Sure, I can do that for the 100th time.
6 Likes
I see a number of posters here jumping to a false conclusion. Putting aside any discussion of he merit of ZHQās calculation, Zwift Points is not a ZwiftPower calculation. Itās on display of all of us to see on ZwiftPower but this is clearly different. Because itās race-based, itās also not the same as zFTP, which counts every time we log on.
To emphasize, Zwift Points are not based on previous ZwiftPower figures or zFTP.
Using essentially a USAC ranking system for this is so funny because not even USAC uses it for categorization! Thereās Ranking Points and thereās Upgrade Points. The Upgrade Points are the true results-based categorization, and tell riders what races they enter and who their competitors are. Ranking Points are a cool shiny thing to chase, but have no bearing on organization or matchmaking of races. Why you would use one for the otherās purpose, other than the fact that it was just there already for bogus ZP scoring, is mindboggling to me.
2 Likes
I dunno. I havenāt been racing since the end of ZRL and my ZRS is 20. 
1 Like
Maybe his score is really high right now, but if he were to race a few more times (and lose to the riders who are stronger than him) then his score would level out to what it really is/should be. After a little while of racing, scores for everyone would begin to even out and then begin to represent peopleās abilities more better.
Then race organizers could say, I want to have D=0-250, C=251-500, B=501-750, A=751-1000, or whatever combination or number of cats that they want.
1 Like
I donāt think anyone is saying itās a zwiftpower system (though unless theyāre going to apply DQs, thatās would actually be better). But itās in essence the same as the zwiftpower race rating system (avg of best 5 results in 90 days, based on race quality and position, just now higher is better.
Sweet, I am almost an A racer because I finish high in some C races! Seems legit. I only have to Race all A races for 90 days to get a lower ranking.
2 Likes
You will never get this until they fix the group drafting dynamics, itās a joke. I can ride 1.7 watts in a large pack and do 47KPH. I can do an extended effort at 4.5 off the front and get caught by a large pack with someone doing 3.2 on the front. #NoBreakAways.
3 Likes
Here is my brief explanation of why this is a case of square peg > round hole. Iāve explained this in a lot of detail elsewhere as we developed ZwiftRacing.app and refined it to what we have today.
There are, principally, 2 ways to rank participants in multiplayer events. One is a rating system. This is known from Elo in chess, and various takes on that in MMO online gaming. The idea is to assess how a participant performed versus the skill of their opponents. Beat higher ranked opponents, move up. Lose to lower ranked opponents, move down. In ZR.app there are a lot of modifiers applied to this principle, for example seeding initially by zrCS (a reliable power metric for predicting success), podium bonuses, anti-sandbag measures, confidence considerations, etc.
What is great about a rating system is that its principle objective is accuracy. It wants to get every rider to the right point in the global rider list as quickly and accurately as possible. If someone is in the wrong place, as soon as that is identified there is a strong corrective movement. That means itās not really about incentivesā¦ to get a higher rank, you need to get stronger and perform better. So it doesnāt necessarily drive participants to pick certain events, or race more often. But as it is accurate, it is a superb way to split riders in to pens for competitive and enjoyable racing.
The 2nd way to rank riders is a points system. In a points system, your best results are taken in to account to give you a score. This means you can maximise your score in numerous ways, normally competing more frequently, doing well in tough fields, varying the events you participate in, etc. Points systems are great to incentivise participants to race more. ZR.app also has a points system. For many, this is more of an incentive to race, as they chase points to get the best score possible. Weāve found that many riders are only interested in improving their rating (rider versus self) whereas others are interested in maximising their points (rider versus others).
When you place riders in to pens, the only sensible system to use is the rating one, because it is the best sorting of riders by capability. Points are subject to too many variables.
Zwift have chosen a points system, which would be fine if they werenāt using it to sort riders in to pens. However, they have made clear that is the aim, in which case this is completely the wrong approach. Square peg, round hole.
Edit:
Just adding that the correlation in the graphs isnāt made up, it is evidenced through data. That stuff scientists likeā¦ it is proven through the reality of results versus the system prediction.
20 Likes
what you are describing is a ranking system (e.g zr.app) this is not that scores donāt go down when you lose with this new ZHQ system.
No, because score cannot go down (unless it times out). You canāt drop your score by doing badly.
3 Likes
Oh, maybe I misunderstood. It seems weird that Zwiftās system wouldnāt allow you to drop w/o timing outā¦
so does this mean that CE and by extension zFTP/MAP is no longer needed ?
The fact that James canāt see a problem with a 700 point D and tried to justify that the system works I say is a cause for concern.
I can already see the type of inflation at works when a 700 point rider consistently lose to 300 points riders.
1 Like
it only drops when one of your best 5 results goes beyond the 90 days.
A bit like the crazy old days where your average of 3 best results giving you a zp cat ranking.
1 Like
Thatās because itās a points system not a rating system. See post above for explanation of the differences. This is something I tried to point out when they first mentioned they were considering a modification of the ZP rankings, many months ago. The two approaches were well debated within the zr.app community.
3 Likes
Aha, I get it now. Your graphs now make much more sense. It helps when you read everything first 
1 Like
The lightweight junior topping the old Zwiftpower ranking pen D has ~797 points and they pretty much exclusively race in the A pen!
So how is the system an improvement over ZP ranking.
I see it changed the scaling from 600 to 1000 points,
They inverted the score. So 1000 is best vs 0 is best.
But what else was done in the months of testing?
2 Likes
That 700 point rider would need to finish in the top 10 of the race for his score to have any effect on others in the race, makes more difficult to tank to inflate.
If you are racing other 700 point riders your not likely to improve your rankā¦ post 90 days you will start to head where you should be without Sandbagging bonus pointsā¦
1 Like