Currently, if you want to, it’s very easy to not show your full ability over 5mins and so get a lower seed score and potentially end up in a weaker pen.
The fact that your ZRS can drop 15% below your 5mins seed score, increases the probability of being allowed in a weaker pen, especially for those in stronger pens eg. 15% of 700 is 105 while 15% of 270 is 50.5.
Dropping below 100% of your seed score, over whatever duration, simply should not be possible.
Not many routes have hills that take 5mins for stronger riders to climb.
Not many routes used in races have 5min hills in them.
But over the weekend, Tiny races used the ~1Km tarmac side of Sgurr for all four races, which took stronger members of pen D approx 2mins to climb.
Which made me think again that a shorter duration aerobic evaluation for seed score would put more racers in their true ballpark ability.
For a simple example, threshold W/Kg for pens A-E…
A: 6+
B: 5+
C: 4+
D: 3+
E: Under 3
Or is 2mins too short? Should seed score use at least 3mins?
I think in general 5 mins is completely reasonable, and would work as a great seeding mechanism if the other aspects of racing score worked properly (score movements better reflect who you beat, anti-tanking, etc). They should also move to less static categories, or have some races that use dynamic pens, to reduce the single target powers sandbaggers can keep under. It’s true there are not that many 5min climbs in races but at least putting in a max effort of 5 mins is probably a bit more likely than a maximum effort at higher time lengths (such as 20 mins which is what we originally had), so going longer isn’t going to help.
In terms of shorter times.. Not sure, but 2mins I would think contains too much of an anaerobic component, so some folks who have a huge 2min power could be put in a category they could not at all keep up in and once upgraded even in a flat race would be completely dropped every time.
3 mins might be a bit better, not sure what the data would show on that.
Edit: I don’t think it needs to be impossible to drop below seed if the ZRS movement works correctly. The seed puts you about in the right spot, and you can move up or down based on results. That said, there needs to be a minimum based on the seed as part of the anti-tanking measures.
Yeah, there’s lots of other things wrong with ZRS, but I think making 100% seed score of whatever duration would at least stop some racers playing the system by dropping to a weaker pen.
I’ve also wondered at times about whether the Zwiftpower Zpoints gives a good reflection of rider strengths for seeding, but as you say, judging a 20min effort opens up lots of scope for sandbagging…
1min: 1 point per Watt and 60 points per W/Kg
5mins: 1.2 points per Watt and 85 points per W/Kg
20mins: 1.5 points per Watt and 1.5 points per W/Kg
Quick scan on zwiftpower shows 3600 to be the highest score per oldskool pen and most pens have at least one racer with that score or something close.
The 5min w*w/kg Compound Score is really good. Works great with vELO.
And a results-based system is supposed to move you down below your seed score if you lose, I don’t understand why people think it should only go up from your seed score. Some people’s seed score will be too high, especially if they’re not a sprinter at all.
The issue is that ZRS doesn’t work well as a results-based system. It seems like it’s 100% assuming that the categories are correct and just looking to drop or promote people who are winning or losing all the time - but most of us don’t race more than once or twice a week, so it just isn’t fast enough.
Races come down to short power as long as long power is close enough. Power bests aren’t necessarily a good measure as some people can recover and hit their best again much more quickly than others.
I think I misspoke on my response. The 15% thing I was referring to was the issue where a new 5min PB would not set you directly to a new seed, but stay 15% below. Now that I think about it I think they fixed this issue, so I’ll edit my post. The seed “should” be a starting point, and does not need to be perfect as long as the ZRS movement works properly, there should be some anti tanking, and a minimum you can drop to, but there can be movement both up and down.
Preventing people dropping below their seed is primarily an anti-tanking method. To me is a worthwhile one as the seed score also sets an established ability. It’s impossible to say whether the score is based off a maximum effort, someone may be capable of performing better than their seed suggests but they can’t be worse.
