New to this and my data is still being chewed on and analyzed. But I noticed some entries that give numbers that are (way) off from what Strava reads - any idea why, so I can understand and/or correct?
See for instance https://intervals.icu/activities/2283718812 (I guess the link to Strava can be followed from that activity, but it is https://www.strava.com/activities/2283718812).
The chart shows the right values, but the average/NP and the training load do not.
Thanks!
I see what you mean. I will look into this. Tx for the report.
No biggie - itâs old data anyway, but Iâm curious. Recent data all seems okay 
Activities with gaps of more than 8 secs between data points have this issue. I am going to deploy code to fix all of these just now. Luckily they are quite rare.
I have too many to go through them all, but it looks like this mainly/only occurs during indoor rides. I had a Tacx Genius (upgraded to Smart at some point) until late 2017 and then got the Neo. But, it also looks like a lot of them were recorded using the Rouvy (then CycleOps Virtual Training) softwareâŚ
Yours have been fixed now. You had approx 20 rides with the issue.
BTW, how did you fix that? Anything I can do?
I just changed the power and HR interpolation code to accept bigger time gaps when the ride has big gaps in general. I can identify rides with big gaps easily so not hard to then fix everyones stuff.
Have no idea where that massive average power came from. Going to look at that now. At least avg power isnât used for anything much.
Okay, thanks. As you can see, I have a lot of data, but most important is the data from the last year or so. I cannot easily select for (possible) discrepancies, so I just go through the data randomly as it gets analyzed. I suppose it will be best if I just ignore these, so you can do other/better things 
Please send in anything you notice! I am going to go run queries to find âway out of line avg powerâ but I wouldnât know to look if you didnât report it. I used to use avg power from Strava until a month ago (as Strava has zero for some activities) so its quite possible there may be issues with it.
Fixed. The ârecording timeâ calc didnât like your rides with huge time gaps either, I forgot to check it before. And average power = work in joules / recording time 
Code is currently running to fix all such activities.
By âhugeâ you mean how long? Indoors, my brakes hardly are longer than 5 minutes (mostly 30 sec to 2 min), but during long days in the saddle outdoor, I may take a lunch brake 
Lunch breaks are ok. Long gaps (more than 30s) between the normal samples were the problem.Intervals.icu considered those to be âlunch breaksâ and ignored the data
I see - Iâll try to minimize those 
