Hi
I notice I see different percentages for decoupling for specific intervals / sessions on Intervals.icu than I see on Training Peaks. Anybody seen this before and have an answer as to why it happens?
Also see the timings in the screenshot below from a short session last night.
5m3s and 30m1s is 0:35:04 , not 0:34:56 like intervals.icu reports. No pauses / stops in session.
Training Peaks reporting correct elapsed time.
Is this incorrect time reporting a potential root cause of the decoupling inconsistencies? Is Intervals pulling the average power + hr from the correct time intervals?
When you say zeros what do you mean exactly? There was no pause in that session. No coasting. No zero power or cadence.
Edit : Unless I didnāt pedal at the start for a few seconds. I will check that out
Edit 2 - 4s at the start but no other pauses or drops after that.
if you really want to check it out. You would need to either share the actual raw FIT file. (links to your activity in intervals.icu is only accessible to your followers / david)
off the top of my head, I donāt know y the difference in timing, but, in terms of drops / cadence drops / power dropsā¦ you canāt spot it unless looking at the raw FIT file at the per-seconds data.
other than that, what @MedTechCD says is what I usually notice as well.
Even power data (averages & TSS) are different between TP and intervals.icu (just few watts)
Thanks folks
That answers the timing question but why does Intervals.icu reoprt a NP of 143 watts for those first 2.5 minutes v 150 watts from TP.
If I omit the first 30s and analyze from 00:00:30 - 00:04:55, then both sites give an average and NP of 150. But including that 30s at the start, intervals.icu reports the NP as 7 watts lower than TP.
It looks like something to do with the zeros at the start but why? And which data should I believe is correct? This also affects the decoupling percentage because decoupling formula uses NP as opposed to average power
The last screenshot you posted there @app4g shows a NP of 150 also. (how were you able to get that data in that format?)
Go to the intervals data page and if that first interval is a āWorkā interval, you will see a pencil at the right. Click that and edit out the first 6 sec of the interval. You should now have the same result.
If it is a recovery interval, go to the activity page and click āOptionsā at the bottom. Check Hide Recovery intervals. It will now not show as an interval. If you want that interval as a Work interval, place your mouse cursor at the beginning of that part and push the āaā key to add a new interval. Drag the borders to select what you want as interval and make sure that the zero power points are not in it.
It is fairly uncommon to have a āWorkā interval right at the start of a workoutā¦
They are both correct, just different
TP excludes the zero power points from an interval, while Intervals includes the zero points.
I prefer the way Intervals is doing it because, if you have zero power output for 10-15 sec during an effort, and those are excluded, you will have an over-estimated Power for that interval.
Excellent. Thanks again. Ya, the interval itself is utterly insignificant, just turning the legs over really for a short warm up. I was just wondering why the decoupling discrepancy between here and TP because I have noticed it on a lot of files Iāve just gone back through. Thanks for the answers though. All cleared up now!! @MedTechCD and @app4g
Hi folks
Back to this again after todayās session. Something does not seem right to me regards NP (and consequently decoupling)
Session was as follows (5mins warmup, 10 x (2mins 200watts 60 rpm / 1min 300watts 105 rpm), 3 mins cooldown. Started pedaling well before I started the Wahoo so I would not have any zeros. No drops in power or cadence throughout the session.
Now look at the last 2 min interval. Intervals.icu is reporting the NP for this interval as 221 watts. I just cannot believe that that is correct. I have attached an excel file showing each second of data for that interval (1924s to 2044s). It just canāt be a NP of 221. TP reports what looks like correct NP of 203 v an average of 202.
Can someone please look into this for me and see why Intervals.icu is reporting NP of 221 for that interval ( I havenāt even looked at other intervals tbh because I just happend to see this first and went investigating from there) (Edit - Theyāre all pretty much the same - hugely exaggerated NP for those 2 mins intervals)
Original fit file + fit fit file data also at the link below.
You say that you are concerned about NP differences because they affect decoupling measures. Where are you getting decoupling from?
Intervals.icu does not use NP in its decoupling measure, as far as I can see. Decoupling is computed from 60s moving average power as a percentage of eFTP and HR as a proportion of HR reserve. These seem to me to be a matter of AP not NP.
Furthermore, there are differences in the way in which decoupling is computed in different sources. Intervals.icu calculation is different from the Seiler method [https://bluecattechnical.shinyapps.io/EnDuRA/], for instance.
Iāve always found decoupling by dividing the NP by the HR for each half of the ride/interval. Then itās subtracting the first half ratio from the second half ratio and dividing that answer by the first half ratio and multiplying by 100. Itās not complicated and doesnāt require rolling averages or HR reserve. How does Intervals.icu know my HR reserve in any case?
Whether or which, the NP from that data Iāve posted is incorrect. There is no way that the NP for that highlighted section in the excel sheet is 221 watts. The TP normalized power is correct.
As to your first paragraph,
[1] intervals.icu will do the calculation for you on the activity/power page;
[2] you must have input resting HR and max HR;
[3] everyone Iāve seen uses some version of AP not NP, but since the calculation of decoupling is intended to apply to rides with +/- constant power, this is moot.
[1] - Garbage in means gabage out and I donāt know my max HR which means any calculation intervals.icu would do regards HRR will be inaccurate at best and garbage at worst.
[2] - See above. I only know resting HR
[3]- Iāve always understood decoupling to use NP but my exposure to this has been almost exclusively from Training Peaks. They use NP.
In any case, I have changed my max HR in the settings page and recalculated the data from the session yesterday. NP and decoupling have remained the same which would suggest decoupling is not measured off a moving average or HRR.
But to my original query from yesterday, if anyone can shed some light as to why the NP for each 2m interval is 221 watts Iād much appreciate it (I know itās a v short interval and thus NP calculations are probably best ignored, but a difference of 18 watts between here and TP is still significant and Iām just wondering why there is such a difference)
So Iāve done my own calculations for NP from the excel sheet I attached yesterday and I got 203 watts, the same as Training Peaks.
I took the moving 30 second average of the power data from the interval.
Then I took the moving 30 second averaged power data and raised each value to the 4th power.
Then I got the average of that data from step 2 and then got the 4th root of that average to get the NP. 203 watts.
Heart Rate Reserve is not used for the calculation, it is only used to plot the data (on the Power page for the whole ride) because plotting %FTP next to %HRReserve makes for a much more linear relationship. Have a look at the video linked under the chart.
Would you mind explaining what you do with the Decoupling numbers from such a ride, and per 2m interval? AFAIK decoupling is used to evaluate aerobic fitness and only makes sense for longer semi-stable rides.
Can you post a screen shot of that last interval zoomed in with the RAW power data chart? Just a couple seconds extra at the start and end compared to your other screenshot and RAW power as a supplementary chart. The difference in NP must come from somewhereā¦
Iām not to keen on directly analyzing your FIT file because I would need to change my settings to correctly visualize it.
Iāll post what you need soon. Just out at the minute.
I donāt do anything with the decoupling data on those short intervals but decoupling data and NP data shown on intervals.icu for all sorts of rides is continually at odds with what training peaks reports. Time after time.
I only ask about that particular 2 min interval because NP on that is clearly miles off. Itās just not 221 watts and like I said in the last post I manually calculated it myself and got 203 watts, same as TP.
No problem, we should be able to sort out the NP difference.
When youāre at it, also post a screenshot of the āInterval Dataā page. Just to evaluate if it is one single interval that is off are all of the rest intervals.
