Averaging Coordinates - A Little Confused

[+catcher24]

I have seen several posts that recommend or suggest taking several waypoint readings at the cache site and then "averaging" the waypoints. I just got my first GPS last week (Magellan Meridian) and it will display a message that it is "averaging", but I get the impression that this is not what is meant when someone mentions averaging waypoints. Any insight would be greatly appreciated. Thanks to all in advance!

 

Catcher24

"You see, you spend a good deal of your life gripping a baseball and in the end it turns out that it was the other way around all the time." Jim Bouton

[+sbukosky]

Waypoint averaging is a function of the GPS where it takes a reading every few seconds and averages them thus increasing the accuracy of the end waypoint. Without this feature, you can leave your GPS rest in one spot with as clear a view of the sky as possible and after several minutes, mark a waypoint.

 

Steve Bukosky N9BGH

Waukesha Wisconsin

[Kerry.]

quote:

---

Originally posted by sbukosky:

Waypoint averaging is a function of the GPS where it takes a reading every few seconds and averages them thus increasing the accuracy of the end waypoint ....

---

 

That is correct 50% of the time but incorrect the other 50% of the time. The thing is one doesn't generally know which 50% one ends up with.

 

Cheers, Kerry.

 

I never get lost everybody keeps telling me where to go

[martmann]

I could be wrong but I think catcher24 is talking about taking several readings and averaging them, like in This Thread

 

_________________________________________________________

If trees could scream, would we still cut them down?

Well, maybe if they screamed all the time, for no reason.

Click here for my Geocaching pictures

[+Markwell]

True, but Kerry has the empirical data that suggests that taking multiple readings within a short period of time may just make matters worse, as you could be averaging bad data.

 

After about a year of disagreeing with him, I'm leaning more toward his line of thinking. It still makes me feel better to take multiple readings when I'm hiding. However, the best course of action would be to come back to the location on another day and look at what the readings say then.

 

Markwell

Chicago Geocaching

[+Renegade Knight]

I always average about 100 points. That eliminates any random errors, or comes close enough. Systematic errors, due to satalite geometry, clouds, air temp etc. won't get eliminated in this short amount of time.

 

By way of proof I've gone to a cache with the owner so he could laugh at me while I found it. While I was wondering how in the heck to find it 20' over the edge of a cliff he was verifying his coordinates (discretely) and getting the same odd information.

 

When all was said and done the difference was he hid it during the day, and checked up on it during the day. His coordinates were within a few feet every time. When we did it at night, both of us had readings out into the canyon.

 

So Markwell, you guys are both right.

Also had we re-averaded our coordinates at night by taking a hundered or so points, the new coordinats would be off the other direction and we would be looking in the parking lot.

 

==============================

Wherever you go there you are.

[+briansnat]

I used to take 20 some readings and average them. I've found it's a waste of time. The cache coordinates where I used averaging were no more accurate than those where I didn't. Now I just make sure I have a good sat lock, mark a single waypooint and I'm done.

 

I've had very few complaints about my coordinates and in fact the one cache where I do get complaints was one that I averaged to get the coordinates.

 

"You can only protect your liberties in this world, by protecting the other man's freedom. "You can only be free if I am" -Clarence Darrow

[+Stunod]

quote:

---

Originally posted by Markwell:

However, the best course of action would be to come back to the location on another day and look at what the readings say then.

---

 

Even better might be to come back at a different time of day, since the sat constellation will repeat itself every 24 hours. If you have a poor constellation at noon today, it will be bad again tomorrow at noon.

 

"Just because I don't care doesn't mean I don't understand."

[dsandbro]

The whole theory behind averaging is one reading will put you x meters north of the actual coordinates, the next x or so meters south, then west, then east, etc., so when you average all the computed positions the average is most likely pretty close to the actual coordinates. This works most of the time.