If people perform better, shown by them beating people with higher scores, then their score should increase. If the score movement worked well people’s scores should settle at something close to what the seed would be if they recorded their perfect 5 minute interval.
I would prefer to see the seed score decay after say 30 days, as people may no longer be able to match that effort and so their results should start to move their score down.
It’s important to remember that it’s not just sprinters who should see their score rise. Someone with a flat power profile may be able to hold close to their 5 minute PB for an hour, largely because they’ve also not recorded a true maximum effort over 5 minutes.
And also remember that ZRS did not allow scores to drop below seed when it was introduced. That was changed as a response to the poorly conceived version 1 seed calculation, and wasn’t reviewed when the seed calculation was changed.
I mean, it can go either way. The seed can be “the minimum score”, or give room to move down. Honestly it doesn’t matter if you don’t know the pen boundaries and everyone is under the same rules. You can have a system where the seed is the minimum, or where the seed allows some movement down or up.
There needs to be some anti-tanking either way, whether seed is the minimum, or whether you can go below seed.
I feel like allowing scores to drop below seed is too easy to exploit, and the impact of that is more significant than allowing scores to drop if someone can’t quite perform at the level their seed suggests.
Functionally the current floor works the same way as seed=floor, just that some people enter the system with a score at the floor and others with a score 15% higher.
Moving the pen boundaries, ideally making them dynamic, would be a much more significant improvement. Add in terrain adjustment and score movements that reflect the relative scores of the field and ZRS could become a good system.
Yeah, I think this would be a great change. Having a couple of the bigger Zwift races where numbers are not an issue at least having dynamic pens, and enabling race organizers to use dynamic pens if they would like. Hard to sandbag when you have no actual target to aim for.
Nope. IMHO, there needs to be two components: A short one and a longer one. Say, 2 minutes and 10 minutes. Just 2 minutes would be giving the Diesels like me a free pass; I have no discernible sprint but I can chug along very nicely at a steady 3.3w/kg for quite a while. Just 5-10 minutes gives the sprinters a free pass and we see that in things like the Tiny races. Or any race with a 2 minute climb. ZRS needs to take both into account and fold that into the results. Without the stupid decay function.
All of this is just my opinion but I seem to recall that when they were using two power points to set the seed, we had much closer racing.
That’s ok. Again, dynamic pens would be an option, not the only option… some might prefer the dynamic pens, some might prefer the static pens etc. For GC races an organizer can use more static pens etc.
Why? Again they would be an option, not the only option. Race organizers could decide to use them, and if they don’t like them they can use other options already available.
For a series where there’s not a lot of participation or for GC races it wouldn’t make as much sense but for larger races you could get some great matchups.
Pure ZRS, as it stands today, represents ability really badly.
Two riders with the same ZRS can have very different abilities made up from their seed score and whether they typically finish in the top or bottom half of the pen field compared to the total signups for the pen. One could be podiuming most races, while another could be near or at their 85% seed score floor.
Even seed score (best 5mins at a given weight), as it stands today, represents ability badly. According to ZWIFT ZRS v3 by VirtuSlo , 4W/Kg for 5mins at different weights gives the following seed scores…
30Kg: 239
40Kg: 288
50Kg: 329
60Kg: 363
70Kg: 393
80Kg: 418
90Kg: 440
100Kg: 459
110Kg: 476
120Kg: 491
etc.
That same ability of 4W/Kg for 5mins to go up a virtual climb like Petit KOM is splitting them into at least three pens, when that climb would almost certainly be the selection point for who gets to compete for the race point.
They ought to all be in one pen or two at the very most, especially when you consider how positively crazy Zwift physics are compared to real life.
Simply splitting 100 riders into five equal pens of 20 is asking for serious ability trouble!
If it’s a well attended race this would be unlikely. Dynamic pen logic can also be tuned to ensure a massive spread would not happen. Dynamic pens do not mean each pen has to have the exact same amount of people in it, it just means the logic would split the group dynamically based on the ranking of the people who enter. The top pen might still be smaller than the second pen etc.