Category Enforcement - How is my category calculated? [February 2022]

Sorry to hear James.

You’ve been a fantastic voice for change.

Sheit! How in the world are we going to get to 4k posts on the Anti-sandbagging thread without you to push things along, James? Your contribution will be missed for sure. We all hope to see you back here sooner rather than later.

i dont blame you. anyway regarding CP, i do think the system is fundamentally sound, details aside.

Stay safe and stay well. You will be missed by many. As others have said, great progress has been made but sadly a little too late. Like you said, far more important things to worry about, plus other platforms are certainly making far more progress. I know I’m checking them out after the recent tests and no doubt others will as well

We may not have always seen eye to eye on the best way to approach things but regardless your efforts are appreciated.

What i would say, if you enjoy doing something dont stop doing it - Take a back seat, block the forum for a few weeks but if you enjoy looking after your team and riding keep doing that.
Its your enjoyment, dont let outside forces change that,


Is Critical Power (and the additional MAP component) really a good metric for identifying who should be moved up to the next category? I have always found CP calculations to be very sensitive to the quality of the best efforts included. Especially something like the 3 point test in the Highnorth spreadsheet referenced earlier in the thread.

Examples using the 3-point test on the Highnorth spreadsheet - for a 70kg C@3.12w/kg 95% of 20min
with the same 10 min and 20 min best efforts and a variety of different 3 min best efforts. (I don’t think any of these power numbers are outside of the bounds of possible riders)

Person #1 - C Cat
320w for 3 min (4.57w/kg)
250w for 10 min (3.57w/kg)
230w for 20 min (3.29w/kg) gives
CP=216w (3.10w/kg) W’=18.7KJ and 5 min Power of 284w (3.98w/kg)

Let’s change that 3 min number to someone more punchy and send CP down and the best calculated 5 min power up.

Person #2 - C Cat
335w for 3 min (4.78w/kg)
250w for 10 min (3.57w/kg)
230w for 20 min (3.29w/kg) gives
CP=212w (3.03w/kg) W’=22.1KJ and 5 min Power of 286w (4.09w/kg)

Add 5 more watts to the 3 min best and they have possibly tripped the MAP trigger to the next category

Person #3 - B Cat
340w for 3 min (4.85w/kg)
250w for 10 min (3.57w/kg)
230w for 20 min (3.29w/kg) gives
CP=211w (3.01w/kg) W’=23.2KJ and 5 min Power of 286w (4.12w/kg)

Take away some watts in 3 min power for a less punchy rider

Person #4 - C Cat
305w for 3 min (4.36w/kg)
250w for 10 min (3.57w/kg)
230w for 20 min (3.29w/kg) gives
CP=221w (3.14w/kg) W’=15.3KJ and 5 min Power of 272w (3.88w/kg)

Take away even more watts from 3 min power and you get someone who is catted up (97% goodness of fit, so certainly an outlier, but not an exceptional outlier)

Person #5 - B Cat
290w for 3 min (4.14w/kg)
250w for 10 min (3.57w/kg)
230w for 20 min (3.29w/kg) gives
CP=225w (3.21w/kg) W’=12.0KJ and 5 min Power of 264w (3.78w/kg)

Here, the riders who have been moved into another category are the anaerobically strongest (#3, by MAP) and anaerobically weakest (#5, by a CP above 95% of 20 min). To me it doesn’t really make sense to create a system where an anaerobically weaker rider should be moved up categories before a rider with the same 20 min power like Person #2. I know who I would bet on for most Zwift races. And, this is not even taking into account raw wattage ceilings for lighter riders.

Now the CP and W’ numbers do suggest that Person #5 might be stronger in longer races like a race up Ventoux or a 40k TT where they can leverage their better CP by riding steady and not deplete their smaller reserve of W’ compared to Person #3, but very few Zwift races are long and steady.

One simple fix to this model might be to never let CP in the Category calculation be higher than 95% of 20 min power. At least that makes sense for the edge case where a lower 3 min power moves someone up you would not expect.

Of course, this is probably not the way Zwift is calculating CP, but for most CP calcs I have seen, Critical Power can be higher than 95% of 20 min power for well developed aerobic riders. Those are the exact people that should not be moved up categories if the goal is to move up the riders who too easily win races with the current Category system.


I have now done 2 enforcement races and as I have been switched cat by the algorithm between the 2 races, I will share my experiences here.

The first was way back at the start of the week in enforcement version beta 2 (AND-OR bugfix), where, as a 53-54kg female with a CP of between 185-195W (depending what formula is used), I was in B pen on the Paris loops. I basically clung to the large second(ish) grupetto, doing fine on the hill but of course coming second to last in my group on the final sprint. Someone said that double draft is on for these races so maybe I wouldn’t cling on so well in all races. The pack pace was hard for the first lap and medium hard thereafter.

Then today, with enforcement algorithm version 3 (watt floor implemented) I was able to ride in C, for racing around 2 laps of Ocean cliff side. Several interesting things about this race.

Firstly, someone complained at the start there was a sandbagger, which precipitated me trying to explain the enforcement system in the chat! I too thought there could be no sandbaggers anymore. But the “sandbagger” won! I think that makes him worth looking in to: ZwiftPower - Login Other racers should not a priori be able to predict the winner…

The pack pace on the flat was easier than in the B race. Up the cliffside I could hang with the front group, but couldn’t break off the front. It’s a fairly punchy climb after all. On the downs I had to pedal - sometimes quite hard - to stay in touch. There was another woman about my weight in the race and I seemed a little quicker than her first time up the cliffside, and she a little quicker than me second time. But there was also a male of the same weight (53kg) and he did a big break away on lap 2 and got 30s up the road before the following group got our act together and … we did chase him down.

At the finish someone went long off the front up the ramp out of the tunnel and the front group jumped hard and dropped me and the other woman completely in the dust… so we finished 10th and 11th! I did over 300W at that point, but it was no where near enough to hang on!
The 53kg guy ended up finishing second in the sprint behind the curious “sandbagger” (who may not be human? orange jersey on steel bike with no HR).

So, in short, it was fine, and I think this latest version is a reasonable mirror of the old cat system with some significant improvements. My problems are really just the same as ever problem of mixed races - they make me feel like we women are a different species! Women’s B cat racing - where I spend more time - makes a lot more sense!


i just tested it by putting my own numbers in and my CP rises when i lower my 3 min wattage and leave my 20 min as it is, i’m presuming the algorithm thinks i’m more aerobically developed relative to my actual ability if my power curve is flatter. i’m not using the highnorth spreadsheet because this thing doesn’t have excel on it, i just went to a 2 curve calc from the front page of google

396w 3min
270w cp20
CP 248

345w 3min (artificially lowered)
270w cp20
CP 257

is it punishing riders who have a flatter power curve?

i dont use CP in my training so i don’t fully understand the system, but this seems super weird to me.

i understand enough that the MAP component is designed to put a lid on it to some extent (for sure, i am in A regardless of my CP, though i wonder if that would still be true if i did not intentionally do a lot of max testing at sub 10min intervals on zwift), but given that zwift races already typically are in favour of those with an exceptionally strong minute, isn’t this just just going to punish people with a low anaerobic capacity even more? i think double draft is hiding some gaping holes here

1 Like

We’d also noticed this. It may seem initially counterintuitive but it’s natural enough that increasing your max power on a short time interval can tend to steepen the curve and therefore actually drop the RHS of the fitted power curve where it’s extrapolated beyond the longer time interval data point.

The problem is not so much in the CP calculation but rather in thinking that CP has any direct relevance to zwift racing. It just doesn’t. Its only relevance is indirectly, in what it generally means (but rather approximately) for shorter-term efforts. I think I mentioned I did about 3.1 for my B cat test race a couple of days ago - and a couple of guys who beat me were lower than that. I hardly pedalled a single stroke at threshold, it was either Z1/2 or flat out. The race was entirely decided on the two short climbs and slightly longer drag to the finish (Richmond UCI). Finish order was much more highly correlated with 1 min power than 20 min.


Do you think that is an observation for a specific course (Richmond) and races on other courses will correlate better with a different part of the curve?
How to apply this to all of Zwift?

i was originally going to post this the other day to point out how stupid double draft racing is, but i guess it works here. the entire A field from a pen enforcement race including the few people who were dropped:

2nd lightest guy in the race (me) wins on a very flat course (seaside sprint) with the 2nd lowest 20min w/kg, only a guy weighing 25kg more than me managed .1wkg less for 20min. not because i am better at drafting than everyone else, i’m not, but because i did the bare minimum while everyone else did their best to create gaps and drop people. and they would have certainly succeeded in a regular draft race

i think double draft is probably covering up some big faults in this cp category model. i guess they’ll become apparent if community races start adopting it without DD enabled


I’m sure it will be a bit different for each course, usually the critical element will be anything from 15s (final sprint) to 3 mins for the typical climb where selections are made. But I wouldn’t want any magic formula applied.


Interesting use of a CP calculator and I believe supports what I have read elsewhere. ( There is so much to read how does one know what is right or not ? )

I’m sure you have done all your own reading but if you have not read this Critical Power for cyclists explained | Better than FTP? - BikeRadar as I think it makes easy to follow reading.

In summary it says: ( none of these are my words)

CP is the power output you’ll tend towards when riding at a high intensity, typically for around 30 to 40 minutes.

Similar to FTP, Critical Power provides an indication of your maximum sustainable power over an extended period of time.

CP will generally be slightly higher than FTP (studies suggest around 9 per cent higher on average, although in my experience as a coach it’s typically slightly less, around 4 to 5 per cent) but both can be interpreted in a similar way.

W’ represents your capacity to ride above your CP. Having a larger W’ is most important for short events performed above the Critical Power (such as short hill climbs), and for punchy disciplines such as criterium, [cyclocross] and cross-country mountain bike racing, which require frequent efforts above CP within a longer race.

While more information can be gleaned from a CP test than an FTP test, it still can’t tell you everything. For example, if your W’ goes up, it can be unclear whether this is due to an increase in your ability to produce energy anaerobically (without oxygen), or whether it represents an improvement in your aerobic fitness (e.g. [VO2 max]

However, if you build your W’ too far, then this can come at the expense of your endurance and threshold power. Thus, only cyclists competing in short events such as track sprints or short hill climbs will want to build their W’ much beyond the ranges above.( an optimum W’ would be around 16 to 20kJ. Very well-trained male cyclists do well with W’ values in the low to mid 20s. For females, optimal W’ values are around 11 to 15kJ for moderately trained and 14 to 18kJ for well-trained cyclists. )

1 Like

JamesA_ZSUNR, in my experience the races with people pushing this level of power are very hard.
I raced in Cat B for most of my 4 years in Zwift, so I had many races and events to gain some experience and knowledge.

My 20min power is 260-285W depending on my training and form. Weight between 65-70kg.
In some flattish races, were there are lots of “heavy hitters”, i have to draft all the time, and really struggle to stay in the front. My only chance is to be good in drafting, and close gaps pretty quick.

So if ZHQ puts a 200W (CP now, “FTP” in the old system) floor in Cat B which is a second tier Cat, basically it´s saying that if you can´t push at least 200W FTP, you don´t stand a chance in many of the regular races routes (most are flattish).

Wouldn´t it be logical that ZHQ reflects on the fact that if you can put 340W (exemple) you are putting such a “stress” in the people with low raw power, that basically forces the low power guys to ride at close to 90-95% FTP (or more???) all the race.
How are guys that ride close to FTP, be able to respond to surges and attacks? They don´t and get dropped like stones.

And the same applies to the new “MAP” value. We know that a guy like Wiggins could do 570-580W for a little more than 4 minutes in Athens Olympics (his own words). But in Zwift we can have a guy with 490-500W of 5 min power, and he can still be a Cat B rider? He only have to “play” with his weight.
And let´s remember again that Cat B is only a 2nd tier Cat …

There are physiological limitations to these metrics. So only looking to W/kg and “forgetting” the raw values creates a schizophrenic Cat system.

1 Like

340w for a 90kg guy in the B cat will also be 90-95% of FTP.

1 Like

If you have 3 or 4 guys with that power - and with the churning effect - these guys don´t need to hammer the full time.

Gerrie, think this way: i would bet with you that if you select a group of people with some cycling background, only a low % of those riders would have a FTP of 340W.

I would bet that a much higher % of people can hit 4W/kg

There is a point were you start to have a disconnect between more weight giving you the chance to have more power. There are physiological limiters at play.

Cat B is 3.2 to 4w/kg so for a 85kg rider that will be 272w to 340w FTP.

So for a 85kg rider to be the top rider in B he has to have a 340w FTP.

Yes for a 60kg rider it should be easy to hit 4w/kg or 240w FTP.

This is why it is hard for a 60kg rider to compete on a flat route and a 85kg rider on a hilly route.

What need to be tested is what does it take for a 60kg rider to stay in a pack and compare that to what it take for the bigger guys to drive the pace of the pack, because if they don’t drive the pace it wont be an issue for the lighter riders.

1 Like

Zwift could test it using a ride with all virtual pace bots of differing weights and sizes to see what would be ideal. Pace bots don’t get tired and put out consistent watts. They would be the best testers ever.

Erm… okay. :see_no_evil:


I may have been a bit optimistic. Sorry @Dave_ZPCMR . :rofl:

With some tests we could get some more reasonable numbers. Maybe the current ones is good. We need something to feel like real life.

I don’t want to sound like I don’t care about the light weight riders. :ride_on:

1 Like