"Keep everyone together" explanation?

For the past month or two a group I ride with have been using the “Keep everyone together” feature on meetups and it’s been useful when riding with riders of differing abilities, especially when we’re on a route with longer climbs.

The past couple of days we’ve noticed behaviour that seems different from how it used to. It seems that there is a specific rider/order to the bunch which is very hard to change. For example, on a longer climb there will be a rider at the front that is very hard to catch, let alone pass. He might be doing 3w/kg with a few metres gap to me. If I do 5-6w/kg to close the gap, as soon as I get close he suddenly speeds up (or I am slowed down) and the gap is there again. The same thing for the rider behind me.

We are fairly certain it didn’t used to be like this until a few days ago. We could ride more like a bunch, if you were doing more w/kg you’d get ahead a certain extent (only a couple of seconds) but the bunch would generally be ordered in a way that made sense according to our power/weight. This no longer seems to be the case.

Can anyone explain the behaviour, how it works, if there have been changes recently just so we have an explanation for what is going on?


I have questions concerning the group riding dynamics too, and so I’ve decided to post here in the hopes that we will draw the correct “eyes”… As a result of riding in group rides associated both the the Zwift Academy, Zwift Tri, and other group rides where “keep together” appeared to be in play, I noticed that my speeds are dramatically different (for the same power output on similar terrain geometry). My questions are manifold:

  • What factor determines the group average? Is it the speed of the ride leader? The slowest member of the group? Or is it an average of all the riders? Or is it configurable by the organizer(s)? Other?

  • If speed is affected by the “keep together” rubber-band, are distances similarly affected (YES) - but how? Surely the distance covered by a group over a given amount of time should permit the game to calculate exact distances, no? If I had been riding with Coco for 75 minutes at 191 W average power, I would have covered just over 50 km. As it is now, my group ride only has me traversing about 30 km - a huge distortion!

  • What about “energy”? Calories? Are the calories transformed measured according to my actual power output, or are these numbers also affected by the group ride “keep together” algorithm???

I like to keep careful metrics, and while I enjoy these group riding dynamics, the keep together function appears to throw a wrench into accurate data-keeping.

Here are some specifics. Today, in anticipation of an Academy Tri Bike Workout (2), I rode with Coco Cadence (who most assuredly does not have a keep together rubber band) at power outputs that coincidentally were nearly identical to my subsequent workout. My average power in both rides was 191 W, and while the terrain was slightly different between the two rides, this differential was smoothed by the fact that I only did 20 km with Coco, while the Tri Workout was of a much longer duration and mostly flat.

Coco Ride: Average Speed 40.1 kpg

Academy Tri: Average Speed 24.9 kph

Can anyone shed any light on this dynamic? As I said, I enjoy these group rides, however, I run very structured training parameters, and these “keep together” group dynamics really throw off the calculations. Am I better off just duplicating the workout details into a custom workout, and ride the workout solo? Is there any way to adjust the group rides to obtain correct speeds and distances for the power that I output? Solutions? Ideas?

Thanks for your consideration!


This is a cross-posting from an old thread that has not garnered any replies. I am posting here in the “General” category hoping for answers…


  • Zwift have not declared (from what I know) the algorithm for what determines the overall speed but the consensus, from all the posts on the subject I have read, is that its a combination of all riders (its not the fastest), and its not the slowest.
  • pointless asking about speed/distance in group rides - its irrelevant given you can have riders at the extreme of the “keep together” dynamics, that will cover the same distance (but have significantly different watts and/or w/kg). If you care about accurate distance for the watts you are generating, I’d be doing the ZA solo mode.
  • if you are a numbers nerd, the only thing you can take from these group workouts are the watts - which is kind of the point of workouts in the first place, not distance. Distance is not material and only a by-product (ie, the same workout on AdZ covers a different distance to Tick Tock, but the watts are identical).

Comparing a non-“keep together” ride with a keep together is pointless. One ignores drafting and focuses on watts, one is just a ride and has variable drafting on a different route.

1 Like

Thanks for your reply. I AM a numbers geek (nerds only talk, geeks get it done!) One of the metrics I inquired about was calories/kj… Do you know whether these are similarly distorted by the “keep together” group dynamic or are they calculated on a derivative of power? Thanks for your help.


Calories will be as accurate as can be given they are only determined from power, which is as accurate as your trainer/power meter.

1 Like