Lag in power reporting on effort change - desktop, not mobile app

(Dominic Allkins (BMCC)) #1


Loving Zwift - has helped immensely with my training :slight_smile:

Since the last patch I have noticed there seems to be a lag between change of effort/cadence and the power reported on-screen. Just riding around this wouldn’t be so much of an issue but when doing intervals it makes it tricky to get the level right.

As an experiment yesterday I ran the mobile app next to my laptop and the mobile app reported the changes in effort instantly - no lag.

My set up is Mac OSX El Capitan & connected via Ethernet, not WiFi.

One of my buddies has noticed the same lag on PC as well.

Hopefully an easy fix. Keep up the great work.

(Noel Nunkovich *USMES*) #2

I suspect what you are seeing is by design. The most recent update from Zwift finally introduced power smoothing, something that has been requested for some time. The smoothing is actually a rolling 3 second average of your power number, designed to make the number less “jumpy” while riding. I’m not sure if you use a power meter outdoors or not, but 3 second average power is what most people display on their Garmin or whatever when training or racing with power. The number just jumps around too much to be very useable otherwise. During time trials I usually even go a step farther and use 10 second averaging as my pacing number.

The downside of power averaging is obviously just what you’re seeing: a slower reaction time. The 3 second average is really kind of a compromise. It slows reaction enough to make power useable at steady state while still making it react fast enough to not be a huge problem when switching from one power level to another, like in intervals. 

My recommendation is to just ignore it. I think once you get used to the reaction time you’ll probably find it easier to use.

(malcolm may (WCC) C) #3

However on the road, when you punch it, the bike moves immediately. On Zwift your bike doesn’t move until your power checks in. By the time my watts climb, I am verklempt. So your max sprints take much longer, and it becomes something different. I suspect it means what doesn’t kill you makes you stronger. 

(Michael Henasey) #4

there is always going to be some delay or lag between your instantaneous power changes and how your avatar reacts in-game.

some of this is built-in I suspect while some of it is natural for the information has to “travel”. the built-in aspect is sort of like the 3-sec power smoothing/averaging feature. if our avatars were trying to react to every sudden power/cadence change, the avatars may appear to be very jumpy and this would frustrate users or detract from the sense of “realism”.

now the mobile app and the w/kg display next to your name in the rider’s list reflect your instantaneous power, not your 3-sec avg. power. any delay or lag you see there is simply a result of how the information must “travel”.


(malcolm may (WCC) C) #5

Nice story. However it certainly has no lag time when you ease off. :slight_smile:  

(Michael Henasey) #6

putting zeros into an average calculation has a big impact…

(malcolm may (WCC) C) #7

I don’t go sit in the corner. I said ease off, not immigrate to Tahiti

(Michael Henasey) #8

perhaps you could provide an example and/or a video demonstrating this “easing” off affecting the avg. power value much faster/more significantly than when you are applying more power to hit a target?

(malcolm may (WCC) C) #9

Why? Is it going to influence a change? Or is it just an intellectual exercise for you? I have a number of disappointments, that are liveable as they are consistent. And being a zwift jersey winner is not a goal. 

I have a lemond revolution trainer. Its fan based so takes a little to generate watts. But it doesn’t take 10 seconds. If I jump on a sprint, I may be 50 yards in before my watts exceed 300.

My secondary speed measure is 25 + mph, so the lemonsd is recognizing my acceleration, but not zwift. Eventually I’m holding 1000+ for 10 seconds, but am exhausted when I get there. On the road with my powertap wheel, thats about 400+ short. And I get there much quicker.

So wah for me, however I suspect the calibration is linear on the lemond. Instead of with an upcurve. I provided a power curve which I found on line, and I can ‘feel’ its not the same as the zwift calibration. How do I know? I don’t, but its consistent, so I can live with it. I have the new wattbox on order, which should change the game in that I’ll get a lemond calibration for the trainer.

And then I find that unless you are not on a kickr you get capped at 1200. No issue cos I’m only pushing just over 1100 at this time. And its certainly better than when I was capped at 400 for having a ‘dumb’ trainer.

As it is, I’d proabably get a more accurate ‘game’ on trainer road as they have it set up for the measures and nothing else. However I like the view. Even if I am hammering at 25 + mph up a 3 % grade and the zwift calibration is crediting me with 4 mph. So how do I know? Cos on the weekends I ride much hillyer terrain for a much faster result.

Now you may say its online, and its all relative. And it is. But its fun and consistent. And only to get me to outdoors while I am rehabbing my back. And don’t get me started on the downhill…

(Michael Henasey) #10


(malcolm may (WCC) C) #11

:slight_smile: Its fun. Its imaginary, so as long as its consistent, its all good.

And I have learned a whole lot more about the use of watts since riding zwift. Its a tremendously interesting subject.

(malcolm may (WCC) C) #12

Found that working twice as hard for half the outcome soured the experience.

After finding out that Zwift were uninterested in correcting the calibration for the revolution I quit, and moved to trainerroad.

Not as scenic of course, but more structured (less fun)  and the watts align with my powertap wheel outdoors. 

I’m no longer killing myself on the mtn for half credit. 

Zwift was fun while it lasted.