GPS Accuracy

As a general rule, I try not to use any comparisons with less than three devices.....and neither are being blocked by the other (i.e. I try to never test on same wrist).

The mind boggles...:o

1% is very solid. There are so many nuances as to why a given unit might measure 1% difference, that you'd actually have to dive into every track and see why one is higher than the other. Without doing that, then saying one is right and the other is wrong is no different than doing the opposite.

As a general rule, I try not to use any comparisons with less than three devices. If less than three devices, then I'd stick purely to looking at tracks (never in map mode, always in satellite mode), and then reference that with what my brain says I actually ran.

Also, you'd want to validate both are using 1-second recording, both are using the same GPS/GLONASS settings, and neither are being blocked by the other (i.e. I try to never test on same wrist).

All good points and that's what's been done here, see the previous posts, bar the third device, but with only two arms it gets tricky not to have one blocked if you do that. Maybe an extra long strap to put it high up on the forerarm ?

Again 1% doesn't sound like a lot except if you've memorized your M pace and on race day you have the choice between :
A. your HR is higher at the same "mile marker" pace -> recipe for disaster
B. you slow down by 3"/km and miss your goal

I'm looking forward to seeing what happens on the track to work with a known distance.

I would say it's different for everyone. Speaking as a mountain trail runner, GPS accuracy is definitely what I'm paying for when I buy a high-end GPS watch. I can go several months and sometimes years between my visits to certain trails/trail races, and I use my tracks to re-orientate myself. Most trails and trail race courses present their own unique challenges and hazards. My GPS tracks (and the pace/HR data connected to them) are my record of an adventure and a learning tool for future races. I'm not trying to go on a diatribe, just pointing out that spotty GPS will destroy your tracks on a trail run with lots switch-backs and elevation changes. It sounds like people are having more success on successive runs so it's not really an issue anyways. I've always trusted Garmin to improve their hardware and software.

I agree if the gps is bad for someone then they should get it replaced , but some of these posts over the years comparing gps tracks where it's off by marginal metres is ridiculous. I fish with higher end gps systems than my Fenix and the accuracy variances are no different , we don't have access to milItary gps.

All good points and that's what's been done here, see the previous posts, bar the third device, but with only two arms it gets tricky not to have one blocked if you do that. Maybe an extra long strap to put it high up on the forerarm ?

Again 1% doesn't sound like a lot except if you've memorized your M pace and on race day you have the choice between :
A. your HR is higher at the same "mile marker" pace -> recipe for disaster
B. you slow down by 3"/km and miss your goal

I'm looking forward to seeing what happens on the track to work with a known distance.

For separating on GPS, the 'best' thing I've found thus far is some extender straps that Garmin makes for watches to put over coats - it keeps the watches apart. But then that can dork with optical HR. So these days I'll just attach other GPS units either to the shoulder straps of a CamelBak, or I'll hand-hold them out front.

I do agree that 1% can mess with pacing for a marathon. But I disagree that +/-1% is a failure of GPS. In testing a lot of devices, I'd say that +/- .5-1% difference between units is pretty much the norm for wearables for running/hiking. You can get a bit better in cycling scenarios, and a bit worse in openwater swimming.

Again, in order to know which device is incorrect - you'd really need to specifically analyze the entire route and see where it went off.

Depending on GPS only for pacing is a recipe for disaster in any race. Best bet is to reset your watch with the lap button if you notice your auto-lap drifting off, or better yet turn off auto-lap and hit the mile markers (though sometimes the mile markers can be off too)

Depending on GPS only for pacing is a recipe for disaster in any race. Best bet is to reset your watch with the lap button if you notice your auto-lap drifting off, or better yet turn off auto-lap and hit the mile markers (though sometimes the mile markers can be off too)

I agree - if pace is that important, one need to do the above.

Depending on GPS only for pacing is a recipe for disaster in any race. Best bet is to reset your watch with the lap button if you notice your auto-lap drifting off, or better yet turn off auto-lap and hit the mile markers (though sometimes the mile markers can be off too)

Obviously (and there are even CIQ/DFs that can auto-correct the distance and time ahead/behind) but you're missing the point, if you've trained for weeks at 4'30"/km on the open roads with your F5 and that translates to 4'33"/km during the race using actual/"mile marker" pace you have the two options I mentioned above, neither satisfactory.

GPS vs GPS+GLONASS?

I saw some weird behaviour this morning- I ran on the exact same path through the park this morning during both laps (every second recording on a 5X)

https://connect.garmin.com/modern/activity/1631869175

Obviously (and there are even CIQ/DFs that can auto-correct the distance and time ahead/behind) but you're missing the point, if you've trained for weeks at 4'30"/km on the open roads with your F5 and that translates to 4'33"/km during the race using actual/"mile marker" pace you have the two options I mentioned above, neither satisfactory.

I get your point, but with all due respect; the difference between 5:05 and 5:02 in km/h is 0,12 km/h. For 4:05 and 4:02 it's 0,18 km/h. I doubt that even anyone can run that precise. Not even the GPS resolution (or a footpod with 99% accuracy) can cover that. So it's kinda more a hyperthetic question than realistic.

Use mile/km markers or train on a track if pace is that important. But also note no one runs like a robot ;)

Right, if you're jogging a 3"/km is meaningless, however 3"/km is a 2' difference at the end of a marathon and I'm not going to run 16k at MPace on a track because I can't trust the pace of my GPS watch ;-)

GPS vs GPS+GLONASS?

I saw some weird behaviour this morning- I ran on the exact same path through the park this morning during both laps (every second recording on a 5X)

https://connect.garmin.com/modern/activity/1631869175

I've also seen unexplained variances between laps or out/back, what I called the "wobbly" tracks.

Was that GPS or GPS+GLONASS? GPS+GLONASS uses more power and is going to be useful in challenging environment like urban canyons because the watch will have more potential satellites to work with so it *might* improve accuracy. Outside that type of environment the GPS constellation will have more than enough satellites available to get the maximum potential accuracy given the degraded "environment" of a watch (tiny antenna, RF interference, body obstructions, etc...).