 

Sometimes the first reading will be x meters north, then y meters north, then northeast, northwest, etc, so the average is still in error and not any more accurate. You cannot easily tell when the first or the second is happening, although the ratio is not 50-50. The first scenario is far more common and the second less common, although I have not seen any specific figures.

 

All you can do is look at the satellite screen for a rough indicator. If all the satellites are more or less clustered in the same half of the sky then you are more likely to have the second scenario (but not necessarily!). If the satellites are relatively well dispersed around the sky then the first is more likely, but again not guaranteed.

 

========================================

Friends don't let Friends geocache drunk.

[Kerry.]

Even in the days of SA position solutions didn't jump from so many metres north to so many metres south etc (and still doesn't post SA) but continuously wandered along a continuous (but extremely unpredictable) path, which in other ways was also certainly not random. Selective Availability had some unique and interesting qualities.

 

Averaging (over time) certainly can eliminate some of the unknown errors, spikes etc that occur, 95% of those errors generally occur within a 24 hour period and generally about 30 days to account for 100% of all errors.

 

In the end it generally still comes down to an unknown as one really doesn't know (with certainty) what type of period one is in (without planning), especially over the short term. The 50-50 figures and background is outlined in the following Averaging Example and the followup Averaging Part II followed by the Effects of obstructions on Averaging

 

Far more important these days is being at whatever location at the best time as time of day certainly matters more.

 

Cheers, Kerry.

 

I never get lost everybody keeps telling me where to go

[+catcher24]

Thanks to all for their replies. Sorry I didn't get a chance to respond yesterday, but didn't get any chance at all to even sit at the computer. Anyway, I think Martmann is on the right track. Unfortunately, I saw that thread already and it didn't answer my question. Using some suggestions above to start this example, let's say I take 5 readings at the waypoint on 5 separate days, or 2 hours apart each on the same day. How do I get the average? Same as any average - add them all up and divide by the number of readings (in this case, 5)? Or is there another way to do it using the GPSr? Thanks again for any info!

 

Catcher24

"You see, you spend a good deal of your life gripping a baseball and in the end it turns out that it was the other way around all the time." Jim Bouton

[I am not bob log]

quote:

---

I take 5 readings at the waypoint on 5 separate days, or 2 hours apart each on the same day. How do I get the average? Same as any average - add them all up and divide by the number of readings (in this case, 5)? Or is there another way to do it using the GPSr? Thanks again for any info!

---

 

 

That so stopet to waiste youre time. wat are you dum or somethent?

 

[This message was edited by I am Not Bob Log on April 09, 2003 at 09:37 PM.]

[+RobertM]

quote:

---

Originally posted by I am Not Bob Log:

My advice is to take one reading, mark it and use it. If you then get a large amount of complaints about the coordiantes, go back a and take another reading. But spending time averaging is a total waste of time.

---

 

Totally agree!

[Mrs.Nazz]

quote:

---

Originally posted by I am Not Bob Log:

My advice is to take one reading, mark it and use it. If you then get a large amount of complaints about the coordiantes, go back a and take another reading. But spending time averaging is a total waste of time.

---

Wow, Bob! You sounded just like Briansnat for a moment there!

 

quote:

---

Originally posted by I am Not Bob Log:

That so stopet to waiste youre time. wat are you dum or somethent?

 

[This message was edited by I am Not Bob Log on April 09, 2003 at 09:37 PM.]

---

 

Oh, thats better, NOW you got your accounts straight!

[+catcher24]

OK, thanks for the advice. I will do that. But, oddly enough (and no disrespect meant to anyone who has thoughtfully given a reply in an effort to help - I REALLY appreciate all of the responses), no onw has yet actually addressed the original intent of my question: do you get the average by adding all the readings by adding up the numbers and dividing by the number of readings, or by doing something else with the GPSr? Thanks again to all.

 

Catcher24

"You see, you spend a good deal of your life gripping a baseball and in the end it turns out that it was the other way around all the time." Jim Bouton

[+Stunod]

quote:

---

Originally posted by catcher24:

do you get the average by adding all the readings by adding up the numbers and dividing by the number of readings

---

 

Yup, that's the simplest way.

