Three to 5% is well within the bounds of what one should reasonably expect from a wrist-worn consumer level GPS device. Did you follow the exact path along which the course was measured? Was the course measured with a calibrated wheel?
Like others, you need to align your expectations with reality.
I understand, but comparing to other runners, using older FR's, an Ambit user and iPhone user, they were +/- 50 metres in comparison, and this is checking different parkruns I do.
Also coming from a Ambit 2 and 3, it was usually very good on distance.
I don't want to invest into another GPS watch if tracking / distance is not improved.
From my own experience with the F3, it seems it cannot handle switch-backs or cornering too well. But when riding / Mountain biking it seems very good.
Perhaps a footpod maybe better option?
I understand, but comparing to other runners, using older FR's, an Ambit user and iPhone user, they were +/- 50 metres in comparison, and this is checking different parkruns I do.
Any observation about accuracy is only of value if there is evidence of a properly measured distance using a calibrated wheel in the first instance. Secondly, a guarantee that the exact path of the wheeled course was followed. Anything else is meaningless and subjective.
And what does this mean?
Also coming from a Ambit 2 and 3, it was usually very good on distance.
3% to 5% is better than very good for these devices. You appear to be suggesting that the Ambit devices were better than that. Please post the tracks and associated data from these runs so that we can make the assessment too.
I can only speak for the FR series that preceded the 935 - 305, 310XT, 910XT and (currently) 920XT. But the distance measurement of these has always been very good in general. I did some research about a couple of years ago using the 910XT on a course measured with a calibrated wheel. The course was ~2.5km with 4 right angled corners and open to the sky so satellite coverage was generally very good. Accuracy was reported to be less than 5m (as indicated on the 910) before each recording. Around 2 laps for each trial, distance recorded averaged 4725.83m with a standard deviation of 41.96m from 60 observations. I don't think you're likely to get better than that on a consistent basis. I am happy to be proved wrong. Please feel free to prove me wrong by supplying objective data to support your viewpoint.
Just because you can occasionally get a GPS recorded distance that is exactly the same as, or within 1%-2% of a calibrated course, does not mean that you will get the same result every time. Using the lone example of what can be achieved tells us nothing about the overall performance on repeated measures of the same course.
A footpod is always a better option provided it is calibrated to pace.
I'm still deciding between a F5 and FR935, I current have an F3.
My main concern is the F3 short-changes the distance I run.
I do a lot of Park Runs all of which are 5kms, and my F3 usually has me around 150 -250 meters short.
After seeing the comparisons here, looks like the FR935 compared to the F5 is also showing shorter distances....
parkruns are "about" 5K. There is no requirement for them to be measured to road race standards and many courses of course aren't on the road anyway.
I have done some that were probably a little short.
If you look on Strava you can usually tell if a course might be slightly "out" based on what everybody else gets using a myriad of devices.
Maybe I'm being a bit harsh, but even still 100 metres difference and we are running the same course.
I guess I just don't want to outlay for a new watch, if the GPS is no better.
Especially checking fellrnr review / comparison on the 5X performance to that of F3 seems a backward step, thus why I was looking at the FR935.
Three to 5% is well within the bounds of what one should reasonably expect from a wrist-worn consumer level GPS device.
I'm sorry but this is quoted regularly, frankly wrong, and minimizes issues that some users may be having. Manufacturers quote much larger error margins then what their devices are capable of. Even the Fenix2 (which I thought was the worst device ever) rarely created errors over 3%.
Even if you don't compare GPS measurement vs course measurement, users are savvy enough to compare identical runs.
While there are some things to teach new users about GPS limitations, good devices are usually have less then 1% error consistently.
This one which is "correct" to 0.1 miles look very very jagged and with an Android phone. I am 99% sure that would be overestimating the actual distance.
I'm sorry but this is quoted regularly, frankly wrong, and minimizes issues that some users may be having. Manufacturers quote much larger error margins then what their devices are capable of. Even the Fenix2 (which I thought was the worst device ever) rarely created errors over 3%.
Even if you don't compare GPS measurement vs course measurement, users are savvy enough to compare identical runs.
While there are some things to teach new users about GPS limitations, good devices are usually have less then 1% error consistently.
I agree. My distances are consistently within a 1% variance, if not better. I'd be really unhappy with a 3% margin of error.
So 1% is 100 metres "out" in a 10K race. I don't recall any Garmin device have used in the last few years being significantly worse than that (and often a lot lot better) unless, of course, the course itself was out!
I agree. My distances are consistently within a 1% variance, if not better. I'd be really unhappy with a 3% margin of error.
Not that this will "settle" any questions here about GPS Accuracy, but it may shed a bit of light on the matter. Check out what the US Government has to say about it. (For those in the US, not sure about European/Asian folk)
