I would attribute that speed change to the slope variation. In Zwift sometimes you see grade at 0% when it can be close to -1% or +1% (like -0.7 or -0.8%) that shows only with a single digit in the HUD. I believe the value is truncated to show no decimal cases 
So with a negative slope even if shallow I would expect such speed variations
If you’re referring to about the 2min mark, when the 48-49km/h is hit, I do see the gradient has gone to about -.7% though she has also dropped to 3.9wkg from her flat 4.2wkg target. An online bike calculator for a solo rider tells me that:
At 315 watts for 75kg rider, solo speed is ~40.6 km/h on 0% gradient
At 3.9 wkg ie 293 watts on -.7% gradient, speed is ~43 km/h
So differential in speed is about right, but absolute speeds in the calculator are quite a bit lower. According to the calculator to hit 46km/h would need about 450 watts on the flat without draft.
Don’t we have the majority running 1.33.6 with a significant minority running the new 1.34? As well as the insignificant minority running 1.33.5 or older. Are the RPs effectively on 1.34 too?
No, around 28:10 on the video.
@DavidP look at my grade setting. Sauce is showing gradients with .1 precision and it shows those in places on Tempus Fugit but not at that point.
When I say the latest game version I mean 1.33.6, sorry 
It appears Sauce is not displaying negative slopes correctly. It shows only as -0.0%…
probably because it’s somewhere really between -.01 and -.04% and doesn’t show to hundredths of percent?
1 Like
I promise you it does. It’s showing that because it’s a negative grade less than 0.0% 
This is a few minutes later in the video going through the caves where there is a slight incline/decline
I wished we all could have that Sauce 4 Zwift functions (or at least a few of them).
1 Like
ahh… I see.
It looks like the initial speed burst could have been instigated by that lone rider that group came upon that she looks like she slingshots past.
I still don’t understand the overall speeds though for her riding off the front. Feel free to play around with this tool and see what you come up with as req’d to hit 47kmh?
(add an ‘h’ at the front)
ttps://silca.cc/pages/power-calc
I rode with Maria on Monday for 90’ and, usually, this would be my “easy” ride too (I’m mid-to-low C, 57 kg, ~3.1 W/kg FTP at the mo’). I did find the ride to be a little harder than anticipated but my average power was only 2.44 W/kg for a 23.6 mph average on the Volcano Circuit. I never felt in danger of being dropped, but I did need to put a few hard digs in to keep pace.
Coco, on the other hand, has felt considerably harder for me in recent weeks. I’m not sure whether this is simply due to a loss of fitness (probably the main factor) or the introduction of PD4 or a bit of both.
1 Like
I can definitely attest to this type of unexpected PP surge (with Coco, Yumi and Maria) and I think it’s been going on for a long time, definitely pre-PD4. It’s very sporadic and might only happen once or twice in a 60-90’ ride (or, sometimes, not at all?).
I think the tiny downhill etc is a red herring - if we look at the video from 28:10 the pace partner is in the wind and getting a 200w draft! Off who? That’s more likely to cause the speed up - the guy behind has to put out big numbers just to follow. Definitely looks weird to me.
1 Like
Just throwing this out there; what’s the possibility of the RP’s not being affected by auto-braking?
Now that we don’t get the auto-brake flashes, is it possible that the bots are slinging out the front while a “mass” of the blob is getting brake checked?
Auto braking is kind of a weird mechanic, and how and when it works I haven’t fully wrapped my head around, but I’ve noticed how … easy it is at times to block someone in during say a race. I literally did it on Wednesday on a flat, I eased, and someone behind me started to put in an effort, but their avatar just sat behind me as I continued to ease, and they continued to ramp. (Which I know this is unrelated but, this was kind of an ugly situation to me, as it was just two of us, so plenty of road to “get around”, but their avatar decided to get stuck behind me despite a solid 1-1.5+ w/kg increase from the person behind me; eventually they backed off when they realized it wouldn’t let them pass for whatever reason.)
Is it possible what we’re experiencing is just a traffic jam of auto-braking, and the pace partners aren’t experiencing this?
The RP game build is using the same physics code, so they are subject to exactly the same rules. However given that they keep a constant pace, only on downhills when they reduce their power they might be slowed down depending on the conditions.
This is also the gamification part of Zwift. Consider them just as a guiding beacon where you need to “master” your power output to keep up with the group.
From my investigation so far, I believe this is being caused by lag from Sauce and all the network ins and outs. Sauce doesn’t refresh the interface every second.
If you watch the stream there are also parts where the opposite occurs, the pace partner is inside the group drafting and the reported draft savings are 0 watts. Same for other riders. Sometimes it takes a few seconds to update. With all the riders being overtaken free riding in the world, it is quite possible it causes this.
So far I haven’t found any possible reason for things to behave differently for the Pacers. They run the same physics code as everybody else.
1 Like
I hope this isn’t the case. The last thing anyone wants is a return to the stronger “sticky draft” of old, even if it’s just a similar effect for “legitimate” reasons.
So you’re telling us that random accelerations due to phantom draft is due to relative latencies of game client and robo pacers? That would have to be multiple seconds of difference though…
No I think what is being said is SAUSE is not as great as people think it is.
This reads more like a personal beef more than an objective answer. I am looking for a plausible explanation for random surges from stable speed of 47kph or so to 49/50 which I can’t visibly explain, even with the game HUD up which is 0 delay. It didn’t make sense, and Sauce data made it seem even more off. Calling it an inferior product doesnt’ give me an answer
1 Like