Race Scoring (Ranking) - a new discussion

Continuing the discussion from Competition & AI Tech Update: July/August 2022 Recap:

Looks like Zwift may well be about to use a Race Scoring system we are familiar with.

Would anyone like to comment on the ZwiftPower / USA cycling ranking system and how it might be improved upon?

What extra add ons or changes might make it more suitable for Zwift racing?

some hurdles, not taking into account the many and varied possibilities of manipulating your rank to suit your own purposes:

race quality and availability of similarly ranked racers for whatever event you want to ride might mean you end up in an empty pen in which case why bother racing at all

race length: some people race short races (i have averaged around 12 races a week for a few weeks) and would generate disproportionate amounts of rank compared to someone who only does 70-100km races

hardly anybody races courses with 20+ minute climbs in them, most people want to race on flat courses. sprinters are still going to win everything anyway

team races: what if a local club wants to ride together but one or two of them puts 300 hrs a year into zwift and the rest only use it for zrl in the winter or TTTs or something. they would have to split up and join some huge zwift clan like coalition or dirt or whatever to find people if they wanted to race in a league

1 Like

Most of these criticism seem entirely spurious:

How is this any different from the current system? If the boundaries were set dynamically (which wouldn’t exactly be hard, though we know how long these things take here) then it could easily be better.

The ranking system that’s currently in existence uses your best 5 races, compared to the power-based best 3 and the CE system that uses all of your rides (well, the calculation will take however many points they use for the curve fit, probably 3-8 I would guess, might be all in one ride but it seems unlikely). It doesn’t seem a radical change to me.

This is a problem how? They will be up against weaker sprinters with higher FTPs which might give the latter some chance of breaking them. Surely you’d agree this might be more interesting than rolling around in Z2 waiting for the sprints?

Already a problem with riders of different ability in a small club. That’s the way zwift - and the real world - works.

Your comments seem to be predicated on the assumption that ranking based on results will be less accurate than ranking based on current methods (ie, it will generate poorer matching of race performance). I don’t see what your basis is for this belief.

1 Like

Should Zwift produce the best Scoring system they can to benefit the great majority of honest racers.
Or
Produce a system geared towards preventing sandbagging, to the possible detriment of everyone else?

Does that beg the question, for Zwift or Race Organisers, as to what pen sizes make for the best races?
( the ability to have lots of pens doesn’t mean race organisers need to use all these pens)

If pens are split by Race Score then do you not bring that range of rank to the race irrespective of whether you race 14 times a week or once a week.

( I do, however, suspect that those who race very regularly and well, under a ZwiftPower system, do eventually generate an ever decreasing/improving ranking based on that 90% factor in the Race Quality calculation)

they are predicated on the assumption that there are logistical hurdles to consider first, the point is it’s not an entirely obvious fix to all of the current issues so they will need to take stuff like this into consideration if they are going to implement it.

they can categorise us however they want or delete the entire category system entirely if that’s what the majority want as far as i’m concerned. who am i to speak for anyone else

1 Like

Yes agree short flat Crit courses where they all end in a sprint even though most of us have no chance of a sprint win. But we all feel we are in a race until the last minute.

From a Race Score perspective:
Those who do well in Climbing races will get their good ranking there.
Those who do well in flat short races will get their good ranking there.
But a good ranking may not transfer well if they try and race in their weaker race format. Possibly racers with good rankings may have to choose the races which suit them best if they want to finish on the podium.

1 Like

My main concern is the suggestion that CE will go away and we will move to a fully ranking system. Which I’m unclear what that means for new or returning riders.

If a new A rider comes into zwift does a social ride at 4 w/kg then thinks about trying out racing do they have to grind their way from the lowest rank up or possibly a better idea have some sort of CE applied minimum per rank so if zwifts knows you can do 4 w/kg there should be a minimum rank,

Also it’s unclear how volume or racing works. If i race 7 days a week does that gain me more points than say a rider that only races occasionally. There is huge proportion of Zwifters that fall into the occasional bracket that need to be ranked correctly. Where as I i race 5 days a week would i find myself in some stupid rank based purely on volume of races done.

No - Not if we use the existing system that Zwift Power uses (based on US Cycling).

that makes sense, what about a w/kg floor will that be applied?

Nearly everyone with be a social/free rider first before trying out racing are they going to be able to go smash a cat d/low rank race at first to earn some ranking points before moving up?

Only if you are finishing in the top half of a field of roughly equal ranked riders. If you place in the bottom half you will not improve your ranking no matter how many races you do each week. The most ranking gain is obviously made by finishing first with then gradually diminishing gains until you reach the rider who finishes middle of the field.

(UNLESS of course you are finishing bottom or towards bottom of a race with riders of much higher rank than your own, and probably a fairly small sized field. I believe theZP style of ranking system works best when races are contested by riders of broadly similar ranking. Encouraging riders to race in much higher quality fields than their own ability can lead to a slight skewing of the scoring system.)

I do, however, suspect that those who race very regularly AND WELL, under a ZwiftPower system, do eventually generate an ever decreasing/improving ranking based on that 90% factor in the Race Quality calculation.

We really don’t have the answers for these questions yet.

So this will all be dependent on what time zone you tend to race, if you ride in a very quiet time Zone you may be limited in what Ranking you can get so if you then ride in a popular Zone you will be a “B” in a “C” or “D” category.

correct… a couple of years ago the top ranked racer, probably kirchmair or someone at the time, had a race ranking of about 100. now there are people in the low 60s-high 50s. i couldn’t begin to explain the math behind why but i have seen it mentioned a couple years ago and then observed it myself

What needs to change, is the seasonal decline in points. For the last month I have climbed back to about the points I had last season. I assume this happens for all returning racers. This means that, if we had categories based on these points, the lower categories would be littered with stronger riders until/if they manage to reach their actual points level.

3 Likes

Yes agree, but isn’t that what we have now in that in ZP after 3 months and CE after 2 months your race category can revert to D either by not riding or not riding hard.

With a Race Score ranking system might there be an argument that the ranking of a racer should only change based on race results?

If you don’t race your ranking doesn’t move - EVER?

How about a system which doesn’t just take your best 5 results but possibly takes, say, your best 5 out of last 8 (10 of 15 etc ?). This would provide a system dynamic in both directions not just currently upwards as we have now.

The immediate response will be Ah but this means Sandbaggers can race badly for a few races and obtain a worse ranking and then win in the easier category. Can’t they do that now it just takes them a bit longer.

A system which took best x of last y might provide a better system for those struggling with form.

I favour a best x of last y I’m just not sure what x and y should be.

1 Like

The difference I believe is… For CE it only takes one effort to keep an A rider out of a D race, and that effort can be across any activity. Where to go up ranking in the ranking system you’ll need multiple actual races, which for some people that might take weeks, or the whole season to get slotted in correctly if they don’t race much, and if you start low you are racing against lower ranked folks which won’t increase ranking fast to begin with, so you can work your way up categories a lot slower if you choose races with low overall ranks etc…

2 Likes

Points categories also have the effect that riders with special abilities (climb, sprint), will only be competitive if they race to their abilities. It is kind of self amplifying.

With some additional complexity (scoring riders in races by distance and elevation) it would be possible to have a ranking system that gives sprint-race winners a lower category in climbing races, short-race winners a lower category in long distance races, etc. I don’t imagine this will be in v1 of the ranking system, but it is certainly possible. I think of it in terms of: what would it take to encourage a winner of short sprint races to enter a long distance climbing event? What would it take to encourage a winner of long distance climbing events to enter a short sprint race?

1 Like