Pack Dynamics 4.1 - Test Events

There have definitely been issues in the past along these lines. So I think so, yes. I seem to remember us being told in the past that no, that wasn’t a thing, and then it being discovered by Zwift that yes, actually, it is a thing.

I think slingshot is the best way to describe what I saw on Turf n Surf with Yumi yesterday.

On both climbs I was having to push between 3.2 and 3.7w/kg to stay just ahead of Yumi. Both climbs have sections around 4 to 5% over longish sections and if I held this power I could generally stay just ahead of Yumi.

I think the slingshot happened when the gradient slightly eased. Of course, it’s possible that my power could have slightly dropped as the gradient eased, but I don’t think the slingshot of Yumi was solely down to that.

I had to kick again to catch Yumi, which I did every time, but it did mean there were quite a few surges on the climb.

As I said on my first post, my overall feeling about 4.1 is very positive, so I’m only raising this issue to see if others have seen it - it looks like they have. If it’s possible to address this, then it just makes the 4.1 experience better!

No that’s not possible, unless there’s some problem we’re unaware of with loading server settings. The pacers are updated with the latest 1.45 version.
Besides, if that was the case we would be seeing the same behavior everywhere not just in certain parts of the world.

1 Like

Deja Vu?

1 Like

i didn’t really think anything much of it at the time since i only rode a few test races but one of the things i noticed as a consequence of a stronger draft is that shallow climbs are very easy to rest in the draft on but nearly impossible to ride at the front on

Observations after couple of group rides in (C & D)
You can’t float just off the back of the blob anymore, which as a sweep is annoying !
The speed of the draft drop off is catching people out, they are dropping off fast despite the slower pack speed, and with no draft to help them they stop peddling, as @Andrew_Nuse suggested to me on the C ride.
A “mob handed” sweep seems to work better.
PD3 to PD4 took a few weeks to settle in and get used to, especially the long draft making riders too comfortable off the back of a blob, until the draft ran out and they were dumped 10s off and suddenly they needed a plus 1.5wkg over the blob to catch.

TL:DR its different

The last couple of RP rides since 4.1 switched on in Makuri have definitely generated some strange experiences. If the RP gradually falls into the middle or back of the leading pack, it suddenly surges ahead significantly leaving quite a gap to all riders. That obviously generates a surge from the pack, but the difference is that the RP doesn’t seem to be affected by the new draft dynamics like everyone else. So it leads to quite a volatile ride compared with previously. Not too much impact on average power over an hour or so, but many more surges and slingshots.

Agree. While overall churning in the pack seems reduced, the RPs seem to churn more and don’t seem to follow the same dynamics as everyone else. Today the RP flew past me on a descent which seemed odd since I weigh quite a bit more than the RP and was drafting in a pack doing 0.5 w/kg more than the RP.

They have steering already but it’s troublesome, people get dropouts with Zwift play and it isn’t very smooth operating even with Kickr Bike. On kickr bike the steering is pressing buttons, not very accurate. And with Zwift play it can be hard to get the precise line you want, too easy to steer too far.

With the dropouts, you can get stuck to the far side and have no draft at all. When it happens on a A grade robopacer ride you can use the teleport function to get back to middle of the road, but on a race that won’t be possible.

Zwift play controllers are more useful for controlling other game features like ride ons and power ups, steering in every form I’ve tried is fairly blah.

Sometimes Constance can zoom past a quick group of riders uphill even while we are doing much more power. Constance also gets a rocket like slingshot up the start of small hills compared to everyone else. This is on Watopia however.

Watch my TEST with RP Genie and you will see the problem. My bot had PD4.1 movement…but Genie is like slingshoting all the time.


Same for me yesterday, did and hour with Coco on Makuri and was definitely some weird surges going on when the RP would suddenly be ahead of the pack, when the pack hadn’t change power and the road was flat.

Wow that camera yo-yo effect will drive me crazy. :roll_eyes: :woozy_face:

1 Like

There are fewer than 300 people in all of Makuri right now and the pace partners are all in Scotland and Makuri isn’t on the guest world calendar until the coming weekend. Going to be really hard to get some test rides in.

Showing that RoboPacer is still on the old Pack Dynamics is far more important than Rubberbanding on my Laptop :rofl:

This seems to suggest that the RPs are not exactly just like any other free riding zwifter and need course specific programming and updating as things change in Zwift? Maybe it’s related to specific areas of specific routes only?

Sorry - We know there’s a problem and we know the solution and it’ll be fixed before PD4.1 is formally released.


Is there no way to have the pace partners just read the Gradient and behave appropriately? Would add a lot of flexibility.

They should have been restarted prior to the change. We’ll make sure that they are prior to 1.46.


A little while ago there was an issue with pace partners surging at inappropriate time and IIRC it was caused by incorrect route information loaded to the bots. Why do they need to know the route instead of just reading the current grade? Separate question from 4.1 issues.

I think it was because the pace partner client was running a different version to game. I don’t think it was routing information but could be wrong.

David has already confirmed that what I stated above is correct.

1 Like