GPS accuracy, 6X vs 5X vs real GPS device vs Xiami dual frequency

Conclusion of the test (for those who don't want to click and read):

Conclusion – Garmin fenix 6X GPS Accuracy

Our test shows that wearables like the Garmin fenix 6X Pro are able to deliver extremely good and accurate track recordings that are hard to beat by the best GPS handhelds – under conditions not related to practice, mounted on an antenna plate and located above the body. Wearables are worn on the wrist and therefore have to struggle with a few factors that negatively influence GPS reception.

For outdoor tours, the GPS accuracy offered should be more than sufficient; the deviations of 5 to 10 meters that can usually be observed are quite common even with GPS handheld devices. How much accuracy is needed each user has to decide for himself – a runner has different demands than a hiker!

RISBAC said:

But what was the Garmin Connect distance, that's another story :)

I'm obviously one of the lucky ones (and few?) that get distances I can trust and have since 2015 with my F3:
https://forums.garmin.com/outdoor-recreation/outdoor-recreation/f/fenix-6-series/193098/fenix-6x-sapphire---problem-with-gps-total-distance/1003816#1003816

there are 2 conditions that are very different i think : 

If you run mostly in good conditions for the GPS, and without sharp turns all the time, the Garmins will be fine. The distance calculation is a good fix for GPS errors in those conditions. Otherwise, the distance would be a bit too much. I tested that by retracing my runs in those conditions, and comparing the distance. Then my 5X then 6X are just fine, nothing to complain about.

But when running with tall buildings around, or in valleys, dense forests, with multiple turns, it's another story. I have multiple examples, with another device recording a longer distance, and by also retracing the whole run when possible.

TobiasLj said:

I'm obviously one of the lucky ones (and few?) that get distances I can trust and have since 2015 with my F3:

I don't think we are necessarily lucky or a few - we are not representative of most people who post in these forums only when they have problems. I believe we have more realistic expectations and the watches deliver. I've got runs going back to the FR305 on the same course over the years and all giving results within 20-30m over a 5km loop. I've got results from ultra trail runs of 80km plus that are within a few hundreds of metres of other devices. Very rarely have I had reason to dispute or be annoyed with the distance I've got.

The track on the map? Well at times that's a different matter entirely :-)

I've never had any reason not to trust my distance measurements, either. I had an oddity the other day where the watch was calculating my distance WAY short (showing 0.7 miles when I had actually gone 1.2 miles) for the first half of an activity, but eventually it corrected itself, and the data for the whole activity on Connect was accurate.

My tracks have never been anything short of excellent, I'm very happy with the tracks my 6X records. I use it mostly for hiking. Never had any issues with trees or valleys messing it up more than a few meters. The worst deviation I've ever gotten was by about 50 meters at one spot on one trail one time, and it was the one time I used Galileo instead of Glonass, and I was standing directly under a sheer rock wall in a canyon. So in about 75+ hours of activities I've recorded, the watch has only ever recorded about 5 seconds worth of unacceptable track. I'd call that a success.

I am very happy with my Fenix 6X, but my experience is not the same as yours. My runs are totally fine, but if I am hiking at a slower pace, a 17 mile day hike might jump up to 19 miles. If I’m hiking in more steep switched-back terrain, I’ll see that grow even more, and with a more squiggly track. In canyons I’ve seen tracks that provide a decent approximation of distance, but at times can be worthless to follow. 

Here’s a hike in the Grand Canyon:

connect.garmin.com/.../4362325041

We were hiking slower than I normally would, but there are several places in that track that are worthless. 

I am not complaining. This is a wrist watch, and I’m very amazed at how well it does at this difficult task. I think the solution is appropriate expectations. We must adjust our expectations to meet the limitations of GPS, while using a small, and relatively low powered device that is swinging around every step we take. 

I only see one part of the track that looks bad - in the north-west part where the track crosses over to the other side of the river and back.

The rest of the track looks great. Considering you were in the Grand Canyon, I doubt any other GPS would've done better.

Totally. I think the 6X did admirably. This isn't the only spot the watch wasn't perfect.

There are these as well:

These aren't perfect. Quite frankly they are messy and not a good representation of where I actually was. That being said, this is a wrist watch! I hiked the day before, and the day after without charging my watch! It's doing this by using signals from space! That's amazing! I've hiked enough with my watch to really trust it for navigation, but I understand it's not going to track perfectly 100% of the time. I'm at 74 hours of activities (you beat me) I've found it generally does a pretty good job in these cases:

1. You are moving at least faster than a normal walking pace.
2. You are not against any rock walls.
3. You are not surrounded by very tall buildings.

I've found it performs poorly in these cases:

1. You are moving slowly and/or taking many breaks. (significant over estimation of distance)
2. At least half the sky is blocked out by rock, dense trees or buildings.
3. Hiking on very steep, short switch backs.

For another example of how bad it can be, here's what being surrounded by granite can do:

I see folks complaining on these forums about terrible accuracy from their $1,000 smart watch. Folks buying these watches need to understand where the watch is going to perform well, and where it won't.

You sound more reasonable than a lot of people here ;-) Indeed, what a device weighting 90gr on our wrist can do is quite impressive. I think some are unhappy because older cheaper devices were better for pure GPS performances. But as you said, it's the whole package you need to look at. The battery performances are impressive. I think some would be happy to trade that for a better GPS accuracy though. Which can be understandable. The difference with the best watches isn't huge though, the 6x is doing really fine, especially if you had a 5 model before (the GPS was really really off sometimes). The really annoying part is how Garmin is trying to calculate the distance not just with GPS data. Because it messes with the pace, and for runners it can be a real problem. 

But except that, and even if yes, it's very expensive (just like our mobiles are now...), it's still doing a lot. The maps are great to find alternatives when you are lost or blocked by something. Screen is bigger and very readable. Battery is terrific (almost 3 weeks without a charge !), it's very welcome since we have many devices to charge nowadays. But for sure, most of us don't need 60h of battery for a run. And the Sony GPS chipset are everywhere. So I can understand why some are unhappy because they look at one specific aspect that is very important for them. A 600$+ should do it all, that's the general opinion, and that's understandable.

Garmin should probably over two series of watches, one focusing on battery life, one focusing more of GPS performances (providing Sony has a chipset for that, since Qualcomm has bought the designer of the chipset that was so performant in the Suunto / Polar watches). This way people could choose according to their needs. 

Personally, I'm very happy with the performances. My only complain is the distance / pace calculation which could maybe be improved. But overall, it's still a device doing A LOT. It's my everyday watch and it replaced my Suunto / Samsung / Fitbit combo which was doing a lot too, but annoying to maintain (switch, charge, not all data at the same place...). 

Good discussion.

I agree the GPS tracks are "good enough" for sport use. If I need a hyper-accurate track, I've probably got a differential GPS in a backpack with a mast antenna.

...and the Sony GPSr chip is transitionary.

When a next-gen, L1+L5 band chip becomes practical for inclusion in smartwatches, I suspect we're going to see huge improvements in GPS tracks. (I believe it's the Broadcom BCM47755 used in smartphones - not sure if could be shoe-horned into a watch)