virtual power is way off

(Greg Woitzik) #1

Just got back from vacation this morning and jumped on zwift for a quick spin. When it was locating my devices my powermeter wasn’t awake and under speed cadence I was able to select my trainer (Kurt kinetic). I assume at this point zwift was calculating virtual power and not using my powermeter. What I noticed, is that zwift was showing 50-60 extra watts compared to what my garmin was showing for my powermeter. I even stopped 20 mins into the ride and calibrated my powermeter but the difference was still the same. Is there going to be some way to address the calibration of dumb trainers when using a speed cadence sensor? I liked the extra power I was getting (best time by far on zwift kom) but I know this isn’t real (and won’t be posting to strava) and will make sure only to use my powermeter in the future.

One other note on today’s ride. I saw a few riders going in circles and off the course in certain locations. Never seen this before. The most riders ever were on today (full list +20 that weren’t showing) and don’t know if this was causing issues.

(Greg Woitzik) #2

It’s a powertap g3. Seemed to stay about the same difference apart the entire ride. I did recalibrate the powertap 1/3 of the way through but didn’t make any difference. How is someone to know how much pressure to apply? Is there going to be a roll down calibration or anything similar?

(Noel Nunkovich *USMES*) #3

Virtual Power, based on what I know about it, is always going to have some variability. The key phrase here is “consistent but not (necessarily) accurate.” Because it’s not a direct force measurement, things like tire tread, tire pressure, and the pressure contact between the rear tire and the trainer’s roller all affect the calculated virtual power. As an easy example, heavier trainer tension against the tire causes the tire to deform more as it rolls. That means more of your power output is absorbed in deforming the tire, therefore lower wheel speed for a given effort.

It’s probably possible to “calibrate” your trainer fairly closely if you have a power meter but it’s always going to be a very user-specific thing. Roll-down calibrations are problematic, as I understand it, simply because a wheelspeed sensor doesn’t have enough resolution (a single data point per revolution, where multiples are really needed)

The thing is that Virtual Power is very usable for an individual’s training, it’s just very hard to compare it with a direct force power meter. It’s really only a problem when you have a situation, like I do, where you have multiple bikes you’d like to use on the trainer but only some of them have power meters. I think Virtual Power is always going to be a problem in that sort of situation because it’s really only useful when being compared to itself.

(Greg Woitzik) #4

Agreed, as long as variables are kept the same (roller pressure, tire pressure, etc) virtual power can work well for training-repeatable. However, for competition on zwift, sprints, koms, and strava segments, it’s not going to be very accurate.

(Mike Davidson) #5

Agreed, Zwift’s virtual power is too high for some reason on a Kurt Kinetic Road Machine. I’ve compared it to Trainerroad’s virtual power, using the same tire pressure and pressure against the flywheel drum, and it’s significantly higher on Zwift, similar to the result you are getting Greg (50-60 watts). In the past I have found Trainerroad’s virtual power not far off my powermeter, at least within reason for me not to worry about the discrepancy. Not sure why this would be the case given that Kurt Kinetic has in the past supplied the power formula for their trainers, Road Machine being P=(5.244820)*S+(0.01968)*S^3.

(Nigel Van de Velde) #6

I noticed the same thing for a Kurt Kinetic Road Machine II.

Today i did another Zwift ride and simultaneously captured the data with my garmin device (1s recording, no autopause). After the ride i editted the data of the garmin’s ride and calculated the virtual power for every datapoint with the formula stated on Kinetic’ website (

That way I ended up with 2 files; 1 from Zwift containing virtual power and 1 from my garmin containing virtual power. Following link shows the respective power curves of the 2 rides.

As you can see the profile is indeed a lot higher for the Zwift recording…

I realise virtual power calculation is a lot harder on the Zwift course, as they need to compensate for elevation as well, so they can not use the formula from Kinetic’s website… Still I hope that this can be solved during the beta testing.

(Greg Woitzik) #7

Nice graphical representation. Shows exactly what we have been experiencing.