Load/tss Zwift and intervals.icu don't match

Hi

I have noticed my Zwift and intervals.icu load/tss don’t match.
Normally it has been 1-2 points in difference - so it hasn’t bothered me.

Yesterday the difference was 10 points.
It is lower here than on Zwift.

I tried to remove any warmup/cool down period and reanalyze the data in intervals.icu
The FTP and other user data seem to be the same here and on Zwift.

Any ideas?

Edit:
Workout from 14th of July shows 171 TSS in Zwift, and 161 Load at intervals.icu
From the 15th of July, it is 113 TSS in Zwift, 106 at intervals.icu

Edit2:
I can see the average speed is taken from the whole workout, including pauses. Why not from moving time? Seems strange when the time shown is moving time, but the speed is from the total time.
Can this influence the Load/TSS score?

Load from Power or from HR?
Load from HR is influenced by moving/elapsed time. From power not because not moving means power equals 0.

From power, though the data from HR is also there and a bit higher on intervals, 110 hr load vs 106 power load.

Zwift should just use power, max hr and ftp is the exact same on intervals.icu and Zwift though.
There is not option for threshold HR on Zwift, so they must use they’re own calculation for that.

@david Perhaps you can take a look at this excel file which I took of 1 Ride.

Data

  1. Sheet1 - StreamsCSV downloaded from Intervals.icu
    Screenshot 2023-07-16 at 10.46.57 AM

  2. Sheet2 - FIT file download. Converted to CSV using GoToes website
    Screenshot 2023-07-16 at 10.47.18 AM

  3. Sheet 3 - Fit File downloaded from Garmin. Converted to CSV using GoToes Website.

Used this guide to calculate NP™
https://medium.com/critical-powers/formulas-from-training-and-racing-with-a-power-meter-2a295c661b46

The results.

  1. Intervals.ICU calc NP™ - 146 NP™
  2. Fit File Download from Intervals.Icu - manually Calc NP™ - 154w
  3. Fit File Download from Garmin - Manually Calc NP™ - 154w

For Load Calc
TSS = ((t * NP * IF) / (FTP * 3600)) * 100

W/ the same IF/FTP/time, the NP™ difference should be the cause of the discrepancy? What do you think?

Dropbox Link to the Excel File.

Tx

edit: There is no manual pause to the ride and the TSS was off by 20tss (lower on Intervals.icu)

Does Zwift report the normalised power for the ride? How does this compare to Intervals.icu? Specifically for the 15th of July ride.

The difference could well be time related. Adding a pause in the middle of a ride and including that in the time for the TSS calc causes TSS to go up. You can test this in the workout builder.

Where does one find load in Zwift?
I’ve looked at my rides and it only shows (Ave/Max) Power, HR, speed, cadence as well as calories, elevation and duration.


oh need to add…
Same file in Garmin and TP shows Load about 1-2tss diff
Intervals.icu shows 20tss lower vs Garmin & TP.

1 Like

I remember that once you complete the ride, it’s there on the summary screen.

(Pic from Internet)

2 Likes

I’ll log in later and see if I can see it, in App. The website doesn’t show it.

Edit: I’ll be honest and admit I have never looked at that screen. Once done, I save and log out and head to TP and Intervals to complete my RPE.

I’ll get minor differences on my Garmin vs the import from Zwift; mostly when there are data drops, which happens occasionally, and is really not that important to worry about.

Based on the import from Zwift to Intervals for the VO2 workout I did (dual recording on Zwift and Garmin), both recording on a Kickr Core (Zwift uses the FE-C profile through BLE and Garmin used the Ant+ profile):

  • Ave. power: 199 (Garmin) vs 198 (Intervals);
  • Load: 87.3 (Garmin) vs 86 (Intervals);
  • Duration: 1:09:06 (Garmin) vs 1:09:07 (Intervals);
  • Normalised Power: 229W (Garmin) vs 227 (Intervals);
  • Heart Rate: identical average and Max
  • Coasting Time: 0s (Garmin) vs 12s (Intervals). The raw power in Intervals shows a drop in power, probably when changing front chain ring in a recovery interval.
1 Like

Typically for an Indoor Ride, the numbers differ by 1-5 or so and I just don’t bother with them.

But for outdoor rides, the diff usually is about 5-10 which I also don’t bother much.
Today it was 20tss diff so I decided to take a look.

Hence the excel.

Anyways, per last info from David about how Intervals.icu may “fill in the blanks” or “disregard come data” (that it determines one is not moving etc) This may likely be the case. (I don’t pause my garmin until end of the ride)

I noticed that (in the excel), the data stream downloaded from Intervals.icu vs decoded from the RAW FIT file, I see there are 7088 points vs 6968 points in the decoded raw fit file.

That likely affects the NP™ calc as well as the load.

1 Like

My workout for today showed load in Zwift of 104 while Garmin was 104.8

What are you comparing your difference with?
The file in Intervals and in another app?

Yes mine is also a bit closer again. I think it’s something todo with some breaks I take on some workouts. At least it seems like the difference is bigger when I take breaks.

Outdoor I haven’t noticed anything that bothers me.

I guess even though the calculation for tss should be the same, different programs read data different and that’s where it happens.

And after reading a bit and testing with other apps, I can see nothing is almost exactly the same. Could be nice though, but again, if the difference is not bigger and you know, it’s not a big problem for me :slight_smile:

yday’s outdoor ride was off by ~2 Load wise. It was also a short 1hr ride w/o stopping and as usual, i never pause/autopause my garmin.

I noticed the diff from the data in intervals vs Garmin. So I also checked TP. Which made me go down this path (in this post here)
:point_down: :point_down: :point_down: :point_down: