Why I'm done with my Fenix 6

TheOtherPhil said:

I’m not conflating anything....if your distance is measured incorrectly over a set time frame, your pace will also be incorrect. Time = distance/ speed. Pace = time/ distance.

Sadly, this is not how it seems to work on Fenix 6. What I see repeatedly is that the distance is acceptably accurate (still a bit short), but the pace is nowhere near the acceptable accuracy. My typical experience is something like this: I keep looking at the watch and every time I look I see the pace in 11-13 min/mile range, then at the end of the mile I see a mile split notification that shows me I've finished that mile in under 10 minutes. I don't know where that pace comes from but it doesn't seem to come from distance over time unless the GPS signal strength is ideal (e.g. a completely unobstructed sky). 

silentvoyager said:

Sadly, this is not how it seems to work on Fenix 6. What I see repeatedly is that the distance is acceptably accurate (still a bit short), but the pace is nowhere near the acceptable accuracy. My typical experience is something like this: I keep looking at the watch and every time I look I see the pace in 11-13 min/mile range, then at the end of the mile I see a mile split notification that shows me I've finished that mile in under 10 minutes. I don't know where that pace comes from but it doesn't seem to come from distance over time unless the GPS signal strength is ideal (e.g. a completely unobstructed sky). 

And I’m saying that it doesn’t matter because GPS will not give you useful instant pace...all the time, every time on every watch.

Here’s a little experiment for you. Turn off auto pause, start an activity recording and stand on the spot for 10mins without moving. Save the file and look at the distance travelled and your max speed...ideally it should be zero meters travelled and zero for your speed. If it isn’t, then that again validates my point. Now magnify that error across poor conditions and a swinging arm.

You can fight the physics of the inherent limitations of GPS.

Then Garmin should acknowledge that and develop a solution. Smooth the pace, offer a rolling average in the options.

Even if you are 100% right with what you are saying. Why are some older watches much better than the Fenix6? Many accept a certain difference, like the older watches had.

First, regarding GPS accuracy of distance. I am comparing my Forerunner 230 against a Stryd recording to another device, and assuming that the Stryd is accurate. Here is yet another run where the FR 230 is better than what you say is possible.

For this run, Stryd is logging distance and pace to my Fenix 5S on my left wrist, Forerunner 230 on my right wrist. 9:41/mile vs 9:38/mile. 3 seconds difference per mile. This was a run partly along a downtown sidewalk crowded by 3-story buildings and partly along a trail with heavy tree canopy (granted no leaf cover this time of year). I don't know or care what conditions Fellrnr was using for his/her testing, these are the actual conditions I run under.

Second, regarding instant pace. You are correct that if a device gets distance wrong then that will translate into errors in pace. I agree that is true for all devices. What I am saying is that somehow the Fenix 6 instant pace is broken beyond just inaccuracy in estimating total distance, in a way that previous Garmin devices were not. Pace = time/distance is mathematically true, but almost certainly not how instant pace is actually computed. Garmin is very likely using some sort of Kalman filter, in all of its devices, and I am saying that it is somehow fundamentally broken in the Fenix 6.

2 min per mile is slower than actual pace is defiantly a fault with the watch as this is not the same across all watches as others have pointed out too. Yes there are times when the pace slows/increases but this are momentarily changes I would expect from direction changes or changes in gps quality but correct quickly.

Comparing my forerunner 220 to 230 to fenix 5s to fenix6s I never noticed any materialistic differences in instant pace (alway poor)  but fenix were always slightly short for me.

For propace and structured workouts I agree with it alway seems to use lap pace hence completely at the beginning.

I have noticed since around firmware version 9? That average pace seems to have a longer averaging probably in response to too much variation from the low power chip to be more similar to the older watches. Is this watch you are/were seeing too?

I personally don’t have any materialistic stuff instant pace once running at the workout pace 3min - 6min per km, faster or slower I do notice far more varieties. 

when using other profiles, walking, trendmill, trail run, bike (even at slow aka run speeds) I find instant pace is better for 6min per km + I wonder if the profile is at fault?

