Garmin 405 not measuring correctly?

The beauty of having a GPS watch (I've got the 205) is that you DONT STOP at lights and such, you just keep running sideways or you turn around and backtrack later.

I NEVER stop at lights or intersections and such. I didn't think anybody did either?

Thank you for this commentary, I honestly appreciate your thoughts. I have been looking around to try to get an idea of how well these devices (205 & 405) might work for me. Most reviews I see on the various running forums are about as useful to me as the reviews on MarathonGuide.com, they have essentially zero relevance to my basic use for such a device. I am well past believing in the utility of knowing the distance of a given run, my only real use for a GPS/pace watch would be for pacing LT and AT runs on the roads. I would want something that would allow me to input a pace range and then would sound an alarm if I deviate outside of that range. If the watch cannot even tell what the current pace is within a few seconds then I may as well just be running by feel and keep the $100-$300 for other stuff. I know that the Suunto model is crap from reviews I have read, can anyone say with decent authority whether the Timex Bodylink is useful for my purposes?

wak9 wrote:

The beauty of having a GPS watch (I've got the 205) is that you DONT STOP at lights and such, you just keep running sideways or you turn around and backtrack later.

I NEVER stop at lights or intersections and such. I didn't think anybody did either?

Haha, no, nobody else seems to do this. I run with a group sometimes and they think I'm weird for running back and forth at lights, but to me, this is the great thing about the garmin. If you run back and forth at a light 3-4 times, it adds up sometimes to nearly a mile of extra running.

Yup, lots of hills.

that would make sense

but anyway, regardless, it's not for me, I'll trust my legs to tell me how far I've gone, I'm generally very close

Try a Suunto t6c, either with the triathlon pack or with the GPS pod, for something to use during a tri.

You should be running by feel anyway. You have to be able to learn your pace by feel. If you rely upon technology to teach it to you, you'll always be reliant upon technology.

The Garmin is useful for telling you how far you ran (the utility of that is self evident) and giving you accurate splits -- like mile splits. It is not useful for instantaneous pace. I have no use for that anyway since I already know what my paces feel like.

If you are really looking for accurate, instantaneous pace readouts, you probably need something with an acceleromter. One of those distance measuring devices that uses a footpod. Over very short distances, they can be remarkably accurate, assuming you properly calibrate them.

That is all well and good and your advice is of course solid, especially since I have been running by feel for years and years now (hence no use for knowing how far I ran or splits, unless the split length can be set as low as something like 200m) and at this point I am looking to include critical zone training. (You may have heard of this?) Unfortunately my 'feel' does not reliably translate to the LT and AT paces as set forth by Vigil, especially as varying levels of fatigue affect the ability to 'feel' those pace ranges. So yes, I really, truly am looking for accurate, instantaneous pace readouts to aid in that endeavor. I am okay with being "reliant upon technology" for LT and AT runs in order to do those runs with ideal specificity, out on the roads. I have indeed looked into accelerometer watches - www.tech4o.com - and cannot tell how reliable and accurate any given product might be. I thought this forum might help provide the solicited answer, I suppose this is a sign that I should likely be hunting down Mahon's e-mail address to see what Kastor, Hall, et alia use for those workouts.

dukerdog wrote:

Garmin's statement about their device being accurate to within 15 meters doesn't mean it is accurate to within 15 meters PER MILE. It means that for determining your current location, it is accurate, on average, to within 15 meters. And that fact alone tells you almost nothing about how accurate it is in calculating distance.

Average_Joe wrote:

While I do not dispute any of what you say, I have found in practice that when the accuracy reading is around 20 (I think it's feet, might be meters), I almost always hit a mile split within 3 seconds of the true mile marker. This is on the Boston Marathon course so I trust the mile markers.

I'm not saying they're inaccurate. Based on tests I've done comparing to my own calibrated bicycle measurements, they are actually quite a bit more accurate than 15 meters/mile or 3 seconds/mile. I'm just saying that there is no way to convert Garmin's "15-meter" accuracy statement to distance measurement accuracy.

question for cut wrote:

if your going up a hill for 200m its only going to be about 150m from an ariel view.

Give or take about 49.5 meters.

I don't know how long it has been posted, but there is a new firmware update that, if I can understand the technical jargon, claims to improve accuracy and the pace function. I downloaded it last night, but haven't run with it yet. I found out about it because it also claims to fix a a freeze-up problem I have been having as well as helpfully adding a feature where you can disable the power save timeout while in training mode. I hated how it would switch to power save after 5 min. even if you were in training mode if the watch wasn't running.

To slightly correct Alan, the accuracy to 3-4 feet means that it correctly ascertains your current position at any given moment to that accuracy. It doesn't mean 3-4 feet per mile - it means at any given track point. I am not sure how that translates into mile accuracy. I have a loop I do around Lake Zorinsky here in Omaha with a fair number of twists and turns that has been certified with a Jones Counter at 7.39 and my GPS always shows 7.41 when I run easy there. The odd thing is if I do a timed fartlek, say 2 min. hard/1 min easy with alot of pace changes, it reads a good bit longer for the whole loop quite consistently (and each element of a timed fartlek varies quite a bit in distance on the GPS - unrealistically so - one two minute section might report 4:55 pace and another 5:20 pace and the recoveries tend to read unrealistically long).

https://buy.garmin.com/shop/store/fr405update.jsp

"Yes, under humid conditions, the thing trips out. Crazy."

When it is humid out, lock the bezel before running. Drops of sweat are interpreted as touches on the bezel. Irritating as hell, but alot of moisture on the bezel confuses the touch sensitive area.

the Garmin is useless on the track because of the steady turns - It has to draw a series of straight lines to account for a long drawn-out turn. Imagine that it takes a "pciture" of where you are every second (or half second or whatever) and then draws a straight line between those two points. It can handle a few sudden immediate turns without too much inaccuracy, but a long steady curve and the inaccuracy builds. On the track, mine will register 1 mile fully 50m before I actually finish a mile. If you view your track lines on Garmin Training Center, you can see how it makes its errors.

It also sucks on trails with alot of switchbacks. I ran a trail yesterday and the first two miles have alot of tight turns back on itself and it claimed I ran 9:20 pace for the first 2 miles. If you view the tracks on the mapping feature, you'll see why - it draws lines cutting the tangents and will also sometimes register no motion at all going around a turn. Hopefully the new firmware will fix this somewhat.

"my argument is that the garmin doesnt measure slope, sure it shows elevation but that if your going up a hill for 200m its only going to be about 150m from an ariel view."

Yeah, I thought that too, but we are wrong. I think Malmo set me straight on this year ago by getting me to set up the geometry problem that solves it - assuming the earth is flat is only slightly less accurate than accounting for even a steep slope. Set up a right triangle where the base is 1 mile (that would be assuming a flat mile) and the slope of the hypoteneuse is 5% (that is, it climbs 5' for every 100' of forward motion - a pretty steep hill as this would be roughly 250' per mile). Apply the pythgorean theorem and you'll see the length distance between the flat base and the climbing hypoteneuse is very slight.

I pretty much like my GPS for a few things. 1. Doing tempo runs in odd places that are hard to measure (running paths or rail trails, for example) and being able to quantify spontaneous unplanned hard stuff during other runs. 2. Getting auto half mile splits on said runs. 3. Small local races that aren't certified. 4. I can do a bit of quantification of timed fartlek workouts (say 2 min. hard./1 min. easy) and do them in places that are hard to measure.

I could care less how far my normal easy runs are or the pace of them.

question for cut wrote:

when you ran on trails were there any hills because a teammate and I got into a discussion about this and he says this 10mile loop we do is only 9.5 or so. my argument is that the garmin doesnt measure slope, sure it shows elevation but that if your going up a hill for 200m its only going to be about 150m from an ariel view.

I could have done a better job explaining how to set up the math problem to disprove this. Imagine a mile that climbs 100' - that is a pretty steep mile. We can use the Pythagorean Theorem to calculate the length of the hypoteneus (the distance of the climb) vs. the base (the assumption that the mile is flat. The base of your right traingle would be 5280' and the height would be 100'. So (5280 x 5280) + (100 x 100) = The square of length of the climb. Doing the math, you'll find the hypoteneuse is 5280.946. So the difference between assuming the mile is flat and accounting for the slope is about 1 foot. So while some inaccuracy would build over the course of a very hill trail run, it isn't nearly what you might think. The problem on trails is a combination of low signal strength and switchbacks.

okay we're getting somewhere let me give you some more info just on this particular loop. the first 1.5 mi is completly flat you start on top of the mountain, the next mile is nothing but switch backs going downhill dropping elevation of probably about 200ft and then you gradually decent with lots of windy turns before you start climbing the last 2.5 miles in which you go up about 500ft in elevation, probably about 300ft for the last mile, its pretty steep. thanks for all your feed back I'd say out of what i think is 10 miles there is probably about 4 miles that is either climbing or drops which adds up. now i understand you cant run for pace when your up there but there's no freaking way its going to take me 70 minutes to run 9 miles, we're fairly decent collegiate runners.

It just isn't the hills causing any GPS discrepancy. You can draw it out on paper as a series of right triangles. Just repeat the forumal above over and over and you are only adding several feet per mile even if it is constant up and down. It is the switchbacks. Those make alot of difference. You can google this and see reports from folks who have taken their GPS on trails (and only walking, which is more accurate since you don't cover as much ground between readings) and see the track map it generates and see how switchbacks throw the whole system off.

Although I don't think 70 minutes to run 9 miles on trails like you describe is that far-fetched - that isn't really that slow for a recovery day on trails - trails like you describe are much, much slower than roads.

question for cut wrote:

when you ran on trails were there any hills because a teammate and I got into a discussion about this and he says this 10mile loop we do is only 9.5 or so. my argument is that the garmin doesnt measure slope, sure it shows elevation but that if your going up a hill for 200m its only going to be about 150m from an ariel view.

While you might think that slope is a huge factor in the accuracy of the GPS, it is actually very minuscule. Calculate it out for yourself--use the pythagorean theorem. If in fact the 10 mile loop that is measured at 9.5 miles is off by .5 miles because of slope, you would have had to climb/decended 16,486 vertical ft. Correct me if I am wrong, but I doubt your 10 mile loop climbed/decended over 3 miles worth of vertical ft. Also, a 200 m hill would have to be 133 vertical ft high for it to be inaccuratly measured at 150m--Thats 88% slope.

There is error in a GPS because it must take a "reading" every second or so and every reading has an error associated with it (+/- 10ft or 3 m at the best). Considering that the GPS takes its reading from satellites that are miles away, that is pretty damn good. But the biggest error comes from this taking a reading every second or so. If you run on a curvy course the error will be more pronounced because the GPS will "cut tangents". Other factors to consider with GPS error are atmospheric interference (clouds, smoke, smog, sun spots), signal strength, and geographic location. If you run on a smoky/smoggy day your GPS will go haywire.

Although GPS's are good tools, they will never be more accurate than wheeling out a course, but wheeling out a course has limitations too. And for this reason, people should understand that the GPS is great for places that you cannot wheel out, but the distance the GPS tells them is probably inaccurate. I know there are a lot of "calculation freaks"out there running (I mean come on, we're runners) but drives me crazy when people start questioning the accuracy of tracks because their GPS said it was not 400m.

Sorry, wasn't trying to sound snotty. I just meant that learning how to hit those paces by feel is a valuable skill. But referring to the utility of knowing one's distance I simply meant that knowing how far you run is fairly useful to anyone keeping a running log that tracks mileage -- a fairly common practice among runners.

Fair enough. I do plenty of threshold and marathon-paced runs in which I need to hit a specific pace. My point was that you should be able to get to the point at which you can hit these paces while only having feedback at less regular intervals -- like every mile. (Note: by "should" I don't mean "ought", I mean that you will be able to learn that if you wish). However you actually CAN set the ForeRunner to give you feedback every 200 meters or so.

I think you might find some valuable feedback on accelerometers here, particularly if you start a thread just for them. In fact, I would be interested myself to hear what the best accelerometer products out there are.

This would illustrate how the Garmin performs on a track, or any course that has many turns:

http://runcolo.com/blog/garmin-405-accuracy/