Zwiftpower’s (ZP) ranks categories based on W/kg. This system has three characteristics –
- It’s completely transparent.
- It’s simple to understand.
- It allows reasonable, if imperfect, categorization.
Following on the third item above, Zwift’s new category enforcement (CE) should give some correlation to ZP. However, the following example (for one) shows that this isn’t necessarily true.
Looking at a certain racer’s profile (62302), from the beginning of 2022 through July 18 he raced in 145 ZP C races and trophied in 33% (20 gold, 16 silver, 12 bronze). Just considering the 20 gold finishes, the average number of C racers was 9. Thus, it wasn’t like this rider had no competition. Thus, this racer is a strong C under the ZP rating system. (During this same period this racer also competed in 16 D races where he got 13 golds - 81%.) Note that ZP rates this racer at 2.88 W/kg at 20 minutes and his power curve shows 1000 W for 10 seconds over the past 90 days.
Now, under the CE system this racer was allowed to compete in four D races where he took gold in each. Thus, I submit that there is a significant problem with the current CE system when a strong ZP C racer is allowed to compete in the D category in CE races.
===========================================
For comparison, let’s look at my stats (profile 636022). For the same period I have raced 31 times – 19 D, 6 E, 6 C (the last are CE). Of the 25 D and E, I trophied 9 times (3 gold, 4 silver, 2 bronze). Thus, I’m a strong D racer under ZP. (Compared to the above racer, I’m rated at 2.54 W/kg at 20 minutes and only 281 W for 10 seconds over the past 90 days.)
Under the CE system I’ve had to compete in the C category. There I’ve raced 6 times where I finished next-to-last 3 times and last 3 times. (The average number of racers was 16.) Since I’m a strong D racer under ZP it’s certainly possible that I’m a weak C under CE; I can deal with that. What I don’t understand is how the above racer can be a D under CE while I’m a C under CE.
Looking at the current CE system –
- It has limited transparency. As far as I know, no one has provided a complete, detailed explanation.
- Because of limited transparency it’s difficult to understand. I suspect it would be difficult to understand even with complete transparency (i.e., how easy would it be for someone to manually calculate their category to verify Zwift’s categorization; with ZP it’s easy).
- At least in the above instance CE seems significantly wrong. Likely there are other instances.
Generally I’m a strong supporter of category enforcement. However, it seems to me that Zwift’s algorithm needs to be corrected.