TheOtherPhil said:

Here’s a little experiment for you. Turn off auto pause, start an activity recording and stand on the spot for 10mins without moving. Save the file and look at the distance travelled and your max speed...ideally it should be zero meters travelled and zero for your speed. If it isn’t, then that again validates my point. Now magnify that error across poor conditions and a swinging arm.

I completed my own experiment for the benefit of this thread. My Tactix Delta was placed on the grass with clear view of the sky. I let it acquire a GPS signal and sit for 5 minutes before starting the "Track Me" activity. Alongside it, I placed my GPSMap 60csx which has a large, high gain antenna. I also allowed it to soak for 5 minutes before re-setting all odometer data and then recording the subsequent track/ data.

Whilst sitting in a static position the units recoded the following data:

Tactix Delta: 10 meters travelled with a max speed of 1.7kmh

GPSMap 60csx: 3 meters travelled with a max speed of 1.7kmh.

The reason that these devices recorded any distance at all was because each GPS sample can contain an error and the smaller the antenna, the greater that error can be. In the case of the GPSMap 60csx, it was showing an error of +/-5m and this is per sample at 5-18x per second. Every single point is potentially recording a "movement" even when it is still as the current sampled point could be up to 5m away from the previous, but it isn't often that extreme due to smoothing.

In the case of the Tactix/ Fenix 6, that error is much higher per individual sample due to the limitations of form factor etc.

nonparametric said:

First, regarding GPS accuracy of distance. I am comparing my Forerunner 230 against a Stryd recording to another device, and assuming that the Stryd is accurate. Here is yet another run where the FR 230 is "miraculously" better than what you say is possible.

I am saying that it is possible to track “accurately” given the perfect conditions and a little luck. I am saying that the precision just isn’t there to repeat that feat, every single time with a GPS smart watch. I’ve already invested way too much of my time into this thread but I’ve compiled a small, non-exhaustive list of recent gps watches with pace “inaccuracies”....including the Forerunner 230 that you own.

Fenix 5 inaccurate pace: https://forums.garmin.com/outdoor-recreation/outdoor-recreation/f/fenix-5-series/166659/pace-not-accurate

https://forums.garmin.com/outdoor-recreation/outdoor-recreation/f/fenix-5-plus-series/163569/gps-issues-with-incorrect-pacing-during-runs---fenix-5

Forerunner 35 inaccurate pace: https://forums.garmin.com/sports-fitness/running-multisport/f/forerunner-30-35/237876/forerunner-35-average-pace-problem

Forerunner 230 inaccurate pace: https://forums.garmin.com/sports-fitness/running-multisport/f/forerunner-230-235/129471/ridiculously-inaccurate-pace

Forerunner 645 pace inaccurate: https://forums.garmin.com/sports-fitness/running-multisport/f/forerunner-645-645-m/193899/forerunner-645-music-distance-pace-inaccurate

910xt inaccurate pace: https://www.reddit.com/r/triathlon/comments/2a77fj/garmin_fenix_2_or_forerunner_910xt/

Quote “I recently bought the 910XT after making the same decision you had to make. I love everything about it except it's "instant pace" for running is horribly inaccurate with at least 60 sec/mi fluctuations both directions. I'm trying to see if I can get used to running in HR zones and lap averages, instead.

Vivoactive inaccurate pace: https://forums.garmin.com/sports-fitness/healthandwellness/f/vivoactive-4-series/212202/erratic-pace

Suunto 9 inaccurate pace: https://besthiking.net/gps-accuracy-suunto-ambit-suunto-9/

Quote “After running for several kilometers, I noticed that something is just not right. The Suunto 9 watch showed that I’m running with the pace of 3 min/km which of course wasn’t true – I’m not Usain Bolt.”

TheOtherPhil said:

Suunto 9 inaccurate pace: https://besthiking.net/gps-accuracy-suunto-ambit-suunto-9/

Quote “After running for several kilometers, I noticed that something is just not right. The Suunto 9 watch showed that I’m running with the pace of 3 min/km which of course wasn’t true – I’m not Usain Bolt.”

