Running Distance Usually Short?

nkrigg said:

but it is real and is affecting some people. Perhaps I'm unlucky because I sent my first one back and the replacement is still the same.

I'm sure some people have the issue - I would be livid if mine did. Earlier on this thread a user had this issue and got a replacement, and said the problem was fixed.

Note that 0.03 miles is about 45 meters which is WAY under some of the shortness claims here - less than 1%.

What are the results from other users in your Strava feed during measured races? - monitoring this allows us to identify a systemic issue which IMHO it's quite clearly not, or issues unique to you (for whatever reason) - I realise it doesn't help you at all if other devices are measuring fine however.

I don't use strava but I've checked distance recorded on various other watches used by team mates after the event and theirs are usually OK. You're right that 0.03 is low but it's just so annoying that it never says 5k or 10k, you know what I mean. Plus over a marathon it'll be well short, almost 2/10ths.

5135968 said:

I'm sorry but that really is a desperate reach to try to support your hypotheses

It is not a hypothesis but rather a well known fact. For example, see this post from DC Rainmaker who is a well known industry expert:

https://www.dcrainmaker.com/2009/03/racing-line-understanding-how-courses.html

There is also a fact that Garmin does some smoothing to address side to side wobbling of GPS position. I've done distance correction hundreds of times, and after the correction the distance (obtained from the recorded track) is always, 100% of time, longer than the originally reported distance shown on device. It is possible that without smoothing the device distance would be too long, but that depends on two factors - the strength of GPS reception and the number of turns. In general weak reception leads to more multi-path signals being received and more wobbling as a result, and that lengthens the distance. However making a lot of turns or running on a wavy singletrack results in cutting corners or  straightening the path and making the distance shorter, especially when some of GPS updates get dropped. That is a complex problem.

My experience is that Fenix 6 tends to smooth the distance way too much which may or may not be a problem depending on terrain and sky visibility. For me it was a problem most of the time until I upgraded to Fenix 7 which has very significantly improved the distance accuracy. 

silentvoyager said:

It is not a hypothesis but rather a well known fact.

How is 'it' a fact when you literally magicked up some imaginary distance than you imagined runners might be running?

You said 1% difference - over a half marathon of 13.11 miles that is 210 meters, or 0.13 miles.

So I'd expect to see Fenix 6's reporting 13.11 and other watches reporting 13.13 + 0.13 miles, or generally differentials in that order of magnitude over a half marathon. That doesn't happen.

No offence but I could quite easily randomly claim that people who report short distance are cutting corners and therefore not running the distance they think they are.

5135968 said:

You said 1% difference - over a half marathon of 13.11 miles that is 210 meters, or 0.13 miles.

1% is what I observed my Fenix 6X being short compared to the recorded track distance when running on road in my area. Many roads are lined with trees. That doesn't mean that everyone would have the same difference.

I never said that that would be how much extra people would run in a race. If you read the linked article it explains where extra distance comes from. 

And I think the smoothing of the distance affects all Garmin models, just to different degree. I observe that even on Fenix 7, just to a much smaller degree.

And again, the difference depends on tons of factors - that's why it shows for some people and not for others.

I have the same problem. On a known 5k run the Fenix 6 measures about 4.6k. This is a significant error. My iPhone accurately measures close to 5K. My Apple Watch accurately measures close to 5K. And most tellingly, the Fenix measures the course at 5k when it is set to walk. There is unquestionably a significant error with the run setting. 

I have a FR245 and the FIT file distance is always slightly under the GPX file distance. This can be quite annoying if I want consistancy across platforms.

Interestingly Garmins FIT file distance smoothing works quite well when you get a GPS anomoly say from going under a bridge or through a tunnel. The GPX will record a wild point which throws the distance out. The FIT file smoothing copes with this.

That said I work off raw GPX files, and not keen on opaque smoothing algorithms that can change over time.

Nighflame said:

I have a FR245 and the FIT file distance is always slightly under the GPX file distance. This can be quite annoying if I want consistancy across platforms.

Interestingly Garmins FIT file distance smoothing works quite well when you get a GPS anomoly say from going under a bridge or through a tunnel. The GPX will record a wild point which throws the distance out. The FIT file smoothing copes with this.

Old topic, but thanks for your post.

Unfortunately, it doesn't quite fit in here. You have an FR245.

A device which, due to its plastic housing, is inherently more capable of achieving better GPS quality than the device this thread is about. fenix 6. The fenix is made of metal (and obviously has a bad antenna design).

Generally speaking, I can say after a long time (I now own a tactix 7 pro) that the 6 had very poor GPS reception. This was often discussed.

What I have learned is the following:

Garmin uses a mix of GPS data and data from the watch's acceleration sensor to determine distance.

The better the GPS reception quality, the less accelerometer data is used to write the distance to the activity file (.fit).

The poorer the GPS reception, the more data from the accelerometer is used.

Nighflame said:

FIT file distance smoothing

And this seems to have been the problem for some people with the FENIX 6. The accelerometer is apparently "calibrated" during the first few uses of the device based on a number of activities. There is obviously a comparison between the GPS signal and other factors provided by the accelerometer - such as arm movement, steps, etc. - that takes place. This obviously has both advantages and disadvantages.
Disadvantages: if this calibration does not work well due to poor GPS quality, it will later have a negative effect on the distance values.
Advantage: This poor calibration can be reversed by a factory reset of the watch and then performed again with a few subsequent GPS runs/walking activities in good reception conditions.

To summarize everything I have learned so far, I can say the following:
1) the fenix 6 series had poor GPS reception (even within the metal models like the fenix series). Obviously the antenna design was not very good.
2) Due to the poor reception, data from the acceleration sensor was more often included in the calculation of the distance traveled.
3) With poor calibration, this leads to a greater deviation in the recorded distance.

The GPX file contains only the data supplied by the GPS. The FIT contains a mix of GPS data and acceleration sensor data.

The poorer the calibration and GPS reception , the greater the deviation between FIT and GPX.

Nighflame said:

Interestingly Garmins FIT file distance smoothing works quite well when you get a GPS anomoly

Based on my own observations and the discussions in many threads, I would like to say that this statement is NOT GENERALLY true.

It seems to be both, device-dependent (i.e. poor antenna design as with the 6) and dependent on other factors.

---

I have now owned the tactix 7 Pro for over 2 years. I have carried out some tests with the 7 and can say, at least for myself, that the 7 has much better GPS reception than the 6. In my tests, there were no major deviations between FIT and GPX. The recorded distance corresponds both to a route created online and to other recordings of the same activity from other sources such as Komoot on the phone or a GPSmap device.

In terms of recorded distance, I am SIGNIFICANTLY happier with the T7pro than I was with the 6X and tactix Delta.

Here is the data of a recorded walking activity with my tactix 7 pro :
First screenshot original FIT. Second screenshot shows the activity exported as GPX

8,00 km vs. 8,058km (4,97mi vs. 5,007mi)

silentvoyager said:

My experience is that Fenix 6 tends to smooth the distance way too much which may or may not be a problem depending on terrain and sky visibility. For me it was a problem most of the time until I upgraded to Fenix 7 which has very significantly improved the distance accuracy. 

I fully agree with that.