Racing against others on the Zwift virtual cycling platform is a fascinatingly unique combination of video games, esports, exercise, and not a small amount of pain. In part one of checking out the Zwift racing landscape heading into 2022, we took a look at some numbers around the total racing population, how it split between countries and categories, and what the most popular events are.
Before we dive deeper into the ranking algorithm and look at how well the ZwiftPower Race Ranking numbers can be used to stack rank the entire racing population, there were a lot of questions around additional demographic stats of Zwift riders. In the below charts, I break down the ranked racing population by gender, age, and weight class – split out by race ranking (which is roughly tied to categories).
[adinserter block=”1″]
As quick context, I wanted to further illustrate that we actually only have good data for a small portion of all Zwifters – those that both compete in races, have registered a ZwiftPower account, and scored points in at least one race in the last 90 days. These riders are naturally going to be biased toward stronger riders – those who care more about race outcomes and are interested in weeding out cheaters/sandbaggers.
We can use the below demographics (and stats from the initial article) as rough approximations for the overall Zwift population (especially for things likely not impacted by rider strength, like country), but most stats related to power output will likely change significantly as the scope expands. In the above bubbles, I estimated the total population and average power output via some random sampling just to get a sense of how much it varies vs. the population we have good data on.
With that said, let’s take a look at the data we do have:
Gender gets less male-dominated as you move through the categories, but overall the population of ranked women is just quite small: ~6.5k total women who have a ranked result on ZwiftPower in the last 90 days (out of ~80k total riders).
When looking at these next charts with Age & Weight data, it’s worth remembering that most (65%) ranked racers have a ranking between 500-600 (the last bar).
Interestingly, in a situation that might be unique in competitive sports, and definitely in esports, most ranked racers on Zwift are solidly middle-aged (40-49) (34%) with another 28% aged 30-39 (especially as you move toward the higher categories). 50-59-year-olds make up the 3rd largest segment with 22% of the total population. There is a chunk of 20-29-year-olds at the top end of the ranked population. while the overall population of U23s and 60+ riders is fairly small (~5k total).
Weight classes are fairly stable across ranks, except for the most competitive ranks where lighter riders (60-69kg) are overrepresented.
As a frequent Zwift racer, I’m always curious about who the humans are on the other side of the avatars, in as much pain as I am. Based on the above stats, it’s probably no surprise that the team with the most riders on ZwiftPower is DIRT (e.g. Dads Inside Riding Trainers).
Questions or Comments?
Share below!
Interesting article…Thanks! I’d be curious what you think of the current category system and what changes you’d make to it given the insights of your 2 articles. It’s also curious to me that almost half the racing population is unranked (they all sit at 600 points), would broadening the ranking system draw more people into racing and grow this part of the platform?
Thanks David – right now the bar to scoring points / becoming ranked is logistically a bit of a challenge – registering for a 3rd party site – zwiftpower, making sure you have a HR, etc. Obviously lowering the hurdles there would grow that part of the platform. Overall, though, the most important way to grow the racing population is just to make sure racing is FUN – fill the top of the funnel by just drawing more and more people in with new formats / experiments. It might be a controversial opinion, but I don’t think most people race to win.
That’s kind of a disappointing look at the racing demographics. Zwift’s best feature, for me, is the approachability of racing for the non-trad crowd, but it seems I might be alone in that assessment.
Older people cheat the most😁
What random sampling did you use to determine the base line Watopia numbers? You just say “random sampling of riders encountered”, which sounds like you were riding around and looking at numbers. If so, this method is heavily biased to discover riders that are slower than you, and will not give you a uniform sampling of riders. To see why, imagine you are on a 1 mile course and there are 3 other riders. You are traveling at 10 mph. One other rider, who is half a lap ahead of you is also traveling at 10 mph. Two riders start next to you, one at 5 mph, and one at 11 mph. You ride for one hour and record everyone you see and how many watts they are doing. You will pass the slow rider 5 times over the course of the hour. You will be passed by the rider going slightly faster than you exactly at the finish line (so once), and you will never see the rider going the same speed as you. So from your perspective 5/6 of the riders you see are going half your speed while 1/6 of the riders you see are going about your speed, but in reality, 2/3 of the riders other than you are going about your speed.
Another version is to imagine 100 riders on a long stretch of road, half are slower than you, half are the same speed. You won’t pass any rider going the same speed as you, but you’ll eventually pass all the riders going slower than you, so you’ll think most riders are slow.
Its definitely a fair point – was highly unscientific. There just isnt really great data out there on it. That said, I have a strong hypothesis that range isn’t that far off from reality – but agreed that we dont have the data to really back it up right now
I’m always amazed how many racers have entered weights near or below 100lbs. I’ll often come in the back half of the pack in a race, but be top 10 in total watts. I’m not that heavy but I do use my actual weight. Real life rides are more competitive for me.
That is egregious! Yeah, a great study of the Zwift Racing Landscape would be how many use an accurate weight (within +- 2.5lbs I’d say is fair to account for empty/full variation).
I’m honest about my weight – measured at some point in the last week, naked, in the morning, after a big poo. For an evening race after I’ve taken on a load of water in the afternoon in anticipation of an hours sweating, it might not be so accurate.
Would be interested in the age and weight breakdown separately for men and women racers if you have that information.