PDA

View Full Version : 310XT Accuracy?



EREBER
08-09-2009, 03:06 PM
I am pretty sure that the 310XT is very accurate, however, I just did the Georgetown to Idaho Springs Half Marathon and my 310XT indicated that I ran 13.28 miles, when the entire distance should have been 13.10 miles. This event was not USATF certified, but that seems like a fairly decent discrepancy. Has anyone else had a similar experience, or an experience that was not like mine?

Thanks,

Eric

DW152
08-09-2009, 03:18 PM
That's about a 1.4 percent difference. I looked over some of the races I've done with my FR305 and 310, and your difference is a little on the high side compared to my results. My largest discrepancy was the Army 10-miler, which I did with my 305, and there was a difference of 1.2 percent, when post-processed with Ascent. However, you do note that your course was not USATF certified, so it's hard to know how accurately your course was measured. Also, it would be helpful to know what your satellite accuracy was during the race (if you had this displayed). Other factors that can affect this comparison was whether or not the race was in the open or under trees or near buildings, and if the course was straight or had a lot of turns.

For a more detailed discussion on 310 accuracy, check out this thread:

https://forums.garmin.com/showpost.php?p=1280&postcount=20

JOHN313
08-09-2009, 05:28 PM
Besides the degree of inaccuracy in the device, on longer races, it's not uncommon to run longer than the official course calls for. The course is measured on the tangent but individual runners often do not run the entire course on the tangent. They may be crowded off of the tangent lines, stick to a particular lane of the road, move from side to side to go around obstacles or catch a water stop, those little variations add up.

MIMISCHRISTOP
08-10-2009, 05:14 AM
Besides the degree of inaccuracy in the device, on longer races, it's not uncommon to run longer than the official course calls for. The course is measured on the tangent but individual runners often do not run the entire course on the tangent. They may be crowded off of the tangent lines, stick to a particular lane of the road, move from side to side to go around obstacles or catch a water stop, those little variations add up.

nice point, enough hair splitting already!

EREBER
08-10-2009, 05:52 AM
Thanks for the reply - that definitely helps. The race was in a canyon, so it wouldn't surprise me if the accuracy was a little off, but 0.18 miles still seems high. In the events that you have done, does the distance indicated by your Garmin vary around the measured distance (sometimes high, sometimes low), or is it always long?

DW152
08-10-2009, 06:23 AM
Thanks for the reply - that definitely helps. The race was in a canyon, so it wouldn't surprise me if the accuracy was a little off, but 0.18 miles still seems high. In the events that you have done, does the distance indicated by your Garmin vary around the measured distance (sometimes high, sometimes low), or is it always long?

I think it's always been equal to or slighly higher than the advertised distance. I don't think it's ever been shorter. This is reassuring, since USATF certification actually requires the course to err on the high side so that it is at least the advertised distance. Also, as another poster pointed out, when I do a race, I'm in the middle of the pack, so I won't be able to run the tangents and the shortest distance possible.

ALAALA
08-10-2009, 10:35 AM
Generally, my 305s and 310 measure short (that is, in nice-ish GPS envos). Unless in urban canyons where signal bouncing on metal and glass might cause "trackpoint excursions". And on tracks, where it seems that they'd rather "forget to turn", than cut corners. A have noticed, though, that the 310 lays a better "track track" than the 305. BUT it still measures with about the same error. :confused:

In the woods, they do tend to measure relatively much shorter than nominal. About 0.7% short in average in my experience.

And again, in nice-ish GPS envos they tend to measure a bit short (cut corners).

I always think of the Stockholm Marathon as 42.6 km due to the difficulty to follow the shortest path, especially around drink and food stations. Not to forget the showers (three years in a row we've had ~30C...). That's a little less than 10 meters added every kilometer. Or a little less than 1 meter added every 100 meters.

KAFOLSKE
08-10-2009, 04:02 PM
...., but 0.18 miles still seems high. .....


1) How do you know you followed the exact same route they used to measure the course?
2) How do you know they accurately measured the course?

Given 1 and 2, I don't think you can draw any conclusions about your watch's accuracy.

R.MCLACHLAN
08-17-2009, 03:28 AM
I have noted a few times that the recorded 310xt distance is different from that in SportTracks , Ascent, and rubiTracks. Apparently these software packages recalculate the distance from the dataset to give a more accurate track distance. Interestingly Garmin Connect does not do this. e.g recent 5Km run measured 4.99 on the 310xt & GC, but in SportTracks and others it was revised to 4.93. Admittedly a very small difference of approx 1%, but derived from the same dataset nonetheless.

NDURANCERIDER
08-17-2009, 09:31 AM
MotionBased recalculated distance from the data points stored rather than report what the unit reported. Similar to the other 3rd party applications.

One of the questions often asked at MB was "why doesn't the MB distance match my watch." For Garmin Connect, I guess they decided to try to minimize that worry by reporting what the watch says.

I can see arguments for both ways of doing it.

Initially MB had to calculate based on the trackpoints because the non-fitness units didn't (don't) report distance, only basic GPX trackpoint information. Those non-fitness units will be supported in Garmin Connect later this year, so it will be interesting to see if fitness unit users will be able to choose the non-fitness algorithm or use the unit-reported information.

April

SEILOGRAMP
08-17-2009, 09:37 AM
One of the questions often asked at MB was "why doesn't the MB distance match my watch." For Garmin Connect, I guess they decided to try to minimize that worry by reporting what the watch says.

It would be nice if the user had the choice. Perhaps the default should be to report exactly what the watch says. For users that understand it, they may want to have the data corrected by GC.

ALAALA
08-19-2009, 09:51 AM
Generally, my 305s and 310 measure short (that is, in nice-ish GPS envos). Unless in urban canyons where signal bouncing on metal and glass might cause "trackpoint excursions". And on tracks, where it seems that they'd rather "forget to turn", than cut corners. A have noticed, though, that the 310 lays a better "track track" than the 305. BUT it still measures with about the same error. :confused:

In the woods, they do tend to measure relatively much shorter than nominal. About 0.7% short in average in my experience.

And again, in nice-ish GPS envos they tend to measure a bit short (cut corners).

I always think of the Stockholm Marathon as 42.6 km due to the difficulty to follow the shortest path, especially around drink and food stations. Not to forget the showers (three years in a row we've had ~30C...). That's a little less than 10 meters added every kilometer. Or a little less than 1 meter added every 100 meters.

I'd like to update a bit on the distance measuring part. Having some more
experience now with the 310 I'd say it does tend to overshoot in general
(the notable exception being trail runs, but there it doesn't really matter, since the Forerunners generally undershoot in the woods, so it might just
show better values there, less undershooting so to speak).

The 310 measures track work longer than the 305. Way too long: 20-30 meters per kilometer, whereas the 305 generally kept the difference at 10
meters per kilometer. Of course 0 meters would be the best, but it shouldn't
have to be worse than the 305.

But also, the 310 measures typical street runs in nice-ish GPS envos a bit
longer than the 305, perhaps 15-25 meters per 10 kilometers. This might
seem like minor problems to most, but why should the 310 be worse than
the 305? That is just plain unnecessary, to me.

SEILOGRAMP
08-19-2009, 11:25 AM
But also, the 310 measures typical street runs in nice-ish GPS envos a bit longer than the 305, perhaps 15-25 meters per 10 kilometers. This might seem like minor problems to most, but why should the 310 be worse than the 305? That is just plain unnecessary, to me.

Agreed. Maybe (hoping) these issues can be resolved in an update.

ALAALA
08-19-2009, 12:13 PM
Agreed. Maybe (hoping) these issues can be resolved in an update.

Perhaps they could introduce user selectable modes, like so:
- Track Mode (less aggressive calculations, tend to choose to undershoot)
- Normal Mode (305 style)
- Trail Mode (more aggressive calculations, tend to choose to overshoot)

Track Mode would possibly also be nice when running in urban canyons, which tends to lead to overshooting.

KAMIKAZE_RUNNER
08-25-2009, 08:45 PM
I have found all my Garmins* to be very accurate on distance. I am speaking of the Training Center mileage and not MB or GC. I run marathons and have NEVER had a 26.2 mile race. The closest I came was 26.3 and that was because the race had a blue line marking the course. It is so easy to mess up on distance...take the turns too wide, weave from side to side, don't run tangents . . . any number of possibilities.

My money would be on Garmin being accurate.

Jim

*One 201, one 301, two of the 205, and now a 310XT. They die after a year so I figure it is the cost of being active, $1 a day in long-term depreciation! :p

ALAALA
08-26-2009, 08:27 AM
I don't know about you, but I'm talking about wheel measured and/or calibrated bike measured and/or corrected GPS tracks in ST (often against known measured distances) and/or track runs (which are certain to be correct). There is no doubt that there is a consistency in all four Forerunners I've used (2x305, 1x405, 1x310); they all measure long on tracks (and in cities), short on trails, and surprisingly exact on roads having few GPS disturbing parts. I have experience from more than 1000 runs with Forerunners.

CORHAA
10-14-2009, 01:33 PM
Yesterday for the first time used a 310 on one wrist, and a 305 on the other. According to the 310 I ran 9.69 km (6,06 miles) and the 305 showed me 9.49 km (5,93 miles). The run was mostly in the forest. That is a big difference, and which one should be more accurate? You can't tell. With the 305 I have ran an officially measured 10km road race, and it showed me then 10,1 km. So I would say, that the 310 is less accurate. Next time I will do the same, but I will wear them on the wrist on the other side. Maybe, that arm-movements will have influence.

RVDOWNING
10-14-2009, 01:49 PM
I would expect the 310XT to have poorer measurements than the 305. With the 305 worn conventionally on the left wrist, the antenna is pointing up toward the sky. The antenna is in that extended lip portion of the watch. With the 310XT, the face of the watch is pointing over to the left, horizontal to the ground. No satellites in that direction.

DW152
10-14-2009, 03:36 PM
I would expect the 310XT to have poorer measurements than the 305. With the 305 worn conventionally on the left wrist, the antenna is pointing up toward the sky. The antenna is in that extended lip portion of the watch. With the 310XT, the face of the watch is pointing over to the left, horizontal to the ground. No satellites in that direction.

How do you know that the antenna is on the face of the watch? It could still be below the face. I noticed that the GPS accuracy on my 310 improves if I tilt it so that the bottom is pointing upwards, like it would be when on my wrist and when my arm is in a normal running position. This leads me to believe that the antenna is still below the face.

Besides the antenna position, I think there are other factors that need to be considered when comparing the 305 to 310 accuracy. ALAALA and SEILOGRAMP did quite a few comparisons when those two first started using their 310s.

EKUTTER
10-14-2009, 09:23 PM
Yesterday for the first time used a 310 on one wrist, and a 305 on the other. According to the 310 I ran 9.69 km (6,06 miles) and the 305 showed me 9.49 km (5,93 miles). The run was mostly in the forest. That is a big difference, and which one should be more accurate? You can't tell. With the 305 I have ran an officially measured 10km road race, and it showed me then 10,1 km. So I would say, that the 310 is less accurate. Next time I will do the same, but I will wear them on the wrist on the other side. Maybe, that arm-movements will have influence.
When running on windy trails, especially with tree cover, my distance almost always is short as compared to a wired bike computer. So I suspect the 310 was more accurate. On the road, the 310 is remarkably accurate and consistent. On windy trails, my 310 often comes closer to matching my wired bike distance than my Edge 705, which I would have expected to be better with more room for the antenna. On the road, all 3 tend to come to within 0.25% of each other.

CORHAA
10-16-2009, 02:38 PM
I would expect the 310XT to have poorer measurements than the 305. With the 305 worn conventionally on the left wrist, the antenna is pointing up toward the sky. The antenna is in that extended lip portion of the watch. With the 310XT, the face of the watch is pointing over to the left, horizontal to the ground. No satellites in that direction.

It does not make any difference, on which wrist I wear the forerunner; its antenna is allways pointing upward. But maybe there is a difference in swing of my arms. The forerunner itself is not moving at the same pace as your body. When swinging forward it goes faster, and when swinging backward it goes slower. And, when the swing is a little bit longer on the right, than on the left, than it could measure a few centimeters difference in distance. It would be a nice experiment to run with the forerunner on your wrist, and then the same route with it on your body somewhere. Why don't we try, and report here?

CORHAA
10-18-2009, 06:29 AM
Today I have put both forerunners on my car's dashboard and drove for 26,7 k. At my arrival they showed the same distance. Then I ran a 10k road race with my 310. It measured the same distance as the race: 9,99k. So we may conclude, that there is no difference in accuracy between the 305 and 310, and that your armswing is influencing the distance.

DGKREKELER
11-28-2009, 04:39 PM
I too recently upgraded from a 305 to a 310XT and have noticed that the distance is off by about 2% Last week I did the JFK50 and it registered 51.24 miles. I talked to a couple of other people that had 310XTs and they also showed over 51 miles. I thought that maybe the 310XT was so accurate that my arm swing was influencing the distance but I did the dashboard test in the car and it showed the same 2% error. The elevation gain/loss seems way off as well. On the tow path (which is flat) for mile 26 it shows 367 ft. of elevation gain and 448 ft of elevation loss. Could this influence distance? It looks like this has been a problem for a while, I would expect Garmin to fix the software in the near future.

JFK50:
http://connect.garmin.com/dashboard?cid=81459
http://connect.garmin.com/splits/19480535

(this is my first post so I'm not sure it this is the correct way to include a run)

DW152
11-28-2009, 05:40 PM
As far as the elevation goes, it's widely known that the Forerunners don't provide good elevation data. I ran a couple of miles on the C&O canal this morning, and it recorded something like 150 feet of elevation gain just in this stretch.

As far as your JFK experience goes, I will point out that ultras are not USATF certified, and rarely claim to be accurate. I did JFK a while ago, and I believe the starting and ending points are historical, dating back to the first time the race the race was held in the early 1960s.

EKUTTER
11-28-2009, 05:53 PM
The elevation gain is not fixable as long as they use GPS instead of barometric pressure for the altitude. That is why they don't let you select "ascent" as a field on the 305/310xt as they do on the Edge 305/705, which use BP. For the 305/310xt this value is being computed by Training Center or what ever software you are using to analyze your file. There are many threads that discuss how elevation impacts distance since all the GPS devices I am aware of ignore elevation for distance. In most cases, it is very minimal, a fraction of a percent.

CHRISDSIMS
12-16-2009, 10:14 AM
My 310XT froze last week so I did a hard reset. Prior to the reset, the distance measurement during my runs was very accurate. After the reset, it is way off--I did a certified marathon last Sunday and it measured 26.69--each miles was pretty consistently off .02. Any ideas?

DW152
12-16-2009, 12:10 PM
My 310XT froze last week so I did a hard reset. Prior to the reset, the distance measurement during my runs was very accurate. After the reset, it is way off--I did a certified marathon last Sunday and it measured 26.69--each miles was pretty consistently off .02. Any ideas?

It could have been on one-time thing. You should see if it continues to happen. It also could have been that the course was long, although an entire half-mile seems a little extreme. Also note that unless you are an elite runner with complete freedom to choose your path, you will not be able to run the tangents and will run a longer distance than the certified distance. Again, I think 0.5 miles is a little extreme and I don't think you really ran that much extra distance. Finally, look at your track and see if your 310 ever lost lock on the GPS satellites and gave you some screwy sections.

JASONANDREWS
01-09-2010, 03:13 PM
I just completed the Disney Half Marathon, both my friend and I have 310XT and we both had a distance of close to 13:28 (mine was 13.29, my friend's was about the same)...it seems to be a constant difference..

TRYING2GOFAST
01-14-2010, 09:33 AM
A certified race distance is the shortest distance around the course. Unless you're sure that you nailed every single corner and never drifted around the straightaways, to and from aid stations, etc., your GPS is probably accurate.

JANTILA2
08-02-2010, 05:15 AM
Let's raise this thread up again.

I compared the 310XT against a professional GPS test device used in vehicle industry and testing.

The 310XT showed constantly 0,8-1,0% higher distances. In some cases (straight, open course) the 310XT was only 0,2-0,3% longer.

I guess this has something to do with the software: how to calculate the distance in turns and/or poor satellite visibility.

Software 2.90 seems to be better. The previous versions calculated even longer.

DARTONP
08-02-2010, 06:02 AM
On garmin connect does the red line accurately follow the route?

I have found if you turn the 310 on and use straight away the first half a mile points are not in the correct place. I have started turning the 310 on 10min before using. I now give the 310 enough time to get it fully positioned.

DW152
08-02-2010, 11:10 AM
The 310 (and all GPS units) take a while to figure out where you are and where the satellites are. Your position accuracy improves as it converges on a solution.

It's standard practice to turn the device on for several minutes before you start your activity. Try not to move around (or start your run) too much. At a minimum, you should wait for the "acquiring satellites" bar to go completely across before you start. I often leave mine outside while I'm getting dressed or stretching.

WADIASOFT
08-06-2010, 11:53 AM
310 6.30 miles
405 6.22 miles

310 7.34 miles
405 7.19 miles

310 10.16 miles
405 10.03 miles

One watch on each wrist. Clear view of the sky. GPS accuracy in the mid-20 feet for each watch. I have worn the 405 during many marathons and it always goes long. This is expected due to factors such as tangents and measuring the course a little long for margin of error, etc. Based on that I do not think my 405 is coming up short. I think my 310 is going long.

