Garmin 405 not measuring correctly?

CTZ wrote:

Um, what exactly does 5:00-5:52 mean then?

It means that you haven't been following the entire thread...

Brian says: It isn't a problem of satellites, it is a problem of trying to draw curves using straight lines and inaccurate left-to-right placement of each point compounding the problem of curvature.

Hey Brian: You have the problem wrong. The problem is that you are an idiot. Tracks aren't one mile long. They are 1600 meters long.

I'm going to have to agree with Brian here. This is about as good an explanation as I've seen as to why he's right. Dogman, forget about forward and backward arm swings 'canceling out' each other; they actually don't matter at all unless you're moving very, very slowly.

If the Garmin GPS units were actually able to continuously track your position, then yeah, arm swing would come into play. The GPS unit would show that you traveled a lot farther than you actually had because it would show your arm swings as forward and backward movement, essentially doubling your distance travelled. BUT - the Garmins only track your position at certain intervals (about every 4-5 seconds on average I think), so all that matters is where the unit is located at that very instant the GPS records your position. As Brian pointed out, as long as you are moving fast enough so that each tracking point is further along the route than the previous tracking point, arm swing has only an extremely tiny effect on accuracy.

Here's an illustration: let's assume you've set up the Garmin to track your position every second. Let's say that you're walking slowly at a constant speed of 4 feet per second. For simple math, assume that you're swinging your arm 1 foot behind/ahead of your torso on the backswing/frontswing. You start the watch as you begin walking, which records your begining point - A. One second later, the GPS records your position again - B. Let's say your position was recorded when your arm was at the forward-most part of the forward swing. So the GPS will say you've traveled 5 feet, when in reality you've only gone 4 feet. Another second later, the GPS records your new position - point C. This time, you were at the back of your backswing, so the GPS shows you've traveled 7 feet total, when you've actually gone 8 feet.

You can keep walking as long as you like, but the GPS is never going to be more than 1 foot off in total distance. Sure, the individual "splits" or distances between tracking points will vary slightly; they will all be between 2 and 6 feet (assuming the specs mentioned above), when they should all be exactly 4 feet. But the TOTAL distance will only be off by 1 foot, at most.

Again, if the GPS tracked your position CONTINUOUSLY, Dogman would be right. But this doesn't happen at all - the GPS only records INSTANTANEOUS position at certain intervals. It doesn't at all matter what movement goes on between each interval. You could be zig-zagging left and right between each tracking point, but the GPS would still show your movement as a straight line. The only way arm-swing would in any way affect the accuracy would be if you were moving so slowly that your forward arm-swing at tracking point "A" was AHEAD OF your backward arm-swing at tracking point "B."

I honestly cannot believe this conversation is still running.

Think for a moment. The Garmin takes a measurment EVERY FRICKIN SECOND. That is pretty close to continuous. To demonstate this, put your Garmin on, start it and while standing perfectly still, simulate a normal walking pace arm swing for :30 second, then stop the device. How far does the device say you traveled?

The FACT of the matter is that arm swing can affect the measurement.

Still don't believe me? Strap one one and walk a 10K. Log your measurements. Then drive the same 10K course at the same speed and compare measurements.

Anyone who cannot understand this should honestly go back to school. This is not rocket science people. I don't mean to insult you, but if my freshmen students can understand this concept, why can't you? Unless of course you are still in high school.

Stop the madness!

Dogman wrote:

To demonstate this, put your Garmin on, start it and while standing perfectly still, simulate a normal walking pace arm swing for :30 second, then stop the device. How far does the device say you traveled?

This is why you're wrong Dogman. Yes, if you were STANDING STILL and swinging your arms back and forth, the accuracy would be affected; the GPS would show you had moved when you actually hadn't (and actually even if you stood absolutely motionless, it would still show you had traveled a little bit, simply because the GPS is not 100% accurate as to showing your exact position).

But the whole point is that, because you are always moving forward when you're running or walking, each successive "tracking point" that the GPS records will be ahead of the previous one. Think about it: the GPS only "knows" where you are at one single instant every second. In between these single instants, you are "invisible" to it - it doesn't know what type of movement you are doing between those points. The GPS simply assumes you move in a straight line in between points. It doesn't measure CONTINUOUS MOVEMENT of the unit between each point, only the INSTANTANEOUS POSITION at each point!

I will admit that the Garmins are less accurate for measuring walking distances than running or biking (or driving) distances. But this is not in any way due to armswing. It's simply because the GPS is not completely 100% accurate. Someone earlier on this thread mentioned an accuracy figure of 15m. What this means is that at any given point that the GPS tracks you, you are within 15m of where the GPS says you are. So assuming you're moving on a relatively straight path, the slower you're moving, the less accurate the distance measurement will be. This is simply because there are more tracking points, which means there are more chances or opportunities for error. More tracking points might be better for curvy routes, but less is definitely more accurate for straight routes.

