NP lower than Average Power (ZwiftPower)

That’s very odd.
It looks like a discrepancy between the live data and the fit file. Gap in the fit file data maybe, or the peak being trimmed. No idea how to prove that though.

ZwiftPower power start about 15 sec into the race. it odd time line to start.

I’m not 100% sure so a bit of an educated guess.

Your 1 minute power matches across Zwift race results and also ZP results so you would expect your 5 minute power to match.

I think the change is due to originally on ZP you were looking at the green lightening bolt / Live data. Overnight it then changed to the .Fit file data but that, in theory shouldn’t have changed it much. However if you look on your ZP power profile for the race you will notice in the last 5 minutes of the race you experienced a lot of dropouts in the last 5 minutes. (But not within your max 1 minute power).

I wonder if when Live data became .Fit file data these drop outs caused the 5 minute power to reduce.

Suggest downloading your .Fit file and see what figures appear within last 5 minutes of your race.

Hi @Jeroen_Diepenbroek_T welcome to the Forums!

@Doug_Ox Thanks for sharing information.

I would like to explain that once you reach 90 days this might change and your data can change. However, in order for us and give detailed information and diagnose if there’s something else happening internally, I invite you to contact us and we can provide one-on-one support and check your account with our ZwiftPower team.


I would like to report a what seems to be a up with Zwiftpower. It has happened a few times to me already. After the race, my results appears to be correct but after few hours, my NP, average HR and other stats changes (gets way lower). I checked other apps, such as and the stats are correct there.

Here are the links to both sites (its a C pen and my name is Veronika Vo):

  • ZP - /events.php?zid=4361779
  • Zwiftracing app - /events/4361779

Could you please look into what is causing my stats getting changed and lowered?

Thank you

Hi @7_Veronika_Vo_GXY_CS Thanks for your submission! I’m Norman from Zwift. Sorry to hear about this issue, sounds like it may be isolated to your profile unless some more people start reporting the same issue.

Still, your experience is very important to us! Could you do me a favor and share these details with the Zwift Power team? You can email them directly to, I’m certain they will give you a hand.

I think this is most likely the same issue.

Common thread is the sight of a lot of HR and Power dropouts to zero on the ZP ride analysis.

The Zwift support team wrote me this:
"The results on ZwiftPower are for the event settings (e.g. time, distance, etc.), and this can differ than what’s in your FIT file. There can be time in the pen and riding after the event saved in your FIT file. ZwiftPower will remove that extra data and only reflect for the time within the event. Your NP, 20m, etc. on ZwiftPower will also change to reflect the data within the timeframe for the event you participate in.

With ZwiftPower, it can take up to 24 hours to pull data, trim the results to reflect your event riding, calculate, and update your results. What you’re most likely experiencing is your FIT file get trimmed to be within the event time range.

This appears to be the case with your Stage 2: Bag That Badge - Hilly Route Reverse and other events you participated in. "

Im sorry, but I don’t understand that reply. Are my heart rates in the pen being counted in the race statistics? Are my watts in the pen being counted in the race statistics? Does Zwift want people to join the race at the last minute? Because then the data on ZwiftPower and now on the ZwiftRacing app are skewed. I have an average heart rate of 143 on ZwiftPower. On the site „“ I can see that heart rate didn’t drop below 151 bpm. So, how is the correct value that I had an average heart rate of 143 for the race?

1 Like

Here is your HR and power analysis from ZwiftPower for that race.

I’m sure you didn’t have that many dropouts.

1 Like

Yea, I didn’t have any dropouts at all. To double check, I downloaded the FIT file from zwift com/feed from my activity. I used fitfileviewer com to analyze it. There were 1801 records found and none of them had zeroes for power and/or heart rate…

1 Like

i’ve seen it happen to many people now. as far as to what causes it i’d just be guessing and i’m not a website developer, but it’s a zwiftpower issue so it’s nothing wrong on your end, i’m sure of that much.

This answer does not make sense. (Speaking as a site reliability engineer) I would want to gauge the scope of the problem with a database query that selects all activities from ZwiftPower where the fit file has been processed and NP < average power. See how that result changes over time to possibly pinpoint when it changed significantly, and look for commonalities in the affected riders and activities in the hope that a clue leaps out of the data.

Yea, does not make sense…

  • They were saying that I had dropouts.
  • I analyzed the file with fifileanalyzer com, there were no zeroes.
  • They said that the tool excludes zeroes.
  • I analyzed the FIT file via python fitdecode module, there were no zeroes
  • They replied that it seems that the FIT file is only partially synced in between Zwift and ZP and that I should wait.

But I had another activity with same symptoms on May 2nd (19 days ago). How long do I have to wait? It seems like a bug to me rather a miss-synced file…

I’ve just found this thread. My zwiftpower numbers are all off on my last event due to dropouts that appear in the zwiftpower file only. My fit file from Zwift.and Strava show no dropouts and therefore higher numbers. I don’t suppose there’s anyway to have a zwiftpower file re-uploaded. Rather annoying that one of the data points in my list of results is incorrect. And it’s the one that upgraded me to B based on my 5 minute power.

This issue should now be fixed in ZwiftPower and also in the upcoming game 1.67 release.

Please let us know if you see more instances of this problem or other fit file issues.