Skip to content

Garmin Forerunner 405CX: Several runs later

February 2, 2010

This is the third part of my iterative review of the Garmin Forerunner 405CX (all leading up to a more comprehensive single-post review); click on the links for parts one and two of the review. I’ve been using the device for a couple of weeks, so I’m getting more used to how it works and what it can do.

Holly (who is also blogging her run-up to the Paris Marathon, here) posed a question on a previous post:

Hi there – just wondering if you’ve found that you’re suddenly running any faster with the Garmin? I find that there’s quite a big difference (especially on longer runs) between a route measured on an online map compared to a route measured on GPS. It’s obvious why this would be, but just wondered if you noticed a significant difference? I feel like although I may have technically run 10 miles according to the GPS, say, a 10 mile race may actually be a fair bit longer.

I’d read elsewhere that there are sometimes some issues with the accuracy of the distance measurements provided by GPS. For example, I’m sure a distance-measuring organisation (involved in certifying the distance of races) advertises a statistic around how many measurements are inaccurate compared with the measurements produced by their specialist teams. As would be typical, I can’t find that reference anywhere for the life of me now…

Rather than racing an officially measured course, or trying to measure a running route on MapMyRun, verify the measurement by driving it and then run it with the Garmin I decided that taking the device out on the same route a couple of times would give a good indication of its accuracy.

If I’m doing a tempo run, I’ll typically take a circuit of Abingdon which will, all in, come out as pretty much exactly 11.5k according to MapMyRun. Obviously, there will always be some variation in this depending on the line taken (e.g. which side of the road, where I crossed the road, etc.), but that shouldn’t add up to anything significant. So, over the past couple of weeks, I’ve done this run three times with the Garmin. Here are the details:

19 January 2010: distance – 11.57k; elevation gain – 153m

29 January 2010: distance – 11.7k; elevation gain – 139m

30 January 2010: distance – 11.62k; elevation gain – 80m

So, we’ve got a difference of 130 metres in distance and a rather startling 73 metres in elevation gain. If my route suddenly subsided by 73 metres, I would imagine that there would be a substantial number of Abingdon residents feeling pretty worried about the structural stability of their houses. The route elevation overlapped gives a clearer example of the extremes of this.

A comparison of the elevation profile from three runs over the same course

Spot the difference: three runs, same course, rather different elevation profile according to the Garmin

So, returning to the issue of distance, let’s assume this route is precisely 11.5k. At its most, so far, that would put the Garmin out by 200 metres. That would give the device an accuracy tolerance of something like 0.0174%, or a potential to gain around 17.4 metres per kilometre.

According to Garmin themselves:

With the GPS accuracy the satellites can place your unit in a 15m radius around where you actually are 95% of the time with a clear view of the sky. From time to time your accuracy will drop. GPS works on ‘line of sight’. The the unit does not have a clear view of the sky, your accuracy will drop and sometimes you will lose reception altogether. This is because the signal cannot travel through solid objects.

There are, of course, challenges around giving the unit a clear view of the sky when it’s strapped to your wrist. However, GPS works on triangulation, for which you need at least three satellites in order to provide an accurate reading. Presumably, the accuracy of both the distance and the elevation relies on the strength of contact with those satellites. In which case, I would assume that the variations in the elevation profile above were a result of one or more of the satellites losing contact with the device.

Although Garmin Connect shows you the route you took on Google Maps, it doesn’t retrospectively cross-check the elevation and distance. That information may be provided if you map a run in Google and then import it into the Garmin, but I need to find out more about that. If anything, the lesson from this is probably to map a run and use the Garmin to judge your pace (although that may be a second or two off the reality).

5 Comments leave one →
  1. Wazza permalink
    February 4, 2010 11:55 am

    Hi Lewis, couple of comments. Firstly your latest run looks like a classic “hitting the wall” scenario. Hope you can avoid it for Paris.

    If anything I would say the Garmin reads a little short. I’ve done 2 “certified” runs with mine -a “half” which read 21.27 kms and a 15 km “Great Run” with a large number of international athletes in the field, where my Garmin read 15.24. Maybe I didn’t run the tangents that well but I noticed the gap between my auto split and the official km marker got progressively longer as this race wore on so I can only assume a relatively consistent error all the way.