I ran with Suunto 9 for a year a half before switching to Fenix 6X. While Suunto 9 did have its share of issues that eventually made me switch, the pace wasn't one of them. Yes, the pace on Suunto 9 had a tendency to be a bit fast, perhaps 0:15/mile too fast. But overall the pace was dependable and it was never something that irritated me and made me complain about on Suunto forums (and trust me I was one of the most active members on Suunto forums).

Fenix 6 pace is nothing like that - to me it is a constant source of irritation and in my opinion is the most significant weakness of otherwise decent watch. In you other post you commented that speed = distance / time and that if the distance isn't accurate the speed (pace) can't be accurate either. Well, the problem with Fenix is that pace accuracy and precision is an order of magnitude worse than the distance accuracy and precision. The pace on Fenix is clearly not derived from distance over time, and I don't even understand where it comes from. It is not only unstable resulting is wild swings, but it is also consistently biased on the slow side. I know that because I wrote a Python script that I used to analyze a bunch of my activities. If Garmin simply did nothing and calculated each sample's instant speed and the difference in distance divided by the difference in time, the resulting speed would still be quite unstable, but it at least, when averaged, would be consistent with the distance. But in case of Fenix 6 the averaged instant speed is always (100% of time) slower than overall distance divided by the overall time, which tells me Garmin manipulates the data in some way, but the algorithm is buggy.

Finally, when I raised this issue with Garmin support, they acknowledged it and confirmed that other users have complained about this issue, and that Garmin engineers are working on it. So please stop telling everyone that it is how it is supposed to be and that nothing can be done. Your contribution on this thread isn't helpful.

silentvoyager said:

Your contribution on this thread isn't helpful.

Certain users continued assertions "that their watches are accurate", despite evidence to the contrary is laughable. I have given explanations backed up by data. You have discarded that with your "beliefs". Your Suunto wasn't accurate and precise, it gave you data that you preferred to see in comparison to another device. If your watch cannot measure distance accurately, it will not produce an instant pace accurately, and certainly not within the +/- 15s that you claim....which is a 2.5% error rate at a 10min mile pace.

"The aim of this study was to evaluate the accuracy of the recorded distances of eight sport watches under different real-world environmental conditions for various speeds and reveal the predictors affecting measurement quality. Our results showed that the V800 was the most accurate watch overall, with a systematic error of 3.7 m, a MAPE of 3.2%, and 80.6% of all distance recordings within ±5% of the reference values. Notably, the V800 can use only GPS satellites due to its antenna implementation. Consequently, it is questionable whether the number of satellites and combination of different GPS + GLONASS or Galileo affects measurement quality that much, which was previously questioned [1]. In contrast, other devices showed a significant systematic error of up to –101.0 m and limits of agreements of over ±400 m, overall MAPEs of up to 6.1%, and less than 50% of the data falling within the tolerable range of ±5% (Figure 3). Overall, the recorded distances were underestimated in all watches, and the variance and some outliers were rather high."

"Recent research highlighted the broad use of GNSS-enabled watches in runners of different fitness levels and that users trust the data of such devices [3,4]. However, our study showed, depending on what device was applied, that from 80.6% (29/36) to as little as 44.4% (16/36) of the mean recorded distances fell within ±5% when compared with the reference values. In particular, running over walking and cycling activities were shown to impair the GNSS accuracy in the recorded distances. Nevertheless, the use of all the investigated sport watches can be recommended, especially for distance recordings in an open area. Yet in case of training monitoring and regulation based on recorded distance data, one must be aware that the recorded distances might be underestimated by up to 9%."

References:

Eysenbach, G., Pobiruchin, M., Wiesner, M., Gilgen-Ammann, R., Schweizer, T., & Wyss, T. (2020, June 24). Accuracy of Distance Recordings in Eight Positioning-Enabled Sport Watches: Instrument Validation Study. JMIR mHealth and uHealth, 8(6).doi.org.acs.hcn.com.au/.../17118