There's a difference between the 'accuracy' of individual assessments of position (which GPS devices are very good at) and the cumulative totals (of both distance covered and ascent gained) which those devices attempt to deduce. If the errors suggested by the single piece of research to which I previously referred are due to the irregularity of the routes taken on long walks (and off-road trails used by e.g. mountain bikers), then it may be that the differences between various carried or worn devices are due to the frequency with which they continuously compute those totals. The GPS is accurate, it's the totals which are not: perhaps as much as between +1 and +3 percent for distances, and a (plus ?) factor for heights which is still to be determined.

That's why I'm suggesting that the distance and ascent totals computed by the hand-held GPS devices carried by as many as 100 (?) of the 500+ participants in this year's Hundred be compared together. This might be the first time such a large-scale study has been done (who else does hundred-mile cross-country walking routes producing more than 15 sets of individual checkpoint arrival times for each of 550 people ?). Other than anecdotal evidence, at the moment we don't know which of the many competing GPS (soon to be Galileo, Europe's 2016 equivalent) devices is the most reliable on the very varied terrain typical of a Hundred, or if, as has been suggested, there's not much to choose between them other than battery performance.

To return again to Janet's original (2007!) question, the total distances computed by mapping software are likely to be more reliable than those found by hand-helds, because their figures are not influenced by the vagaries of geography (buildings, valleys, forests and the like) and satellite movements. Since such software does take account of gradients, it may be true that their totals tend to be higher than the 'map miles' traditionally used, but not by very much. In contrast, the discrepancy between distance totals computed by GPS devices and the 'real' distance 'on the ground' may be as much as 3.5%. The equivalent error for ascents may as yet be unknown.

Iain.

Please ignore the 'blank' posting from me dated 3rd Feb - a software glitch, moderators remove it ?. (Yes, it was my fault.)

In a similar discussion on another thread someone pointed out that GPS is used to land planes and in my mind this seemed to settle the accuracy question.
If you sit still for a while bogus mileage will accumulate due to the data calculation causing slight shifts, but this is very small with a good signal but can be a problem if you go indoors for a coffee. So I stop the device if I do not have a good signal when I have a rest..
Sometimes a path which on the ground is a straight line is quite wiggly on the ground and this adds distance. An extreme case is on the roman road north if Jedbergh, St Cuthbert's Way where the path zigzags considerably and the GPS correctly records a higher walked mileage than a line on the map indicates

Now that a sizeable proportion of people are carrying GPS devices, it may be a good opportunity for a reasonably large-scale comparison. The obvious 'test bed' would be the coming Hundred, where there might be as many as 100 different devices whose results might be compared for both distance and ascent (and the subject of Janet's original post, namely the accuracy with which ascent is accounted for in total distance measured).

Surprisingly, this does not seem to have been done, at least not over long distances. One 2014 cycle-based test* by Jeff Barber ("GPS Distance Accuracy Test: Smartphone Apps vs. Dedicated GPS", URL below) compared results from each of ten different GPS devices over 10x2.5Km laps of a velodrome. He found that though the average error rate (for 9 of the 10 devices) was 3.7%, only 4 of the 10 were within 1%, and one - the Garmin 60CSx - was almost 10% out. Interestingly, all of the other 9 'accurate' devices over-reported the distance, a discrepancy which cannot be due to ascent (he kept to the centre lane in each trial).

According to OpenStreetMap**, GPS accuracy is determined by several factors including weather, but principally receiver quality, satellite positions, and surroundings. The latter can include buildings, tree cover and valleys:

"GPS requires a direct line of sight between the receiver and the satellite. When an object lies within the direct path, accuracy suffers due to reflections and weakening of signals. This is particularly problematic in urban environments, within valleys, and on mountain slopes. In all three situations, the objects (buildings and the Earth itself) are substantial enough to completely block the GPS signals. When weak signals are received, they may have been reflected [off buildings and] the surrounding landscape. Reflections generate multi-path signals arriving with a small time delay at the receiver. Even when the object is less substantial (tree cover, car roof, your body), reflection and weakening of signals may still occur. This can sometimes be observed when viewing your recorded GPS track logs on top of aerial imagery [...] When carrying a GPS device, generally, the higher the antenna is fixed, the better the reception. Good positions include the shoulder strap or the top pocket of a backpack ..."

It's not so surprising, then, that measurements might prove be relatively inaccurate (c.f. the claims made for them) when GPS devices are carried in jacket pockets (particularly those of short people like me!) on the varied terrain (which includes woods, forests and valleys, as well as 'slopes') typical of long walks. It appears to be still true that unless GPS tracking is switched off, continuous satellite movement will create additive readings even when the device is stationary. This might explain why distance error tends to be up (though it's not clear to me why this was true for the cyle-track test cited above).

What does seem clear is that GPS track logs will show a discrepancy when compared with aerial imagery. Since the latter (and NOT satellite readings) is the basis for digital mapping software, it is those maps which are still likely to give the most accurate measurements of distance, if not ascent (which remains as difficult to assess precisely as before).

So unless and until we have a proper comparison of a full range of carried GPS devices over several long distances (see above suggestion for this year's Hundred) you'd be better off taking your cumulative distance measurements from 'the map', whether on paper OR a screen, and reserve your handheld for 'quick and dirty' position checking on those rare occasions when you are "lost", or, more likely, off route. Three and a half miles out in a hundred seems like a lot to me !

Iain

*Devices tested:
2 wristwatches, 2 bar-mounted units, 3 smartphones, 1 GPS-equipped helmet camera, 1 handheld device, 1 tablet. Device makes: iPhone 5 (Strava), Asus Android Tablet (Strava), Garmin Forerunner 405CX, Magellan Cyclo505, iPhone 5(Garmin Fit), Garmin Edge 500, Garmin Fenix2, Nokia Lumia (GPS Logger), Garmin VIRB Elite,. Garmin 60CSx.
source: http://www.singletracks.com/blog/gps/gps-distance-accuracy-test-smartphone-apps-vs-dedicated-gps

** http://wiki.openstreetmap.org/wiki/Accuracy_of_GPS_data

back to part of Janet's question. The GPS unit measures its distance from x number of satellites and resolves that into a position in three dimensions. It may in fact think you are in one of two places, the other being a long way off in space. However it assumes you are one of the (thankfully) majority of LDWA members who are on this planet and throws the other position away. It then places you on its reference theoretical "Earth", called a geoid, specifically the WGS84, giving you a lat, long and an elevation.
I assume then that if you climbed 10' ladder, it will give you the same co-ordinates, but a different elevation and will tell you that you have walked 10'.
I love it when people obviously quickly regret asking questions...

All distances shown on the Long Distance Paths (LDPs) are measured using a computer. Note that Anquet, Memory-Map, Tracklogs and Fugawi all give the same value.

Yes that's right, when converting on Satmap from trail to track it drops about 0.2ml on walk length, that's what I referred to as 1`% or less. When Satmap first appeared and the trail was converted to track, it "lost" 2 or more miles. I confess not to have studied what happens when I transfer the track from Satmap to Memory Map on my computer but I can't recall seeing a difference.

I have been playing abt with my new toy since my last post.
If I measure very carefully on Anquet 1:25k I am getting almost same numbers as my Satmap. When you convert trail to track the mileage drops abt 0.2 miles. If you import to Anquet it goes up slightly when the route is measured by the Anquet software..
The guy at Fishers Keswick said the most accurate was importing to you computer mapping and using that reading.
But this differences are small and disappear if you round to the nearest mile.

Dunno if anyone else mentioned this.....Isn't there a factor related to the GPS, if left on, continuing to receive satellite signals when one stops for, for example, a lunch break. Consequently, some extra yards/metres are accumulated on the gizmo. I recall when I first bought Satmap, 5+ years ago, it invariably registered about an extra 10% in walk length, characteristically it would record 22 on a LDWA 20 ml, but subsequent software upgrades seem to have reduced this to under 1%.

Also, if I left my device turned on, at night on my desk, I awoke to find I had logged up 6-7 miles on the trip odometer whilst sleeping! Quite unnerving first time it happened.

Walks measured with a GPS seem a little longer than those measured on a computer (Anquet or Memory Map), maybe by 5% - I have just acquired a GPS. The computer gives quite a bit more than measuring with a piece of string.
Does the LDWA have a preferred method?

I don't know about the Etrex H, but on my GPSmap 60Cx, you can go to the 'Trip Computer' page, press 'menu', select 'Big Numbers', press 'menu' again, 'change data fields', and then choose 'Location' as one of your fields. The numbers are then at least twice the size of those on the 'Satellite' page.

Try one of those specs cords, it saves the out + away bits or use the lens on your compass if it's got one. Sorry, just trying to be helpful.

Hi,
Does anyone have any experience of making the text larger in the Grid Reference field of an e Trex H GPS ????
all the other fields are fine for me to read.
I am trying to eliminate hassle of glasses out/ on / read / off / away
Thanks,
Brian Russell- 7271

Just as a matter of interest Matthew - why would a fell runner claim to be a member of a Flat Earth Society? Jane

Janet, join my 'flat earth society', probably less stressfull than worrying about all that. Matt.

I see that my square root sign has come out as a question mark in my last post!

As Elton points outs, for a normal cross-country walk, there is very little difference between the actual 'up and down' distance and the 'flat' distance. One can be more precise: if a walk is of length d when measured on the flat (i.e. directly from a map) and the net ascent (assumed equal to the net descent) is h, the actual distance travelled must be at least ?(d²+4h²), but cannot be more than d+2h. The least value occurs when the gradient is always the same (either ascending or descending), and the greatest possible value occurs when the ascent and descent are all in vertical stretches with the walk is otherwise flat (I am assuming that there are no overhangs!). To be more specific, for a walk that is 100 miles 'on the flat' which has a net ascent of 10,000 feet, the actual distance travelled must be at least 100.07 miles but cannot be any more than 103.80 miles. In practice it is likely to be much nearer the smaller value, given that extreme gradients rarely occur on normal walks.

For a walk of 100 miles with 10,000 feet of ascent, if the ascent and descent are evenly distributed, that would be 100 feet of ascent and descent every mile, and most people would hardly notice this. Even if the ascent is concentrated in the occasional steeper part, it does not make that much difference to the distance moved. However, the energy required to move the body a vertical distance is far, far greater than that to move the same horizontal distance, with the energy required increasing rapidly with angle, so the ascent is much more significant in energy terms than in distance terms.

When we see profiles of walking routes, the vertical scale is always very much exaggerated and this makes the rollercoaster route appear very much longer compared with the horizontal than it actually is.

Personally I would always (attempt to) give the distance of a walk as the flat distance - it's certainly what has been quoted for walks for years, and I'd hate to see Hundreds, etc, getting shorter as a result of new technology.

Measurement of distance on walks and in other situations is an enormously complex matter - there are other ways in which it may not be clear what one is trying to measure or how to go about measuring it.

I've just done a test on Memory Map by drawing a straight line across 3 km squares on the 1:25000 map covering Scafell Pike. It showed the distance as 3km but the height profile gave an extra 989m

Both GPS and electronic mapping measure the distance including the inclines, though in practice it makes little difference (thus sayeth Pythagorus). It takes a lot more energy to walk up a hill, so subconsciously we think that should be reflected somewhere, but the actual distance increase is minimal. On elctronic maps I think they call it the hidden distance.

A very good question, Janet. You will find an attempt to relate this to fractal theory in December Strider ...

I hope for the price they are busily calculating the distance up and down dale and capabale of providing a meal and washing the pots afterwards

This is undoubtedly very naive, and i will probably regret raising the subject - but when one reads measurements off a GPS - what precisely are you measuring. Is it map miles - which seems logical as the sattelites are way up high and I assume see the surface of the earth as a flat carpet - but on the other hand the GPS is also capable of giving you a hight reading , so is it busily calculating and actually reckoning how far you have walked up hill and down dale over the surface of the earth. If the latter, does it mean that long distance walks are now shorter as we are now reliant on the GPS to calculate distance whereas in the past we used map miles - and what about electronic mappping. What do they measure when you draw a wiggly red line over them?