    It raises a question of whether you pace yourself off the Garmin . By my reckoning, if the Great Run scenario repeated itself in the marathon I would have to run at at least 6 seconds/ km faster to be safe. I don’t put a lot of store in the elevation graphs and sometimes do a double-take when I see they record me below sea level when I am running right beside a river.

    We have a famous hill “Anderson St” here in Melbourne where a lot of people train. As per my post on Cool Running below I ran up and down the thing 5 times one evening to gauge its length and elevation. The length was within about 3 metres out of 440 average. Elevation change results of the run down were very consistent while those for “up” varied quite a bit for some reason. Like you I seem to get enough variations on my regular run to take the readings with a grain of salt.

    I’ve had only 2 easy runs and the last 2 nights off this week so hoping I can find energy from somewhere to get back into it and keep ticking off the sessions.

    • February 7, 2010 10:23 am

      I guess accuracy is always going to be a bit of an issue with pacing devices – and while the Garmin seems prone to saying runs are a little longer than they are, I think it’s still more accurate than footpod-based devices. Still, that doesn’t really help when racing…
      I’ve always used a fairly bog-standard stopwatch for races, taking manual splits each time I pass the distance markers. I’ve found that approach quite good, but it partly depends on the accuracy of the distance markers, and whether you can spot them! I might try using the Garmin at the Bath Half in March so I can get a feel for whether it works for me in a race scenario.
      For judging the elevation profile of a run, it might be best exporting the Garmin data to MapMyRun and look at the elevation profile for the run on that service. Since that’s based on OS data, it should be more accurate. I’m hoping to spend a bit of time playing around with exporting and importing runs, but it seems a little more complicated than I had hoped.

  2. Julian permalink
    April 21, 2010 12:11 pm

    I use a Garmin 305 and download to SportTracks (free ware- Google it). As you have noted already, it tends to slightly overestimate distances. Most of the club championship races are distance certified by a man walking round with a measuring wheel which will be more accurate than a GPS or a car odometer. I have GPS distances for nine 10k runs and while 2 are spot on 10k, the rest vary from 10.2 to 10.06. Interestingly the Andy Reading 10k was 10.06 in 2008 and 10.00 last year, over the same course. A track 5k race was shown as 5.01 on my Garmin while 1/2 marathons vary between 21.08 and 21.15km. The London marathon last year was 43.41km on the Garmin, but it did get very lost in Docklands due to loss of satellite receprion and in the 2 underpasses.
    Interestingly in the first underpass the Garmin track appears to go straight on, so I guess the software assumes the runner is moving in a straight line. I’m not aware that I’ve ever had problems in trees and like Lewis have run at Shotover and various other cross country races.
    I never bother much with the elevation- the graphs give the general trend of ups and downs which are good enough for me. I took the Garmin skiing this year and the heights given on the lifts were usually within 5m of the Garmin, but it appeared the French rounded the altitude to the nearest 5m anyway. Maybe it was because we were 3000m nearer the satellites- only joking.
    Going back to car odometers, I’m not oversure of their accuracy- certainly the speedos read deliberately high for obvious reasons!!

  3. April 21, 2010 1:10 pm

    Looking at the Garmin readings for the Paris Marathon, it was about 500 metres (!) out. However, I certainly wasn’t running the path marked by the official distance measure, and there was about a kilometre where sateline reception was completed down due to being in a fairly long tunnel.

    I noticed that Garmin’s latest software update retrospectively corrects the elevation gains and losses. So, while the details are recorded on the watch, they’ll be fairly inaccurate, but once they’re loaded into the system they should be cross-referenced against a topographic map. (Not sure where that leaves you with ski lifts, though!)

  4. Dan Becker permalink
    February 5, 2012 11:56 pm

    I have a Garmin Forerunner 405CX that I just bought. So far, on 4 runs, It averages measuring my distances around 10% short of the actual distances measured by car odometer. Before I start, I wait for the satellites to get to 15 to 20ft. I may return it to Amazon and look for something else.

What do you think?

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: