GPS Accuracy

How can he make this assumption if he hasn't tested GPS in other locations?

Basically, this test is ironclad for anyone running this route over and over again and represents n=1. To try and apply universally is poor science at best.

He's one guy trying to be helpful and doing the best that he can. His best is better than most of us can do and he probably is paying to buy the devices to do this for us.

What more do you want?

You could, perhaps, provided a fully scientifically robust method. However I would bet that even Garmin with their VAST resources would not have tested to that level of accuracy (patently not if this thread is anything to go by).

Then you complete the 100s of tests specified in your method and then new firmware is released. then what do you do?

Just take it for what it is.

So I've read all the F3HR thread and the 5/5s/5x about accuracy, so I look at the tracks and yes some look a little off (like a few meters off the side) but I'm wondering what are peoples expectations when they are using this watch? I can post some walks I've had with the F3HR where I call the death valley for GPS but I understood that these tracks were going to be crap but the distance some how was right. So here I am again just wondering what are people wanting from this device that makes it seem so bad yet the 935 that has the same internals as a 5 and gives about the same type of tracks a pass? Are the fenix users just burned so much from the 3hr they want something drastically better or unrealistic or are the FR users just used to the waves and slightly off path track?

So here I am again just wondering what are people wanting from this device that makes it seem so bad yet the 935 that has the same internals as a 5 and gives about the same type of tracks a pass?

I believe the consensus among people who have used both is that the 935 gives better tracks and more accurate pacing than the F5.

He's one guy trying to be helpful and doing the best that he can. His best is better than most of us can do and he probably is paying to buy the devices to do this for us.

What more do you want?

You could, perhaps, provided a fully scientifically robust method. However I would bet that even Garmin with their VAST resources would not have tested to that level of accuracy (patently not if this thread is anything to go by).

Then you complete the 100s of tests specified in your method and then new firmware is released. then what do you do?

Just take it for what it is.

Couldn't have stated it better myself. Jeez, Fellrnr is a guy. Could be any one of us. Just a guy that's collecting a metric crap ton of data points. Are his tests the end all be all? No. Are his tests something that can help people when combined with other data points/reviews/tests? Heck yeah.

From my reading of his site, and updates on either Facebook or Twitter (don't recall which), the only watch he's received for free was the Suunto Spartan Ultra. Other than that, my understanding is he purchases the devices he tests. That's cool in my book.

I believe the consensus among people who have used both is that the 935 gives better tracks and more accurate pacing than the F5.

Not only the FR935, the FR235, FR735, etc...too. It is crucial for runners to have a watch that delivers consistent pace and on the F5 there are some very confusing swings (see the MANY previous posts about this). For hikers (the ORIGINAL target of the Fenix live) this won't matter of course, and the faster you go, the better GPS works so it won't impact cyclists much either. So the problem really is that Garmin is pitching this watch as a "universal" watch when it's clearly not adapted to competitive runners at this time.

I would go so far to say that I can ignore his assessments for the following reasons:


From Fellnr's methodology,

"I ran the same route repeatedly"

I don't run the same route repeatedly and neither do most people. I don't run HIS route and I don't run at HIS location. My watch certainly doesn't use the same satellites that he uses.

"The course is challenging for GPS, with lots of twists, tree cover, power lines, turn arounds and goes under a bridge"

Challenges in one location will be different based a number of factors: latitude, time of day, how you wear the watch, GLONASS, surrounding structures...etc.

"I believe that it's reasonably representative of real-world conditions"

I don't

"probably less challenging than running in the city with skyscrapers."

How can he make this assumption if he hasn't tested GPS in other locations?


Basically, this test is ironclad for anyone running this route over and over again and represents n=1. To try and apply universally is poor science at best.

I agree with one of your observations, and I'll qualify that, that his *Garmin* GPS assessments should be ignored. However, if he starts testing Garmin watches using "every sec," then I believe his tests will be quite valuable, and will not be diminished in any way by your additional points. It seems as though you'd be satisfied only if he tested his watches running next to you on your runs...anything else should be ignored. That's obviously absurd. Yes, he runs the same routes, but ALL of his watches are subject to the same routes. What he loses in variety he more than makes up for in the consistent test conditions. He's making these conditions as consistent as is possible.

IMO his GPS tests are exhaustive and very well done, (the best available) EXCEPT for the fact that he has failed to test the Garmin watches using the every second setting.

He's one guy trying to be helpful and doing the best that he can. His best is better than most of us can do and he probably is paying to buy the devices to do this for us.


What more do you want?

What would I want? I would want some who claims to be an expert to perform tests to support his conclusions. If you're going to claim that you're testing methods are applicable to most people, then I expect your methods support that. The problem is that people quote his studies as gospel, when in fact his assumptions are flawed.

You could, perhaps, provided a fully scientifically robust method. However I would bet that even Garmin with their VAST resources would not have tested to that level of accuracy (patently not if this thread is anything to go by).

Then you complete the 100s of tests specified in your method and then new firmware is released. then what do you do?

If you're going claim that your methods evaluate GPS performance, then yes you need to repeat thsee tests after updates. It's tough being an expert.

Just take it for what it is.

And I do. Just like I stated above, his results are useless for 99% of the things I use my watch for.

So I've read all the F3HR thread and the 5/5s/5x about accuracy, so I look at the tracks and yes some look a little off (like a few meters off the side) but I'm wondering what are peoples expectations when they are using this watch? I can post some walks I've had with the F3HR where I call the death valley for GPS but I understood that these tracks were going to be crap but the distance some how was right. So here I am again just wondering what are people wanting from this device that makes it seem so bad yet the 935 that has the same internals as a 5 and gives about the same type of tracks a pass? Are the fenix users just burned so much from the 3hr they want something drastically better or unrealistic or are the FR users just used to the waves and slightly off path track?

I expect the watch to be at least on par with my 735xt, I expect it to be on par with for example the 935. But I simply see that it isn't... It's not about the track being a few meters off overall, it's about the track jumping around a few meters all the time. This results in incredibly crappy pace.

These are also things easily compared and shown again and again. Perhaps good enough for many, but not for everyone. And while the 935 might have just about the same internals it apparently performs differently because both the pace graphs as the gps tracks are simply better. Again, not a feeling but easy to see even with the eye.

So yeah, for me the F5 did and does not perform as I hoped and expected and so it is going back. And while I haven't had the 935 in posession yet (still waiting on availability) everything I have seen adn read about it shows it's the better performer and looking at it's tracks and graphs it WILL make me happy and delivery everything I hope and expect.

Just to check in. Here is the slower morning run with F5s and Runkeeper. I am quite happy with the results. The accuarcy is on par with FR630 and Suunto Spartan Sport, I used before.

An instant pace tracking can be improved, but if you care about the average pace, while doing intervals, it's pretty good.

http://www.mygpsfiles.com/app/#MttbJleA

For the reference, here is the similar run with Spartan Sport and FR630 (not at the same time, so not completly the same track)

http://www.mygpsfiles.com/app/#MsHoxkDl

This is a run on my local track, I did exactly 19 time on the 3rd track which means 19x415.33=7891.27m and my F5, set on SMART recording and no GLONASS reported 7.89Km which I find it very good. Clear sky no clouds. However, if you check the GPS, you will see that is ok, not perfect, actually SportTracks which recalculates the distance after GPS reported 7.82Km. What I am trying to conclude here is that the watch managed to report the actual distance by using SMART recording. Not bad. Also used a FP, for those who want to see the pace graph.

https://connect.garmin.com/modern/activity/1683976466