Yep - I’ve made this very point and it’s been dismissed.
I was riding with Miguel yesterday (1.6 w/kg IIRC) on the Watopia Waistband route and, even though it’s relatively flat, you still need to be on your toes to stay with the RoboPartner group. It needs a lot more concentration than might be expected. I also think it’s possible that trainer lag plays a part.
FWIW I also think the power gap between Miguel and Maria is too large. On yesterday’s RP selection screen it was 1.6 w/kg and 2.4 w/kg. I don’t think any other two RPs have a 50% difference?
FWIW, the screen shows what they’re currently riding at for wherever they are on the course, not their listed flat output. When you viewed the selection screen Maria was going up a hill and Miguel was going down a hill. They’re set at 1.8 and 2.2 on flat.
Taylor = 1.1 W/kg, 83 W
Bernie = 1.5 W/kg, 113 W
Miguel = 1.8 W/kg, 135 W
Maria = 2.2 W/kg, 165 W
Coco = 2.6 W/kg, 195 W
Yumi = 2.9 W/kg, 217 W
Jacques = 3.2 W/kg, 240 W
Genie = 3.7 W/kg, 278 W
Constance = 4.2 W/kg, 315 W
All pace partners are now 75.0 kg and 1.75 m in height.
They will use dynamic pacing, giving an increase of up to 10% power going uphill, and a decrease of up to 20% when descending.
Yeah… I see what you’re talking about. Currently the last few times I checked, they’re listed actually at 1.7 for Miguel (on both routes) and 2.5 on both routes for Maria. Maybe @James_Zwift can comment?
Nonetheless they are 1.8 and 2.2 robo pacers if you actually view them riding. Recent write-up here:
ttps://zwiftinsider.com/robopacers/
I wonder if reducing the weight of the pace partners to 65kg and adjusting power down to keep same w/kg will help with reducing the speed.
This is just a red herring, and breaking a second thing to try and make up for the first thing being broken.
Fixing descending is a general-case problem, nothing to do with pacebots in particular. The descending model is simply based on bad physics. That’s what needs fixing… Not making compensating tweaks to further break pacebots to try and compensate. The entire “dynamic pacing” solution and all further tweaks to it is like constructing increasingly elaborate bridges and bypasses over and around a pothole… rather than just fixing the damn pothole.
How would it further break pace bots? You didn’t give reasons for it.
If we remember the Anquetil bot was around 65kg and didn’t go excessively fast - it was easier to stay with everywhere compared with the current Constance bot. The pace was slower in Tempus as well because Anquetil had less outright power.
Still being worked on. I don’t know off the top of my head when it’s meant to be out officially, but I can tell you I’ve seen some tests internally of it, so it’s still very much alive and kicking.
They get dropped but pedaling through the crest in races and rides too,
but now MORE
as the bots slowing down at the crest trains then to do the wrong thing.
Coming back to this comment a couple months later, because it was revealed in the new testing for pack dynamics 4 that there is a bug that puts the brakes on if you go below 10W (or coast) during a group ride (including robo pacer groups), which accentuates getting blown out the back if you try to ease on the downhill to let the group catch up.
Did I read though that you’re removing that code in the next update, as mentioned not having anything to do with PD4? I don’t suppose you came across any rationale for this to have been written in the first place?
