%
and now with the reset
+18%
Some drastic swings, but I’m still an admittedly pretty weak racer.
I cant wait till October 7th when suddenly all community races will use ZRS and everyone who hasnt been in these test events notices the gaps in cats are even greater than before and as a D you have to race C’s and B’s now and for all C’s its a new ballgame against the top B and even A’s.
Guess we will some increased complaining here 
1 Like
Hey all, I wanted to provide a quick summary addressing the main themes in this thread.
First off, thank you for the feedback! Your passion and this community are what make working at Zwift so enjoyable. And yes, we race too!
We know there’s been a lot of focus on the seed scores, but we want to emphasize that this is just the starting point. As you race more, the scores will naturally level out. That said, we hear your feedback and appreciate it!
Seeding algorithm and use of sprint vs. longer power durations
We understand the concerns with using 30s power. The 30s power interval wasn’t chosen randomly. It was identified through a comprehensive data analysis that found the optimal power durations for predicting the winners of a race. This finding is further validated by the fact that a 30s sprint is often a major factor of a race outcome in the case of a pack finish.
We also understand the concerns that the 30s power best may have come outside of a race context with fresh legs resulting in an inflated 90 day best. This is addressed by a couple of things. Firstly, 30s power is weighed much lower than 10m power in the seed formula. Additionally, we separated the seed from the floor allowing scores to drop after the initial seed. Again, the seed is just the starting point, from there the rankings are result-based.
No method is going to be perfect, and some riders may feel over or under-seeded. However, our approach aimed to reduce this by analyzing historical race data and minimizing prediction errors. We appreciate your feedback as we continue to refine the system. Our goal is to keep Zwift fun and encourage all activities, not just racing.
Changes in Scores After Reset
In some cases, your score may be higher or lower than before the reset, which is expected. Since we didn’t carry over previous race results or apply the current decay effects, changes in your score were likely.
Transparency of seed score algorithm
We understand the desire for openness about the seeding algorithm. While we’re keeping the specifics close for now, please know that we’re actively monitoring and making refinements to ensure it remains fair and competitive. Your feedback plays a crucial role in this ongoing process – keep it coming!
Scoring Algorithm Fairness and Accuracy
We acknowledge that some riders feel the current scoring system doesn’t accurately reflect their performance, especially across different race types and rider profiles. We are committed to continually iterating and improving the algorithm to better represent your results. Recently, we added podium bonuses to reward top finishers, ensuring that top results have a more significant impact on your score. We will continue to make enhancements like these to account for the diverse strengths of all riders. Please keep sharing your thoughts and experiences!
What’s next?
We’re working on implementing pen race score enforcement at the time of the race, which will ensure fairness by placing racers into appropriate ranges based on their current scores – not when they signed up for the race.
We’re also working on detecting a minimum effort in races to prevent scores from dropping if racers don’t meet a threshold. This will prevent racers from intentionally joining races to drop their score – and ruin the fun for everyone else.
We’ll share more details on these, and many other improvements we’re planning as we get closer to release.
10 Likes
Prove it. What’s the algorithm? You promised to make it public once testing was complete.
My experience is that setting a 30s PR in an otherwise Z1 ride inflated my score far more than improving on my ten minute number.
2 Likes