F

naahhh...52 seconds wasn't a typo. it had nothing to do with the accuracy of the garmin, but more to do with the pace of the runner who said that they ran 32 meters in 6-7 seconds. from the math, it seems like he is running a tempo in the 5:00-5:52 minute range. the comment just had to do with what pace he was running at. if he is doing tempos at 5:30, then i don't think that the garmin is going to be the difference whether this guy goes to the olympics or not.

anyway, they are just TEMPO runs.

ahhhh wrote:

It's all about context, genius. How did you reconcile "2%" and "a few seconds" with the 5:00-5:52 anyway? Did you even bother to think about it? It was certainly obvious. A shame you couldnt figure it out. But then again, figuring it out wouldnt have given you the opportunity to satify your infantile ego by calling the other guy a "buffoon" now would it?

So glad to hear that Joe Vigil thinks that a 50 second pace range isn't "fine and dandy". No one here thinks that either, which you would have understood had you thought about it for a second.

Sorry, but the jackasses on this site have proven to not deserve the benefit of the doubt nor have threads such as this one followed closely.

Dogman wrote:

Did you ever go to grad school and do research? I made a hypothisis and proved it to be true.

The results you cite as skewed are real and unaltered.

Lastly, I clearly acknowlegded the usual difference in arm swing between running and walking, BUT, I also acknowledged that running adds a tangent variable. That is why I stated that you should drive the course to eliminate both variables. That is when you chimmed in and said the arm swing would not make a difference.

Well you probably found evidence to support your hypothesis. As you probably know (as a grad student that can't spell chimed correctly) a hypothesis is either supported, or not supported, it's not proven.

Up front I will say I have not read all of this thread so maybe this has been answered.

It has been said here that the Garmin measures straight lines between points on the inside of a turn. My old 205 and the previous posts have said that running on a track consistently give a result of over a mile. I would think it would be less than a mile because all the measurement lines would be inside the curve of the track.

Any ideas?

See this for an illustration of behavior of the Garmin on a track (for a runner staying in lane one the entire time):

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

Yeah, and the Garmin measures 50m+ short, not 9m. You are the idiot. When my watch switches a mile when I am just past the 300m mark, that is clearly not the problem. Look at the much eariler post where there is a photo of tracks from on a track. It does not handle the continuous curve well.

And, incidentally, there are at least 2 440 yard tracks in the town I live in. There are still quite a few around if you live in the midwest.

Actually, I used the Garmin training center for the first time today and it only takes a reading every 4-5 seconds (I counted 50 ticks in a half mile one every .01 miles). That surprised me - I really thought it updated every second.

Even at every second, there is no way the watch would ever be retrograde of the previous reading only owing to arm swing.

good enough for me wrote:

If he is doing tempos at 5:30, then i don't think that the garmin is going to be the difference whether this guy goes to the olympics or not.

anyway, they are just TEMPO runs.

Then why buy a Garmin at all? If "not going to the Olympics" is the standard for concern (I am almost 40, afterall), why run or do workouts at all? You are just being pointlessly contrary. I mean, beyond a certian level of accuracy, no information is better than bad information. (Of course, I think the Garmin is accurate enough).

the reupper wrote:

Dogman wrote:

Did you ever go to grad school and do research? I made a hypothisis and proved it to be true.

The results you cite as skewed are real and unaltered.

Lastly, I clearly acknowlegded the usual difference in arm swing between running and walking, BUT, I also acknowledged that running adds a tangent variable. That is why I stated that you should drive the course to eliminate both variables. That is when you chimmed in and said the arm swing would not make a difference.

Well you probably found evidence to support your hypothesis. As you probably know (as a grad student that can't spell chimed correctly) a hypothesis is either supported, or not supported, it's not proven.

First off, I could give a rat's A$$ how I spell on a message board.

Secondly, You most certainly do "prove" support of the hypothesis.

Lastly, have a great run.

Can't we all just get along?

Brian wrote:

Then why buy a Garmin at all? If "not going to the Olympics" is the standard for concern (I am almost 40, afterall), why run or do workouts at all? You are just being pointlessly contrary. I mean, beyond a certian level of accuracy, no information is better than bad information. (Of course, I think the Garmin is accurate enough).

we are in agreement and i am not trying to be a contrarian. what i am saying is that the 6-7 sec/mile variation we might have from using the garmin to judge tempo pace isn't that critical to most runners. the garmin should provide good enough information for most of us.

but i would disagree with your statement about information. even 'bad' or inaccurate information can be helpful, as long as you understand its strengths and limitations. it is all about the interpretation. our discussion has commented on the inaccuracy of the garmin 405 but nobody has bothered to bring into the discussion the issue of precision. the garmin may very well have good precision (i have no idea actually, but my track workout results suggest this). in that case, if you are doing the same thing over and over, you can probably make legitimate comparisons of those activities. thus, the information may be judged as 'bad' because of inaccuracy, but may be 'good' because of precision. and i would say that in that case, some information is better than no information.