Zwift Pack Dynamics - Physics fudging/drafting

IMO when people complain of unrealistic drafting in Zwift, it’s not so much the general drafting algorithm that they’re responding to, but some ways in which it seems Zwift may be breaking/cheating the full physics model to make packs look more realistic.

Specifically, the lack of collision detection (i.e., riders are allowed to occupy the same space) in Zwift would mean an accurate physics model would create a very unrealistic-looking pack - basically, everyone would be squished together on the 2nd row of the pack, briefly rotating to the front before getting kicked back to the 2nd row.

In order to stretch the pack out, I believe Zwift is fudging the physics to keep riders moving up/back.

If e-racing is to become a real thing, I feel Zwift needs to be transparent about their model - they need to publicly document all of the ways their model works, including these fudging cases (sticky draft being one of the very obvious physics “fudges.”)

My hypothesis on Zwift pack dynamics is that there is a bias being applied to riders to keep them moving through the pack.

I’m not certain about this but I feel that once you hit the front, you get penalized until you hit the back of your group. I’ve observed that I’ll often drop far back in the group even while maintaining the same power level that brought me to the front.

In order to save energy, I’ve taken to letting myself hit the back of a group and then maintain just enough power to very slowly drift toward the front. It seems much easier to move up than to maintain position at the front.

For people who’re used to racing in real life, this is completely counter-intuitive. Your instinct is to try to keep yourself up front - you never want to give up positions if you can avoid it. Zwift’s algorithms, I believe, are actively punishing you for this - because you don’t move toward the back of the pack, Zwift’s bias stays in effect, forcing you to maintain a high power level to keep your position in the pack.

IMO, this is a case where Zwift may have taken the wrong path: It makes for a realistic-looking visual image of the pack, but is unrealistic from a perspective of race tactics - diminishing its value greatly for people who want to race online and offline.

In any case - I may be wrong about my theory, but there’s definitely some “black magic” in Zwift’s algorithms they haven’t revealed. I feel they owe it to the racing community to be open about this (and to give the community an opportunity to give feedback).


I don’t know mate, but I wouldn’t expect too much in this direction - intellectual property and all…


I dont think Zwift even know what they’ve done…but thats an interesting angle to this discussion:

1 Like

I’ve been paying attention to the group drafting the last month or two and I’d say that my experience was the opposite of yours. In the AHDR rides (B cat group rides with around 200-300 riders) I have found that riding in 50-100 place is usually the easiest. Here you are not too close to the front while still being surrounded by a lot of riders. 200w is usually enough to stay in this position with an occassional increase to move up if you find yourself drifting back.

If you do end up drifting to the back of the pack there are less riders around you and I need to be doing 240w (3w/kg) just to hang on. Doing 200w at the back will get me dropped. If I want to move up back to the safety of the middle of the pack I need be doing 260-280w.

It’s like Zwift ignores the 300 riders ahead of you that should be creating a massive draft but only count the number of riders in your immediate vicinity. So if you’re at the back where things are a bit thinner and you’re all struggling in single file to hang on there is less draft than if you were up the front surrounded by people riding 10 wide on the road.

I agree - in my effort to keep my post short, I didn’t post all of my observations.

I think there’s another phenomenon at the back of larger packs - probably induced by the sticky draft. The sticky draft makes it very easy to get tailed off the back of the pack when weak riders drop. Basically, you get “glued” to the rider ahead of you as they drop off the front, even if you’re doing enough power to stay in the draft.

While this is annoying, this actually is semi-realistic, because irl you’ll have a similar effect when the weak rider whose wheel you’re riding gets detached. You have to jump around them in the wind to get back on. I still think it’s easier to get tailed off the back of a Zwift group than a real-life group, but it’s not as contrary to real riding (and I do find, for example, that tail-surfing a small group is a way to ride pretty efficiently - you only have to avoid the back if people are getting dropped).

In Zwift, though, you’ll find that once you get back in the pack (which you should do with an aggressive jump - i.e., big burst of power), you can dial the power way back and still move up. I think you’ll find that your 260-280 is not necessary to move up. As long as you’re patient, once you’re in the safety of the thick of the pack (not in the back where people are getting tailed off) 210w might do just as well - you won’t move up quickly, but you’ll move up steadily.

I think in addition to the effect I mentioned in my OP, there is also some w/kg ranking that is used to determine your position in the pack - i.e., you’ll pass people who’re doing a lower w/kg than you but will get passed by people doing higher w/kg. This tends to predominate in very large packs, where the behavior I mentioned is more evident to me in smaller packs (up to 20-30 riders). It’s even possible that Zwift uses different algorithms for those situations - again, something I’d like them to make transparent.

