On two wheels

I did an FTP test and it turns out I’m spot-on the average middle-aged man in lycra

I did an FTP test to determine my current performance level. It's rather average. But that's a fuzzy term.

I realized I haven’t been blogging about anything bike-related for quite a while. As the season came to an end, there was less to talk about. I’ve done some late rides in the November sun, but now I have returned to Zwift (and I’m lucky enough to still pay the less expensive old price). So here we go.

Last week, I did two quick workouts to get used to cycling indoors again. Zwift is a virtual world in which my avatar cycles together with people from around the planet. It’s a great training tool. Allegedly, Matthew Hayman owes his Paris-Roubaix victory of 2016 to his virtual training that he did whilst he could not ride on the road – he was recovering from a shoulder fracture.

Zwift is also a fascinating marketing case, but they seem not to have understood their in-game advertising potential yet. (In the game, one can ride real bikes, and bike choice has an influence on speed. Imagine you ride on a Trek and you notice the virtual advantages. Next time you buy a bike in real life, will that have changed how you see Trek as an option? And how much would I pay to bike my favorite bike in-game?)

Those two initial workouts led up to an FTP test on Sunday. For cyclists, that’s the much criticized standard to determine how challenging a workout should be. Basically, it measures what power one could do over an hour with a short 20min time trial, embedded in a longer workout that contains some warming up and cooling down. The resulting value, in watt, is then divided by weight. Watt per kilogram can be compared across riders like horse powers of a car.

I had expected my FTP (functional threshold power) to be a some bit lower as compared to my last test in March. Back then, I ended the trainer season and went out on the road. I had a great start, besting virtually all my personal records around Milan. However, in those last rides of the season, my form had dropped, and I had found myself in a post-season dip in which my fitness (that’s the orange line in the next graph) had fallen from 100 back to 50, and more importantly, my form had started to fall without adding fatigue as well. * And that’s okay. It’s the end of the season. Some extended rest for the body is a good thing.


* Form, fitness, and fatigue are one established method to monitor training progress and training stress. Fatigue is the cumulative training stress (as a measure of workload) over a period of 7 days. Fitness is the cumulative training stress over a period of 6 weeks. Form is the difference between the two. It should have a slightly negative value in training, and it should be (slightly) positive (and increasing) ahead of a multi-day event. A too negative value indicates over-training, and a too positive value indicates an extended break that will likely result in a loss of power.

Ignoring expectations, I performed surprisingly well in that FTP test. I started with a power output that was just a few watt above what I did during my last FTP test. My heart rate kicked up rather fast, and rather high. This is a 20 minute test, and supposedly one should reach near-maximum heart-rates only towards the end of it. Instead, my heart rate exceeded 180 only 3min into the effort, and then never went back. I’m a bit skeptical about it. I don’t think anyone would be capable of doing 17 minutes with that high heart rate, but without major respiration issues. I was breathing stronger, but more like I would with a heart rate around 170. I suspect that there might be a measurement issue.


Anyways, I finished the effort with an average power output of 211w. You multiply this with .05 to estimate your functional threshold power. Zwift rounded it up to 201w. Just above the magic 200. Four watts more than in spring. I had actually improved myself.

I then compared this performance to the average middle-aged man in lycra. Turns out that I am that average middle-aged man in lycra. Or at least almost. My FTP is some 30w smaller than the average, but I’m also noticeably less heavy than the average (67-68kg vs. 78-79kg), so adjusted by weight, my FTP is that of the average guy or girl (doing indoor training).

Of course, there are a number of obvious limitations to that claim. That population I compare myself to includes a huge number of users that just never tested their FTP and use the default value of 200. It also includes both men and women, and a gender-specific distribution might look different. And, obviously, it’s only users of a dedicated trainer platform and excludes all those cyclists who do spinning in a gym or don’t give a shit about winter training. And that’s not talking about all the arguments against the relevance of FTP as a single metric (which, occasionally, can be a rather heated discussion).

It does, however, add a different perspective to another very commonly cited graph in power-based training. That graph plots the typical w/kg of the many tiers in American amateur cycling (and international pro cycling). When I started power-based winter training last year, I was utterly devastated when my first FTP test classified me as novice – after all, I’ve been riding for ten years! Meanwhile, I made it to (the bottom of) category 4. *


* The idea of that table is to understand an individual rider’s power profile. It contains the power one can deliver over very short, medium-short, and long duration. A sprinter will typically perform better on short efforts, while a time-trialist will see the opposite pattern. The majority of people have a flat profile, though. Personally, my power profile points at time trials. Not a big surprise. Without any power analysis, I also know that I suck at sprints, but that have a well-developed Diesel engine. Once I’ve settled into my rhythm, I just don’t get tired easily.


So, in a comparison with fellow indoor cyclists, I’m very much average. In a comparison with amateur race cyclists, I appear to be below average. However, population sizes in these race categories are obviously not even. For instance, there are about 1,500 cyclists (worldwide) with a professional license – that includes domestic pros. Being a shitty category 4 racer – well, yes, that is the average.

Next, let’s speak about power-based race categories in Zwift. You can do races in Zwift, and you can select against whom you want to race based on your FTP. The threshold values are not the same as in Andrew Coggan’s table. They are markedly lower. To make it into Zwift’s highest race category, it’s enough to be an average cat-3 racer. I’m currently a mid-level C racer in Zwift, but with good training, I might make it into the B field by the end of this trainer season.

At the same time, that’s why these static values are utterly useless. Consider that I am not the only C racer in Zwift. If I train well, it is fair to assume that so does everybody. As a consequence, with a few years progress, higher race categories with static boundaries in Zwift should swell up (age-related power degradation will push some people back into lower categories). If they don’t, indoor training is proven to be useless. (Or at least less useful than what we thought.) But if they do, racing becomes a more frustrating experience for many. In a Facebook group of Zwift racers, I’ve advocated for dynamic boundaries in the past. As we gather all the data, it should be easy to adjust the category limits every season in order to simply divide the population in four even-sized chunks. Then, those who move up one category are those who have earned it by training better than their peers. Sounds fair to me.


  1. The whole race category thing in Zwift is a toughie. I’m not entirely sure what the solution is – or if t here even is one.

    I actually prefer the races where all the Cat levels start together or where it’s just one open category. As a very ordinary mid B-packer I’ll never be winning any races, but I love beasting myself to try and stick with the faster racers off the line. Trying to draft the big guns for as long as possible to open up a gap is an awesome workout!


    1. I agree that there’s lots of value in open category races. I do think that performance-based categories can be helpful as well. For instance, I like races where the weakest category starts first and has to defend its lead. To that end, category limits should be updated monthly based on the FTP distribution of the user base. However, all of this is not compatible with the CVR Zwift rankings. Which is another big thing to debate about for hours.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: