Robo pacers don't provide draft

just a few minutes after the ride in makuri i joined constance in watopia. there was one more rider with Constance so compared with the previous ride i was drafting a real person for the whole duration of the ride.

drafting a real person: 276 watts (volcano circuit)
last comparable ride: 266 watts (volcano circuit, 2023 jan 6)

drafting constance: 328 watts (country to coastal)
last comparable ride: 268 watts (country to coastal, 2023 jan 3)

this screenshot pretty much sums up the experience during the ride:

I’ve experienced the same as Marek with pace bots, a definite change in the past week. I’m very familiar with the dynamic pacing and expect to work harder uphill, but the effort required to hold the wheel of the bot on all gradients is now higher. I don’t have as accurate numbers as Marek to share, but I’d also estimate around 15-20% more effort. Less noticeable when there is a big pack, but when only a couple of riders following, it’s a very significant difference.

1 Like

Steven sums it all up as above.

Is a Zwift developer or engineer not able to say if there have been any recent changes to the Robopacers in terms of draft mechanics or algorithms?


I got dropped by Yumi after 10 mins on Tick Tock (normally do 60-90 min rides without issue). Not been riding much recently so thought it was just me; maybe it was :thinking:


I tried today but struggled, I had to keep doing 7.0w/kg to stay with the 4.2-4.6w/kg bot. 5km of that was enough. There were 5 riders with the robo-pacer. I went back to Coco group on Tempus.

In the old days of Anquetil I used to do 70km rides with that group on the Epic KOM and radio tower climb no bother at all, it was quite a lot more enjoyable back then. It’s not aimed towards lighter riders these days.

1 Like

Big thanks to everyone for the additional reports, it’s really appreciated! Hope it will make it easier for the team at Zwift to troubleshoot this. Even if it might be difficult to link this to a specific change done by developers, comparing the data from recent rides with historical values should clearly show if there’s a difference as significant as 20% or more.

@Chris_Denton_LACC just to add to what you said, this latest change affected me too, so it’s not only about lighter riders. I’m heavier than the pace partner but now i struggle all the way - uphill, on flats, and even on downhills. In makuri today i had to stay between 4.0-4.9 on the downhill section to stay in touch with Constance. I couldn’t do a proper screen recording, but i tried to take a short video using my phone (sorry for the poor quality).

Yeah, I know (I sound like a broken record), but I feel there’s not much at all on Zwift aimed towards lighter riders (with mediocre talent = me) unless you can sustain impessive raw watts and W/kg. I’m feeling increasingly out of my depth racing as a cat- C and, TBH, I’m not really missing it after a 2+ month hiatus.

I’ve also experienced similar. With Maria on Volcano Flat last week I had to surge pretty hard on the short climb up into the volcano just to stay with the bot, noticeably harder than it should be. But then yesterday with Maria a similar surge on the climb out of the desert on Tick Tock put me way off the front. Pack size was close to the same for both.

What kind of watts was that? Cadence is 75kg, so at 4.2 that’s 315W, and on flat that is what will matter most. What kind of wattage were you doing on flat to keep up? Did it feel like there simply was no draft with Constance?

If they had the previous weight of Antequil with dynamic pacing the swings in wattage then hills would be near impossible for heavier riders, so I think it’s a good thing they standardized on the weights of the bots so moving up or down a bot should be more consistent gaps in effort as a result.

My last pacer ride was Yumi for 90 mins, and my wattage was 199W for the ride, so actually less than I was expecting, but then again it was a large group, so drafting the bot itself was not much of a consideration.

I was having to push well over 400w at some points, (no TT bike, before someone drones about that). As soon as I’d drop to 3.8 or 4.2w/kg I’d start to get dropped again then have to push big watts again.

As for Anquetil - uphill watts would have been 300w at 4.6w/kg if I calculated right. Not excessively hard. Yes she rode routes that were hard for heavier riders, but that’s how it should be with an A grade bot.

Now it’s just crazy - there is no point for me to even try to ride in that group anymore. So for a harder ride I ride alone on Epic KOM or use competitor product with real world hills in it.

Ok, so instead of doing a second tour de wattopia ride today I decided to spend 22mins with Constance on Volcano Circuit. I’ve never ridden with the red bot before because I use workouts and races for my higher zone work, but this was for science.

I think the group I was with varied between 2 and 4 people, and it was rare that someone was trying to push the pace ahead of Constance.

For the 22 mins my average power was 271W normalized to 279W (on My weight at this moment is just a bit heavier than the bot at about 75.3kg.

My biggest issue with the bot is I had a super hard time keeping power constant and staying with the bot, I was always surging and overshooting etc, I felt I needed to surge in odd locations, not just on hills, so I probably need more time practicing with small pace bot groups.

Here you can see my averages and also how chaotic my power was for the ride.

thanks @SeattleSauve! from the screenshot you shared it looks very similar to my experience:

  • being the first person behind the bot, you get very little draft and have to do almost equal wkg as the bot - in your case it’s 4.1 wkg

  • the people behind you benefit from your draft just as before and can keep up with the group at 3.4-3.8 wkg

If you don’t surge you quickly end up behind the group and then have to put a big power output to get back to the group. You can end up in a bit of a roller coaster ride of power up and down.

On Tempus it is a bit easier because the power is constant, aside from the small downhills where the speeds go up massively - usually with the people at the front rolling fast (they are probably all heavier than me).

@James_Zwift can you please confirm that there has been no intentional change in how the pace partners behave?

also, i understand that this is just a minor issue and that it only affects a very small amount of riders, but is there any chance that someone at zwift will look into these reports to see if there’s a problem or not?

should i report it through customer support, or is this forum the best place to solve issues like this? thanks!

I haven’t made any changes to the Pace Partners and noone else would have done so.

I don’t know if there is actually a problem, or if it is just different rider behaviour on different rides causing the different riding feelings.

Support wouldn’t be able to help.

@James_Zwift, is there a specific way i can help you to test if there’s a problem?

so far i could only provide numbers from comparable rides from the past - it’s in this post.

in the same post there’s also a screenshot where you can see that the first rider behind the bot has no draft and the second rider has usual draft.

i also posted a short video showing my pace while being alone behind the bot.

if you need anything more please let me know, i can provide links to specific activities or try toa take a longer and better controlled video of a ride…

as i said i’ve been riding with constance almost every day and very rarely was it less than 100km. now i can barely finish 50km if i’m alone with her and the difference is 30-50 watts.

i double checked with most of the regulars too, everyone struggles to keep up and find it significantly harder now than before. and i’m not comparing constance with amelia, it has all changed from last friday.

1 Like

Sorry, I’m having a bit of a week here :joy:

I’m about to head off on hols for a couple of weeks and once things have calmed down a bit I’m happy to try and look at a bit more depth.

I suspect that when we roll out PD4 (assuming that we do) this will probably help to stabilise things but that’s yet to be seen.


Okay, thanks. Then we put this on hold and wait.

I’ve never heard about them.

I averaged about 152 watts with Maria on Tick Tock today. Standard.

Then I tried it on a gravel bike. 190 watts. Did not like. That’s 10 watts more than riding with Coco on a road bike! :grimacing: