Related
This discussion has been locked.
You can no longer post new replies to this discussion. If you have a question you can start a new discussion

BUG: Pace & GAP values are inconsistent with other activity metrics, are dangerously broken during suggested workout

ZombieSloth
over 1 year ago

I have started using the suggested workouts for my runs as of a few weeks ago, and I set them to use heart rate targets as opposed to pace. I also have an HRM-Pro Plus that I use during workouts.

My predicted max HR is 186, and my LTHR is 167 as measured by the HRM-Pro Plus. My HR Zones are based on LTHR%, using the default percentage bins from the device. As of now, a base run has a target of 139bpm, and currently averages out to ~9:00-9:30/mile over a 30-40 minute base run.

Today I got the first anaerobic assignment. It was 15:00 @ 139bpm, then 7 rounds of 1:00 @ 6:34/mi + 3:00 @ 126bpm. Simple enough I figured, albeit a little odd that there is a pace based target in there since I set it to use HR. When I was doing the fast intervals, it quickly stopped making any sense to me. I was in nearly all-out sprints, HR >175bpm, and the little indicator screen telling me that I was running slower than 7 minute miles for a significant portion of the time. Based on prior experience, I know that I am moving faster than that at that effort level.

Here is the pace field for the first 4 sprints, showing my fastest reading was 6:09.

Here is pace and GAP. GAP looks like it would average out to about the same as pace over the intervals (there were some little underpasses on the trail which I sprinted through).

Here are the stats for the intervals, which do seem to have correct average pace readings. The Avg GAP values look totally bogus, and kind of look like what the workout screen may have been showing me.

This seems completely broken. At best it is obnoxious, and at worst someone is going to get hurt trying to hit a broken real-time metric. Do I have something incorrectly configured on my watch, am I misunderstanding the metrics, or is this super broken right now? I would be very happy if this was just user error and I could fix it with some setting so any input there is welcome.

If this is a bug, then I will do anaerobic workout days in a DIY style based on HR. Does anyone have a suggestion for a target HR to go for in the fast intervals?

Thanks! Other than this, I would say that the watch has been great overall.

  • 0 over 1 year ago

    There may be 2 things going on:

    Firstly, my experience has been that the watch struggles in real-time with pace during intervals that short. It takes the watch a little while to accurately display the pace when you jump into a sprint. It usually catches up towards the end of the interval. As a result, I try to focus more on RPE for short intervals like this. 

    Secondly, the big difference between pace and grade adjusted pace suggest that you were either running in a hilly area or there’s an issue with your watch’s barometer. Try cleaning out the barometer hole on the back and see if that helps. 

  • 0 over 1 year ago in reply to jambongris

    Thanks, I will check the barometer opening. The run is on an overall very flat paved foot path, and the elevation plot values all look correct. There are some small underpasses that go under streets, but all of the 1:00 intervals spanned both the downward and upward ramps. I note that in the pace plot, the pace and GAP values probably average out to almost identical values (area under curve looks about equal), but the final stats in the intervals is really different.

    Plots of values recorded during activity

    - Pace: ~7:00 average (eyeballing it)

    - GAP: ~7:00 average (eyeballing it)

    Post-run summary stats for intervals

    - Avg Pace: sub-6:00 (all matched perfectly to the distance covered in 60 seconds)

    - Avg GAP: ~7:00 average, seems to match what is shown in the plots

    It appears to me that the average pace for the intervals is a post-processed value computed from [actual distance / actual time], whereas the average GAP is computed from [mean GAP readings during the activity].

    I am going to pull the TCX file and import the data into Excel or Matlab or something to try to determine how exactly the values are being computed; I am not sure how responsive Garmin is to issues like this, but I'll try to post as much detail as possible in the off chance that there is an opportunity to improve functionality.

    The one thing that I 100% think that Garmin needs to change is the workout suggestions. If a user sets them to target HR values/zones, then the workouts shouldn't have pace targets (doubly so if pace is known to be sort of an unreliable metric in general, and particularly during short intervals like this).

    The good news is that the workout still got me TE values of 3.4 / 3.6, so other than the UI being a bit confusing, the result was still a little productive.

  • +1 over 1 year ago

    OK, I think that I figured out what the problem is. Call it 80% user error on my part, and 20% poor documentation by Garmin.

    The issue is my HRM Pro Plus strap. I had set it to be the source for pace measurements immediately after pairing it with my 955. This is not the correct thing to do as the strap will not auto-calibrate in this case. So, I am 99% sure that this is why I have been getting really bad pace readings for most of the month, and why I would get really weird jumps in the reported pace value on steep descents as my cadence and gait changed.

    I guess that a silver lining is that I got to put together a fun little TCX / FIT analysis script so I can play around with the recorded data when I want to now. Here are some findings from two runs on the exact same path, but a couple of weeks apart and without/with the HRM Pro strap.

    This run was with no strap / FR955 doing all of the pace calculation per its built-in functionality.

    Blue line is the pace reported in the TCX (trackpoint > extensions > ns3_TPX > Speed) and presumably what is shown on the watch face.

    Orange line is recomputed directly from the GPS data (total recomputed distance is within 1% of the recorded total). Since the raw GPS-pace values are really noisy, I applied a 10 second moving average window to it to make the plot legible.

    The two series track well, and their mean pace values are within 7 seconds of one another. The mean of the blue pace series is 9 seconds slower than the average value reported in connect (which should always be correct as it is [total moving time / total distance]), and the mean of the orange line is 2 seconds slower.

    Here is a more recent run on the same exact path, but with the strap which is (apparently!) not properly calibrated but set to measure the pace.

    Blue and orange lines represent the same parameters as above.

    There is about a 1:00 difference between the means of the two, with the orange line's mean being much closer value to the "true" average in Connect. Anyway, I think that my issue is with the strap, which I have disabled for pace measurements for now.

    As for my original post's run during the anaerobic workout, I think that the huge discrepancy during the fast intervals makes more sense now. As I switched from a jog to a sprint, the uncalibrated HRM strap accelerometer data was probably going bananas. I have another anaerobic assignment on Friday, which I will do with the FR955 calculating the pace this time, and I am optimistic that it'll be a better experience.

    Just in case anyone else runs into this, you can see the instructions for HRM Pro calibration here. Basically, set the pace measurement function to Indoor or Off, and then follow the procedure for manual or automatic calibration. I am going to go with the automatic method since I do not run indoors ever, and I'll report back once it's complete and I can run this same route again with it measuring pace. TBH I my just leave the pace feature off if it looks like the watch does a better job with its own GPS+accelerometer when outdoors, but it'll be fun to run the comparison "for science."

    Cal instructions.
    https://support.garmin.com/en-US/?faq=8B0cpkix4E8kJHHhs2OzY6

     FYI the HRM calibration process is not mentioned anywhere in the owner's manual for the HRM Pro Plus product. It seems like something that should be there since it led to some confusion and frustration. If I am interpreting the cal instructions properly, it seems like the HRM straps should require an initial calibration to be performed before they can be enabled for pace measurement. Whatever the case, it sounds like calibration MUST be performed by the user first if good results are desired.

    https://www8.garmin.com/manuals/webhelp/GUID-57B75051-8E96-44B8-A89E-470B3E3BCD32/EN-US/GUID-0972DBF1-06B1-4D35-85DE-FAE22AA67CEB-homepage.html

  • 0 over 1 year ago

    It is correct that anaerobic runs specify pace and not heart rate. The  reason is simple: your heart rate takes too long to get to the right rate and you won't be able to run at the right pace