Zwift racing - what is valid power data?

Newbie on Zwift racing, but I have done some hours on Zwift (level 26).

There are always discrepancies in readings across power meters. I own eight power meters (brands are SRM, Assioma, power2max, Powertap and an Elite Direto X trainer). I like to compare them with the power analysis tool on Zwiftpower. The agreement is pretty good, but anyway, there are differences up to four percent, which can be pretty significant for racing performance. Even one percent could give a winning advantage.

I see racers have to submit dual recording of power data. How much difference is tolerated, and what is considered “the truth”?

With all do respect… I mean… maybe ask the Zwift forum.

Bottom line zpower is not legit. Basically turns a cat 4 into cat 1 for racing purposes. It might track fairly well ‘on average’ but that means little with the volatility of racing power data (as opposed to say a TT steady effort)

Thats not at all what he is talking about.

ZwiftPower (that he mentions) is a separate deal that also has an option to compare power from different sources.

He is using that compare data from actual power meters and smart trainers, not the virtual power from Zwift (that is confusingly named ‘ZPower’).

1 Like

Interesting read. I did my own analysis in the past dual recording my hammer smart trainer and stages L side power meter. They tracked closely, but ass power levels increased, the meter with the higher readings switched. I don’t recall exactly what is was but suppose

hammer is at 200w / stages reads 195
Hammer is at 300w / stages reads 290
Hammer is at 400w / stages reads 410
Hammer is at 1000w / stages reads 1100w

You might look into some rule books for serious eRacing leagues. If it’s not a rule, it’s not considered cheating. I doubt it’s a legit rule yet. Any smart racer would choose the higher meter to link to Zwift. Showing your second file just proves your meter didn’t have unusual power spikes. I also imagine at the big races in person they prob have everyone on the same smart trainers to make it fair (like with the Covid series of Super League Triathlon)

To my knowledge, it’s not published.

The spread is all over the place.

You can look at the analysis tab on the major races and you can see wattage difference between 5-20w on the 20 minute averages: Zwift Power - Login

1 Like

I’m sure I read somewhere that you have to use the power meter in the turbo trainer as your main source of power for racing. Seems a good idea to me as it means you won’t have people getting free watts just because they’re measuring through a power meter in their pedals. If you’re racing at the very top level there’s a list of sanctioned smart turbos that you’re allowed to use.

Having said all that whenever I’ve raced there always seems to be someone putting out more power than they should be for the level of race I’m in, even after clear cheaters have been removed by zwiftpower! The racing is fun though if you accept its limitations

Just found the rules here:

Section 2.5.1 covers what you’re after

Thanks for the information. From the rules:

“Only trainers and smart bikes with a manufacturer claimed power reading accuracy of
+/- 2% or better shall be permitted.”

Lucky riders have power meters that read +2%, not -2%…

My understanding is that the dual recording is less to do with power meter accuracy and more to do with combating “saw-toothing”.

When on zwift you get a few seconds free power when you stop pedalling - a feature put in place so riders don’t notice short drops in signal.

E.g. if you were to pedal at 300w for 5 seconds, stop for 2 seconds then pedal at 300w for 5 seconds and repeat then zwift would record that as a constant 300w.

Dual recording catches this as the other device would record the pauses in power and the average would be significantly lower.

1 Like
  • Do you have a source for this info?

In my experience, the “sticky watts” has more to do with the power device in use. Power meters (especially pedal based ones) react fast for ups/downs, while smart trainers have more delay in ups/downs. That seems related to the way they measure power specifically, more than some fudge factor Z is adding AFAIK.

Very much just relaying something some random guy on the internet wrote one :wink:

“Sticky Watts: Why They Exist, and How They Affect the Zwift Experience - Zwift Insider” Sticky Watts: Why They Exist, and How They Affect the Zwift Experience - Zwift Insider


1 Like

In my experience it’s device and potentially connection dependent

I have tried to replicate this on a smart bike and it doesn’t seem to happen at least for me. Here is a snippet from some 15/15s which would be the ideal test case. Although perhaps it’s because I’m not going to full zero on the cadence?

1 Like