You missed it…https://forums.zwift.com/t/competitive-ai-tech-update-september-october-2022-recap/
If only someone could have predicted that, eh?
When it was launched I thought it was a step in the right direction and there was a decent idea in there… Currently il put my hand up and say I was wrong.
Im not sure if its failing as its a bad idea or been a shambles of an implementation. But ownership or focus seems to be lacking, so its unlikely to improve any time soon.
If there is no commitment to the roll out of a scoring system (both short & long term) and its follows all the other deployments im not sure its going to be saviour people think it is.
If there’s one thing we’ve learnt from the ZwiftRacing.app work it’s that there is no silver bullet. The biggest win would simply be passing the onus on creating great race experiences to the community. Give them the toolkit to define how races are organised, to be able to communicate that structure in game, and to enforce the pens in the way that the organiser has prescribed. A ranking system is a cool way to split riders, but there are many other ways too. That’s the not the crux of the problem. Static,. poorly understood, miscommunicated boundaries are the problem.
It think the idea of the CE system is great. Not using just one power number is a huge win. CE probably need some refinement after some of the found limitations.
The thing is, they have been refining it. Secretly. And now no-one really knows where they stand any more and it’s a confusing mess.
A simple webpage that shows your calculated CP and MAP and where you sit within the current boundaries would solve all the problems. Any query you just point the person there. Simples.
They have not been refining it to improve the system, its been to reduce workload and complaints. As soon as you give up on the principle you started with you might aswell pack it all in & go back to the drawing board.
That makes sense, also two systems does not help either. ZP limits should have been removed and updated with CE.
But one thing that I am thinking is while there is tweeking behind the scenes it is probably better not to publish the data until they have a lot more confidence.
This is an issue, Zwift is not aggressive enough to handle a racing system with strict rules. That is why ZP is so popular, there was a set of rules and it was enforced.
Category Enforcement FAQ was a link given in Flint’s recent Competitive & AI Tech Update: September/October 2022 Recap post, showing the current thresholds for VO2, MAP and CP.
However, unless I’m blind, it doesn’t say what time durations are being looked at for VO2 and MAP.
Am I right in thinking MAP is basically 5min?
“VO2 Max” at https://intervals.icu/ suggests it is also typically based on 5min efforts.
Pen allocation by Compound Score (W/Kg for 5mins multiplied by raw Watts), with no fudges like these silly minimum Watt floors, no +5% buffer would be a good way to use power and everyone can clearly see where they are.
But bring on ELO ranking based on results, CE has been a messy ~9 months.
As does the fact that they pointed us all to intervals.icu in the original CE post to calculate it
(Edit to say: the above is not correct. It’s just the way I remembered it.)
But given how in that FAQ, VO2 is lower than MAP eg. B is 50+ and 200W+ for VO2 and 55VO2+ (4.2W/Kg+) and 200W+ for MAP, then presumeably VO2 is a longer duration than MAP.
‘they’ certainly didn’t. I made the most of the available information to say how best to work out your own values to explain the CE results.
The OP is no longer valid. The rules have now changed.
The use of VO2 at all is absolute nonsense. VO2 max is not a power/duration value. They have only used it to try and link a pretty round number to the MAP result.
Confusing because it was changed. Back in May, I was a B in cat enforcement as I did a 6 minute race and got 4.1 W/kg, which put me over 50 VO2. Now the threshold for B in CE is 55 which is ~4.35.
That I can get my head round. I have no idea how MAP works at all.
I’ll go back and edit that then. Was sure I’d seen Zwift publish a plug for intervals.icu though. Perhaps that was after your initial post?
Just me I think (I’m a bit of an intervals.icu fanboy)
I might be wrong but I suspect @xflintx asked staff to update this FAQ chart and they updated as he asked but put new VO2 figures in the wrong column. I think VO2 figures were originally only shown under the VO2 max column not the MAP column.
they should replace the CE faq page with that picture, scribbles included
CE numbers got changed to match ZP tolerances for some reason to give some leeway, now, I have heard people saying it was because there were certain zwifters who would be bumped out of B and into A, probably inaccurate but still, but I think the change was done to lower the number people asking why they were upgraded, the slight leeway doesn’t get them upgraded and thus less support for Zwift to deal with
should have just gone with pen enforcement as that was what was requested initially to stop people screwing races up by entering lower pens but Zwift went in a whole new direction and have now figured that after 10 months of it being in use and being tweaked/changed etc its still not fit for purpose and are looking at a ranking system, 10 months of resource management that could have and should have been used in a different manner, very similar to to hardware venture that ended up being so successful
I don’t think this was wasted effort. We still need a good system to deal with people that has no racing data.
Few quick fixes that would go a long way to sorting this out:
- Use Compound Score and get rid of the raw watt floor
- Show your MAP and CP values on a web page after logging in
- Show your calculated category based on those values