MWRINCH
09-09-2010, 12:30 AM
I have the same problem I was kayak racing (10km) on the ocean and the 310xt said I made an elevation gain of 350 meters !!! there was an immediate jump near the start line and then it was never able to correct its self. How come it could not do that on the map overlay?

The linear accuracy is probably the result of many little integration errors over the length of a race.

DW152
09-09-2010, 10:44 AM
I have the same problem I was kayak racing (10km) on the ocean and the 310xt said I made an elevation gain of 350 meters !!!

It's been well documented here that Forerunners do not have a barometric altimeter, and only use GPS to provide elevation readings. This is very inaccurate. Similar to your kayak case, I've run along the beach and have had large elevation gains, at least according to my Forerunner.

DARTONP
09-09-2010, 02:39 PM
I've been using my 310XT doing allot of sea swimming this year, before the 2.90 firmware update for swimming the distance would have been double what it is recording now. Even now it still can be about 5-8% out.

What I am trying to explain now think about this from non swimming. In swimming the 310XT keeps loosing satellites when under water and then gain them again. What you end up with is a zig zag line. When running you get a similar effect when running next to buildings, trees etc. In running this zig zag line will be allot less but still has an effect. Garmin do their best to correct this but it will never be perfect.

To get an almost perfect recording you will have to run on a flat course with no objects blocking the satellites. In reality this will rarely happen. So what can we do? Not allot but things like foot pods will help.

Remember course measurers don't use GPS but use reliable measuring tools like measuring wheels etc.

DW152
09-09-2010, 02:55 PM
I've been using my 310XT doing allot of sea swimming this year, before the 2.90 firmware update for swimming the distance would have been double what it is recording now. Even now it still can be about 5-8% out.

https://forums.garmin.com/showthread.php?t=903


What I am trying to explain now think about this from non swimming. In swimming the 310XT keeps loosing satellites when under water and then gain them again. What you end up with is a zig zag line. When running you get a similar effect when running next to buildings, trees etc. In running this zig zag line will be allot less but still has an effect. Garmin do their best to correct this but it will never be perfect.

In practice, the accuracy is within a couple of percent. For most people, this is plenty good enough. I've been running for two decades now, and being able to wear something on my wrist that not only measures distance, but records where I ran, and records my HR while I run, and is available for download and analysis, is pretty amazing. Its accuracy is better than in the old days, when I would drive my car on my running routes and record the distance from the odometer.

PHILIPSHAMBROOK
09-09-2010, 03:36 PM
+1 on that. For what it they are, wrist mounted GPS devices do a good job of providing information to allow for an acceptable level of performance measurement. When you consider that most of us run within a heart rate range of +/- 3to5 bpm, or a pace range of +/- 3to5 secs/km, then the accuracy is sufficient. Perceived problems arise a lot of the time when people expect pinpoiunt accuracy without understanding the limitations of the technology.

At the end of the day, if you are really focused on accurate peformance, it's hard to go past a stopwatch on an accurately measured 400m track.

GOPELAGO
09-27-2010, 10:47 AM
Hi All, here a comparison between a 5km race. Same place:
- edge 2005: 4.92km, -1.60%, http://connect.garmin.com/activity/50776540
- fr 310xt: 5.19km, +3.80%, http://connect.garmin.com/activity/50176575

I ran both the races on the first track, few meters on the second track during surpasses.
Okay, I know, running on an athletic field is not the best to check a GPS precision because of the circular route. But I'd like to put in evidence the red track of the 310xt: it is clearly worse than edge 205.
Maybe the reason could be in the size of the device. Sincerely, before buying 310xt I thougth it was more precise than edge 205 because it is more recent. But I was wrong.

I have other comparison between edge 205 and 310 xt, same results.

Note: when I run with edge 205, I fix it on my arm with an home-made wristband. Both (205 and 310) are equipped with the last firmware version.

DARTONP
09-27-2010, 12:09 PM
Personally I wouldn't use a track to do a comparison check. Due to 50% being run on a curve clipping comes into effect due to gps recording every 1-4 secs. Also the tightness of only 400m per lap causes weird routes on route view. The best way to fully test is over a straight route in an open area. Also left arm or right arm wearing might have an effect. Tracks are measured a few inches from the edge of the track to follow line of the runner.

If you use a road route that has properly been measured then this will be better but also bear in mind over 5K routes can have a tolerance of +/- 10m as far as I can remember. Other distances will have different tolerances. Road routes therefore are not fully accurate.