I’d also like to say that Tim and the folks that have helped make ZwiftRacing.app what it is today deserve a massive shoutout for all the work that’s gone into that project. It’s incredible to see. Well done everyone!
Zwift Racing Score isn’t meant to stamp out ZwiftRacing.app, and it’s also not meant to be a replacement for it. We took a pretty different approach to scoring than Tim with some different goals in mind, though we share a lot of the same goals too.
Tim definitely beat us to the punch on scoring Zwift results, and that also deserves a chapeau.
If ZwiftRacing.app is your jam, then please continue to race with that in mind and check your results there. We aren’t here to demolish everything with a bright orange steamroller, we are just working on making racing better as best we can.
Obviously a great time to grab some popcorn lol. How will this system work with the sandbagger’s/cruisers/ “Race Craft guru’s”? A High race score will be like Catnip to them lol
I’m disappointed this still doesn’t address the issue that in A category, riders continually drop out when they’re not doing well. In this set up, if it’s not going to result in one of their top 5 performances, they will just drop out, reducing the ‘race quality’ and removing the potential for improvement for all the other riders in the race.
No-one is annoyed that it may replace zr.app, just that the learnings from there have been, broadly, disregarded. The whole point behind it was to demonstrate the potential so that hopefully Zwift would eventually do the same in the core product.
A rating system is good for sorting riders in to pens.
A points system is good for competition, something to go after (e.g. seasonal points competition).
A points system fundamentally doesn’t work to sort riders in to pens, as already clearly explained.
The frustration is that the community have discussed these nuances for months and developed an incredibly accurate (key word) rating system, and yet that has been, almost deliberately it seems, disregarded, and you are trying to jam a round shape in to a square hole.
I do hope the few pen A races I’ve joined simply to avoid the Tiny Races “no sniping” rule after technical issues or last night to have >5 racers in a pen haven’t boosted my score, they should be disregarded, just as they are now at Zwiftracing.app
Giving me ~117 points for simply joining A a couple of times and finishing last is a flaw of the oldskool zwiftpower rank system (which was let’s face it, is a meaningless stat that didn’t allocate riders to pens and certainly didn’t rank racers by abilty) and really should not be repeated in the new Racing Score, IMO.
This exactly. I think the majority of contributors at ZR.app would be happy to have the ZHQ implementation replace it, since that would mean the broader racing community ends up adopting results-based ranking and matchmaking.
But the worry is that whatever this official implementation evolves into will become the de facto standard for results-based racing in the ecosystem. And as James says, a points system with no scaling for ability, no downward revision mechanism, easy-to-game holes like Tiny Races, etc is just a miss. I’m hopeful to hear the ZHQ team is open to iterating on the system before the fall Zwift season really picks up, but it seems like we’re a long way off from a functional system, especially when one already exists in a much better form with ZR.app.
As a general consensus, if you keep winning you should race harder or better opponents… I get that’s not how zwift has worked in the past but that is a move in the right direction.
It just might take a good while to balance things out from the starting point of a broken race entry model.
If we race ITT races and a bunch of them, we can inflate the points in the whole eco system. I can see a cat B-A rank points of 700 doing many D races, can cause a systemic breakdown of the whole system.
so… this isn’t the final product, but what’s the ultimate goal/purpose?
What is the measure of success of the work towards implementing this? Is it more enjoyable racing, more interesting racing, talent identification, or athlete improvement? (something else?)
I just cannot believe that the best that could be done is pretty much what’s already there, but on a different scale where higher is better. And bundling Tim’s work in with other highly inferior rating systems seems like a huge slap in the face. It’s another sandbaggers paradise those near pros who don’t race very much on Zwift will come in and clean up in B cat. WTRL will be a farce as usual. Very disappointing.
Zwift Racing App is a net negative for zwift from both a company and a user base. It creates confusion on which system races are using, meaning people sign up based on it or miss that they should have. It has no integration so races are just impacted if they try to do something that isn’t already the base system. It also can’t solve things that zwift is trying to solve here with regards to initial category placement.
I agree with the feedback this doesn’t look usable but let’s give them a chance to figure it out. It’s not being used for anything now, and that’s a positive thing. It could be a lot better but there’s a trade-off with getting a preview of an early feature: you get to see it before it really works. My hope is that like zwiftracing.app the ranking will be seeded with power data and continue to be affected by changes in power data over time. A lot could be learned by taking another look at zwiftracing.app’s ranking system and taking on board the lessons they’ve learned. Another vote for the seasonal reset here.
ZR as a matchmaking system is significantly more robust than ZP points. Happy to share metrics that we have that shows this. This Score system is great as a points-reward system, but there are some fundamental flaws that will make it unusable as a matchmaking system.
You’re right that having several categorization systems is confusing. I’d much rather there just be one system directly integrated into Zwift. But that system should be robust and accurate for placement, which this Score system is far from.
Fairly sure ZRL would be a mess on any system. People come back to the platform for it then slope their pedals / alter their weight as needed for their purposes. Very hard to assign an initial category that lasts as they adjust to online racing.
Interesting seeing the feedback here - I didn’t think it would be this negative. I see this as a positive step toward results-based categorization, which we all want. The scoring formula can certainly be tweaked, but getting everything else in place (score-based category enforcement, UI that makes it clear what’s going on, etc) takes time too.
That said, knowing how ZwiftHQ works, I would advise those who are experts in ZwiftRacing.app’s algo to put together a concise yet complete post explaining how it works and what problems it fixes from ZwiftPower’s system.
Just saying one is better than the other won’t get anyone anywhere. Sometimes systems are just “different” but not better/worse. Can someone here make a solid case for ZwiftRacing.app’s system, within the context of what Zwift is trying to accomplish?
Ultimately I want to log onto Zwift, join a race, and not have a choice of category. I don’t care what happens behind the scenes to make that happen. I care far more about being a part of an interesting race that isn’t a 30min group ride with a sprint. Accurate matchmaking may work against that goal… that’s why I’m asking “why” above.
I’m sure that ZwiftRacing App is accurate and good and all, but it’s essentially a neat toy that exists independent of the racing experience.
Gordon, I agree that this should be based on a “Date-Forward” calculation rather than a “Lookback” calculation. I DO come from a different angle.
After ZRL, I was burned out and decided to go back to an old friend of a workout plan. I’ve only done a couple of races since, not counting TTTs. What with a family reunion, bronchial problems and what seem to be ever-escalating Grandpa duties, my Zwift Score is … wait for it … 20. Based on what I see here and elsewhere, that’s correct.
But I am … or at least my ZRL me was … considerably stronger than that.
Which is why a date-forward system makes sense. I didn’t know this was coming, so I didn’t keep up my race chops, feeble as they are.
Zwift’s is integrated with the platform therefore it is better. This is a good step forward in terms of rankings being integrated with sign-up. I’m more interested in how it alters racing long-term.
