I think I saw old topics showing that there is a difference between XSS shown by Connect IQ data field and final result listed in activities, but thought that this was solved long ago.
But I became suspicious recently when numbers I remember I saw on Garmin were different from what I later saw on the website, so decided to take a screenshot and check.
Turns out that XSS is indeed different - below is a screenshot from Edge 840 taken right before I ended the recording and the values shown on the web site for that same activity - why there is such noticeable difference? Activity was a virtual ride on TPV, dual recorded - fit file was produced by TrainingPeaks Virtual, but from the same source - Kickr Core.
Also - what’s up with extremely low target cadence recommendations? I often see target at mid-upper 6o for endurance rides and find this recommendation ridiculous.
And back to another issue - target cadence. This ride was recommended as endurance, with fairy low power target which is fine, but 67 RPM cadence? You have to be kidding!
When having a breakthrough the values will be different. Xert online shows the xss values based on the new signature. The iq field dont do this and uses the old signature.
Those are reasonable deltas between the Garmin app and XO values.
In your case BTs are a contributing factor.
The numbers will never match exactly since the sec-by-sec analysis on the server is more precise than limited calculations performed on the Connect IQ or EBC apps.
In the vast majority of cases, the ratio achieved (low/high/peak) will match the target Focus for the day.
E.g. 163/11.2/1.55 is Mixed GC Specialist. So is 168/12.6/1.8.
Which sig values changed with your silver BTs?
As @elcomp mentions if BTs are detected during a ride (two on first screenshot; three on the second ride) and confirmed during post-ride analysis, that means the signature you started the ride with is different than the adjusted signature applied in the final analysis.
Ending XSS calcs will differ by design.
XMB cadence values include current cadence in upper left plus a Cadence Optimizer (Bioshift CO) represented by the small arc above current watts. That arc indicates your preferred cadence range at torque. I.e. the rpm values you normally ride at for any given watts.
The mid value (Noon position) is the number shown under TARGET CAD when the Endurance screen is displayed.
With XMB Endurance the goal is Z2 level watts with a default target of 88.5% LTP.
IOW ride easy at low watts which means lower cadence. The idea being your preferred cadence is more efficient but it’s only a suggestion.
I think there’s a difference between “preferred” cadence indoors vs outdoors.
I tend to spin more outdoors. The exception being climbs when I run out of gears.
On occasion I’ll reference the arc to perform low cadence drills staying below my preferred range. That’s done at higher watts though where the goal is to purposely grind on intervals.
I’ll also glance at the CO gauge indoors to remind me to lower my spin during RIBs (rest-in-between). Maintaining a high cadence through rest intervals just generates more heat.
Not exactly - probably, but the should not be that much different.
If breakthrough changes the signature - fine, but the activity was using my previous signature and should display numbers based on that - and only use new signature next time.
As I said - that erodes the trust in the tool because I cannot rely on any numbers - they change whenever they feel like. It is not the first time I notice Xert changes past data - freshness in planner is also a thing in itself (I posted about that before).
Also - signature changes all the time, not just as a result of the breakthrough. Moreover - if numbers in my signature go up, I would expect XSS numbers calculated based on that to go down - but that is not always the case.
Since I can see what Connect IQ app reported in Garmin’s FIT file analysis, I checked some other rides, without breakthrough - and I see similar behavior, for example:
It is very large difference between total XSS here!
As for cadence - no, 60 rpm is NOT my preferred nor normal cadence. I usually ride at ~85 on the trainer and at least 75 (but usually higher) outside. So I have no idea who would, in the right mind, recommend 67 rpm even for endurance ride - this is very bad recommendation overall - 60 rpm is a strength intervals cadence, not endurance or recovery.
That “very large difference”, between 191 and 183.7 (total XSS) is 3.9%. Seems reasonably close to me. We’ve applied some minor skew factors to the data fields, to help bias the data fields to err on the side of slightly under estimating the XSS you generate… since it is better if the server awards slightly more credit, rather than you stop right at your bucket targets and the server awards slightly less and you missed achieving the targets. So, ideally, where there is a couple percent variance, the server is higher.
(Side note - none of the activity charts show MPA data…viewing power data without MPA is missing half the context of the ride)
If you get a breakthrough (or near-breakthrough) on your ride, the XSS has to be recalculated with the new/updated signature, so there will always be some difference in those rides.
Most of the time, it’s rather trivial (within a few percent) and either:
You got a BT ( ) so final XSS is a bit lower (most users are content with achieving a BT)
You got a near-BT ( ), so the final XSS is likely a bit higher (signature goes down a bit, but you get more XSS credit for the ride)
What really bothers me here - the app and site display numbers with one tenth precision but it turns out that precision is fake. I can understand that there might be differences in processing power data online in Connect IQ app and backend may calculate aa bit differently - but I would expect that behavior to be properly documented and precision of displayed numbers aligned with possible variation.
Breakthrough and higher power impacting XSS - sure, but low power produces different XXS too, here are the numbers from my yesterday’s leisure ride (RPE was about 1.5 )
I think you are missing the point I am trying to make - maybe this discrepancy is by design - fine, but then document it properly and stop using these “precise” numbers since that does not make any sense.
BTW, if there is a breakthrough during activity, it should still use the signature which was active when activity started, do not “backdate” the numbers, new signature would be perfect fit for the next activity.
And since Connect IQ app adds the resulting XSS in the FIT file - it may make sense to use it in the analysis - at least display it, otherwise this happens - I remember numbers when I finished the ride, I then go to look at the ride details online and I am met with completely different numbers without any explanation why that happened in the time it took me to park my bike, change and open Xert web site.
) so final XSS is a bit lower (most users are content with achieving a BT)
), so the final XSS is likely a bit higher (signature goes down a bit, but you get more XSS credit for the ride)
)
