Occasionally, I participate in the Cyclenation Corkscrew Chase at Tempus Fugit Watopia. A main group sets off at 1.8W/kg, and the chasers wait 4 minutes and then try to catch up as quickly as possible. Last time, the chase group was over 50 people. Average speed of the chasers was 40-41 km/h. During the chase, I ride at approximately 235-240W (86 kg, 2.8W/kg). I recently showed these figures to an amateur cyclist who competes in real races. He claimed the data isn’t accurate. In a peloton of 50 people (professional positioning), it would be closer to 150W. Unless I’m constantly riding at the front, which isn’t the case. I found the difference significant and submitted the question in ChatGPT. My question was: “How much power in Watts required for a cyclist 86kg, 1.8m tall in the center of a 50 cyclists peloton to average 40km/h?” The answer (partially):
~80–85% reduction in aero drag
3. Power calculation in the peloton center
Assume:
Solo aero power at 40 km/h: ~290 W (mid-range)
85% drag reduction → remaining aero power ≈ 45 W
Rolling resistance: ~30 W
Drivetrain losses: ~5 W
Total required power:
4. Practical range
Accounting for turbulence, micro-positioning, wind variability, and small accelerations:
Lower bound (very sheltered): ~60 W
Typical center position: 70–100 W
Less optimal shelter: ~120 W
5.Final answer
For an 86 kg, 1.80 m cyclist in the center of a 50-rider peloton averaging 40 km/h on flat terrain:
This explains why professional riders can sit comfortably mid-peloton at speeds that would require Threshold-level power when riding solo.
if you say you were averaging 2.8 w/kg then i’d expect the front to be doing quite a bit more maybe 3.5 w/kg or more.
The issue with the wasted watts window is if its irl it would be impossible as you would crash where as on zwift the game just restricts your movement but does mean you can put out unnecessary power
I doubt that I would. As long as there’s a detectable difference between being in the draft and being in the wind, it makes little difference to me if the draft savings are too low, too high, or exactly “accurate”
It’s not something I worry about as I whizz down Alpe du Zwift, waving at the Yeti, as I round the hairpins at 75kmh.
Accuracy in data was my profession. Call it a deviation, if you like. If you want to simulate something, you have to try to approximate reality as accurately as possible. That, in my opinion, is Zwift’s goal: to achieve the most realistic approximation possible. Feedback is therefore essential. It’s not criticism.
Yep. The event starts at 6 PM Brussels time. Exactly 1 hour and 45 minutes from now. You’re invited. You’ll be standing still for 4 or 8 minutes at 0 watts, so please factor that into the data
Pretty sure that Zwift’s primary goal is for people to buy their hardware, subscribe to use their product, and then resubscribe when that first subscription ends.
I expect that making people happy so they become unpaid marketeers would also be up there.
But what do I know? Maybe 90% of the people who don’t renew their subscriptions do so because of inadequate draft? That seems unlikely, though.
Everyone wants you to buy their product. That’s just how the economy works. But I want to start a discussion that can lead to a better product because designers are listening. That’s also the purpose of a forum: to question and improve where possible. That’s called progress.
And I’m not trying to shut down discussion or stop progress. I’m just suggesting, hopefully in a polite way, that your initial assumption - that you’re still holding on to …
Zwift drafting / pack dynamics being more like real world cycling is necessarily (and without exploration of the consequences) better
… isn’t necessarily true. Online cycling isn’t real life. It’s a different beast. Maybe having such a massive drop-off in power requirements when in the draft just feels “wrong” in that context. Or, put another way, a massive increase in power requirement when losing the draft. I don’t know.
What I do know is that Zwift has been around for a while now, and if they wanted to I’m sure they could have implemented a version with more accurate physics. It would be interesting to hear from them as to why we are where we are. We can definitely agree on that.
If we are talking Zwift races, I’m sure I remember reading that some people wanted pack speeds slowed down to Make Racing More Exciting™ and allow breakaways to succeed more easily.
So we arrived at Pack Dynamics changes over several iterations.
The thing with Zwift is average speeds are too fast on flatter courses, and it’s to do with many things, you take corners at excessive speeds without slowing down and everyone just rides everything like a time trial (did that Make Racing More Exciting™?) with no risk of crashing into other riders as in real life.
Aside from that, just in general folks like speed, people want virtual distance on their Strava. Like the one who rides 8 hours per day and 2500km/week.
Perhaps limiting speed to what’s physically possible in downhill corners is a better idea. Then you’ll have to accelerate after each corner to get up to top speed, and the whole thing starts to feel a bit more realistic.