 

"Just because I don't care doesn't mean I don't understand."

[+timberlane74 & pumpkin]

Not to add a lot of confusion to the discussion, but when averaging a relatively small sample, you should be alert for "outliers", data that just doesn't make sense...for example if you sampled the temp at the North Pole for 10 days and 9 readings were between 0-10...and the 10th reading was 45...there is a HIGH probability of that reading being a BAD data point. Bottomline, make sure all the numbers you are averaging, make sense!

 

Timberlane 74 & Pumpkin

LaCrescent, MN

[+catcher24]

Stunod - Thanks! You answered the question. I thought that was probably the proper way - good, old fashioned math. But never having owned a GPSr before, I thought that perhaps there was some other method using the GPSr. The darned thing does so much I wasn't aware of I thought I had better make sure. And T & P, thanks for the tip.

 

Catcher24

"You see, you spend a good deal of your life gripping a baseball and in the end it turns out that it was the other way around all the time." Jim Bouton

[+Searching_ut]

Averaging is one of those topics you'll see debated from time to time regarding just how much of an improvement, if any you may get from it. Your Magellan averages it's position when it isn't moving. In comparing my sportrak and Legend (Doesn't have averaging capability), I've found that in most conditions the averaging seems to help somewhat, but it's not uncommon for the averaging to cause me problems. Personally I think not being able to turn off averaging on my sportrak is a flaw with the unit.

 

There is softare available that can average out your waypoints for you, or you can do a somewhat less accurate version of averaging by plotting your waypoints on a map, and manually picking the middle of the cluster. An example of doing something along those lines can be found at this page on my website:

 

http://searching_ut.home.sprynet.com/cache_coordinates.htm

[+timberlane74 & pumpkin]

MS Excel has the built in functions to compute all the "averages"...using the best available math functions, together with the "limitations" of technology...still will never give you a 100% accurate location! The more I think about the hobby, having really just started last week...if you are in this for the hobby why do you care if you are a 100' off? The fun I've had so far is getting close and using my noggin to try to look around and GUESS where someone else would also think would be a good hiding spot? Just an observation from a mathematically oriented newbie!

 

Timberlane74 & pumpkin

[+Renyoldman]

I'm with you Timber and pumpkid. Math and I are vile enemies so we stay away from each other as much as possible. Most of the fun is actually trying to find the cache. However, I would like to be at least with in 200' or so

 

"Democracy is the belief that 20,000 lemmings can't all be wrong."

[Kerry.]

Can't emphasize enough that being at a certain location at THE RIGHT TIME is by far much more important than any averaging scheme/thinking of any sort.

 

Cheers, Kerry.

 

I never get lost everybody keeps telling me where to go

[+blindleader]

quote:

---

...being at a certain location at THE RIGHT TIME is by far much more important than any averaging scheme/thinking of any sort.

---

 

True, but somewhat useless as advice. It's like applying the surest method for being wealthy - having wealthy parents. It's a great idea but I don't know how to go about doing it.

 

Averaging at a site for a long enough period for the satellites to move significantly is a method which you could actually apply reliably. Most geocachers, including myself, would consider it impractical, but at least I know how to do it.

[Kerry.]

quote:

---

Originally posted by blindleader:

True, but somewhat useless as advice .... It's a great idea but I don't know how to go about doing it.

---

 

Useless? well that's what planning and especially GPS mission planning software is all about. How to go about it? that's easy, use some mission planning software, several free versions available.

 

quote:

---

Averaging at a site for a long enough period for the satellites to move significantly is a method which you could actually apply reliably.

---

 

Reliably? how do you know that. What is significant movement as one also really doesn't know if the geometry is actually improving or simply getting worse with time.

 

As has been said many times before Averaging bad data simply gives a bad data average.

 

Cheers, Kerry.

 

I never get lost everybody keeps telling me where to go