Did the oldskool Zwiftpower rank score ever do anything as regards pen allocation, or was it always simply a random number, that didn’t reflect relative ability for most racers?
For me personally, it has always been a nonsense number, it doesn’t remotely reflect my ability aginst my pen D peers, except perhaps a little for those that race against me in the Tiny Races.
Was a simple formula applied to each rider’s oldskool Zwiftpower rank score, to convert it to an initial Zwift Racing Score, which could then be used through the last 90 days of races each rider did?
Were any race results ignored for any rider’s last 90 day history, because for instance they were a pen D racer and they completed a strong pen A race?
Were any DQed race results considered for adjustments over the last 90 days, because a rider had not followed a series rule, for instance joining races 3 and 4 of the 1000-1100 Tiny Races and getting a win and second (but were then DQed because of the “no sniping” rule)?
Were there any adjustments made to post-race Zwift Racing Score changes, to take account of things such as a rider doing a lot of short races like Tiny Races, but rarely taking part in races that took more than 20 minutes?
What happened if a racer got pen premoted or demoted in the last 90 day evaluation period?
If I was trying to do something loosely based on this Zwift Racing Score, as a starting point for each racer, I would start by splitting the 1000 into four equal parts, one for each “pace group” pen. I would then start each rider in the middle of their pen’s range (875 for As; 625 for Bs; 375 for Cs; 125 for Ds). Every rider starts with a clean slate, no presumptions on relative ranks from the flawed oldskool Zwiftpower rank score. Then work the 90 day race history, ignoring anomalies such as a pen D rider racing in pen A. If a rider changes pen, their rank is reset to that pen’s middling rank.
… Or something like that!
The evolution of the zwiftracing.app ranking system over last winter, with latest solutions to problems kicking in on 3rd June, has been great to see. About the only slight concern I have at the moment is how custom length races are categorised by the “flat; rolling; hilly; mountain” modifier. But the important issues, like flagging riders who go slow or join pens much higher than their ability, all now get caught.