Fair enough. But I don’t think Zwift should expect people to become enthusiastic about e-racing if they won’t do this (Particularly since they have the ability to change the “rules” - i.e., the physics - with any release they do)


Yes, in smaller groups (10-20) it’s easier to sit on the tail (making sure you don’t get stuck behind someone who drops off).

I do think in larger groups (100+) that having riders beside you makes it easier. I would be surprised if Zwift was calculating a draft based on 100+ riders though. Maybe it’s only the nearest riders (maybe only those within 20-30 metres perhaps, it’s hard to know without testing with 100 bots) that are contributing to your draft and if those riders are strung out in single file you get less draft than if they were in a bunch covering the whole road.

But any change in the physics will apply to all riders in the race, so even if it doesn’t simulate real world racing it will still be Zwift racing.

E-racing does not have to try and simulate IRL it is a discipline on its own.

Like track and pump is different than TDF.

They should still be transparent about those rules though. Zwift already run threes types of draft - standard, double and none. If things are being tweaked the community should be told, if the draft is being dialled down then it could make breakaways more successful. Are riders/racers just supposed to guess?

(Hey, Ullrich! Good to see you here…)

The physics in Zwift is definitely strange. Consider the “swarm of bees” – Zwift Insider uses ANT+ simulators to have a group of four identical riders all at the same power, for example 300W. IRL the only way for this to work without crashing is for the riders to be side-by-side. But there’s no crash detection in Zwift. So the Zwift version would be the riders riding on top of each other. But this is an instability, because of drafting changing suddenly, and stickiness and data lag. For whatever reason, it results in riders one-by-one launching off the front at high speed, then fading back, the “swarm of bees”. This is completely non-physical: IRL a group in a line can only go as fast as a solo rider riding at the power of the rider working the hardest (I’m ignoring the draft advantage of having a rider following you, which is significant but relatively smaller than the draft from following someone else). In Zwift a group of riders will go faster than the rider producing the maximum power as long as everyone is sufficiently close in power. “Stickiness” adds another complication. I think RGT does a better job of physics by requiring a power penalty for moving up in the pack. This helps make pack position more important. It additionally keeps track of rider lateral position (ignored by Zwift except with the steering lanes). This means you need to move into the wind to move up on the side.

It’s a cop-out to say “Zwift isn’t a real-life simulation”. It clearly aspires to that, and RGT does a better job of it. But given the flawed physics, riders need to learn how to play the game, something I’m still doing.

1 Like

RGT also has a better collision detection algorithm.

1 Like

I’ve experienced the same thing. It’s so different that IRL. I’ve been racing on Zwift since almost day one, and I think this effect of forcing you to the front, then blowing you to the back is something relatively new they implemented. If all races were double draft (real draft) I think that would help, because you could at least rest in the middle, just like a real race. I find it extremely hard to stay 10 positions back without drifting to the front and then having to almost sprint to regain a draft. The other thing that is missing in zwift is a visual cue how much draft effect you are getting. Why can’t they just say -10 watts or a bar graph…???

I signed up for RGT, but nobody seems to be in the events. This was a while ago. Are more people on it now?
I liked it a lot, better that zwift.

The community is growing, and now there are rides and races every day at different hours.

RGT racing and drafting is by the way simply on another level vs the rest.


Yeah there’s a lot of good events every day. Not massive numbers lien Zwift. You won’t find 1,000 people on a group ride but the racing is much more fun and dynamic.

1 Like

@Matt_von_Wahlde: the draft numbers in Zwift are realistic for a single paceline. They’re not realistic for a giant arrowhead of riders. So it’s not that “double draft is more realistic”. It’s that it doesn’t handle the cumulative effect of having a lot of riders ahead, to the side, and even behind, in close proximity. Additionally, the draft in a “century ride” (or “gran fondo”) isn’t the same as one in a World Tour race, where in the latter case riders are literally shoulder to shoulder, while in the former case there’s huge gaps between riders who lack both skills and mutual trust.

That makes sense. The lack of “pack effect” is definitely a real issue in terms of realism and the feeling of immersion for me. That and the fact that positions are not absolute, and other riders see different things. I only do races on Zwift, so my muscle memory kicks in and it just doesn’t feel even remotely like RL

just think about the WCS on zwift and how many riders have been participating. not much. guess why … this game is far from beeing real. and if you race IRL you will want a game which is real. the drafting logic is far from beeing real… the w/kg logic in the flat is far from beeing real… like other stuff … :tired_face: :pray:

1 Like

Around 10% watts saved if its a line/column of riders and up to 15% if it is a bigger/wide peloton.

Yes, and that’s why there is not any possibility of a breakaway in Zwift.

1 Like

Absolutely true