GPS Accuracy

[ghalfaire]

Originally Posted by Fourputt

All that's required to calculate the distance is the Pythagorean Theorum.  No angular vector is needed to figure the distance between 2 points on a grid when the size of the grid is known.  Sides 'a' (GPS receiver location) and 'b' (data base target) are knowns, so figuring the hypotenuse is simple, and still leaves accuracy independent of the magnitude of separation.

Well FourPutt I think we are just discussing semantics here.  So let me go though an example of what I think happens inside the GPS which isn’t different from what you said.  If we have two point (P ₁ P ₂ ) in a 3 dimensional space each point’s location is defined by an ordered triple of (x,y,z) coordinates.  To find the distance between the two points you use the Pythagorean Theorem just as you said.  To use the theorem you have to find the x,y,z, coordinates of the distance between the two points.  These coordinates are (x ₂ -x ₁ ), (y ₂ -y ₁ ) and (z ₂ -z ₁ ) respectively.  These are the x,y,z coordinates of the vector that goes from P ₁ to P ₂ .  In other words if you’re standing at P ₁ and you walk in the direction of the vector described by these x,y,z results you’ll end up at P ₂ .  So I submit you know the vector even though you don’t use the directional information in calculation of the distance.  As to the errors, the computation of a square root (needed for the Pythagorean Theorem solution to distance) requires some sort of numerical algorithm for the processor to execute to determine the square root.  To the best of my knowledge all of these are some sort of iterative process. The last step in the algorithm is to check the error magnitude of the answer and if not sufficiently accurate (as determined by the engineer/designer) you go though the process again using the last error calculated to better the estimate of the square root (e.g. the distance).  This process of estimating the square root is continued until the error is small enough to satisfy whatever error tolerance the designer used.  There are several of these numerical algorithms and you can do a Google search to find them.  But they all will introduce an error (how large or small I don’t know) into the final distance calculation you see displayed on your device.

So I'm not trying to dispute that the GPS is accurate enough nor that what you say is inaccurate, just that there is more to the story.

[wmiller]

I have compared my laser to first my sg3 and then my current sg5 and the gps has never been more that 3 yds difference. 9 times out of 10, it's within 2. Plenty accurate for me, as I am not good enough to +/- an approach shot 2-3 yds.

[Fourputt]

Originally Posted by NCGolfer

I have Golfshot for the iPhone and it seemed good at first but my confidence in it changed to the point where I now use a Bushnell Rangefinder.  I used the GPS side-by-side and it was correct maybe 40-50% of the time (within a yard or so) but other times it was off by more than 5-7 yards which is completely unacceptable to me.  And it seemed to vary course by course

For the $30 I put into the app, it was worth it but I feel much better having the rangefinder.

This is why you almost never see a GPS at higher level state or golf association amateur tournaments.  I work as a rules official for Colorado Golf Association tournaments, and they allow electronic distance measuring devices at all of their championships.  These tournaments are usually composed of fairly good players, and you don't see GPS with even one out of ten players.  For these guys, that one hole where the GPS might be off by 10 yards could cost them the tournament.  They aren't willing to take that chance.

[ControlJunkie]

Originally Posted by Fourputt

Tee to green distances can't be compared between scorecard and GPS because scorecard distances are figured down the center of the hole around doglegs.  GPS is only capable of reading in a straight line.

Not sure where you're getting that from, but from the teebox, my SkyCaddie SGX traces down the center of the fairway and if the hole cuts right or left, it gives you distance to the bend, and from that point to the center of the green, which is usually pretty close to what's on the scorecard. The only time it's a straight line is if the fairway is straight all the way, or if it's a par 3. I will say I'm not familiar with the GPS apps on an iPhone or older golf-specific GPS models, but mine doesn't just show a single vector from tee to green. Another reason you can't just go by the number on the scorecard - teebox locations change as well.

To the original post - I gotta say, I've had an excellent experience with my SGX. I normally play with a buddy that has a a laser range finder (either Leupold or Bushnell, can't remember), and I cant remember a time where the GPS was way off of what he gets with his range finder. Usually the difference comes from him hitting the pin as opposed to me having distances to front, center, back, left, and right. But the number he gets is always falls perfectly within what I've got from front to back, and ends up being right if you walk off the difference from the front or the back.

[sacm3bill]

Originally Posted by ControlJunkie

Quote:

Originally Posted by Fourputt

Tee to green distances can't be compared between scorecard and GPS because scorecard distances are figured down the center of the hole around doglegs.  GPS is only capable of reading in a straight line.

Not sure where you're getting that from, but from the teebox, my SkyCaddie SGX traces down the center of the fairway and if the hole cuts right or left, it gives you distance to the bend, and from that point to the center of the green, which is usually pretty close to what's on the scorecard...

Rick was responding to this comment: "Also saw that the teebox distances were off quite a bit from the scorecard on several holes". I assumed, as Rick apprently did, that the comment was referring to the "distance to the green" number that all GPSs have. That value is certainly going to be different than the actual length of the hole, unless the hole is perfectly straight *and* the tee markers are set up in the exact location that was used to get the scorecard number.  So yes, an advanced GPS like yours or a uPro for example can include bends, but I took the original comment Rick was responding to to be referring to the default straight line distance that most GPSs give you.

On another topic: Rick and ghalfaire, you guys are in violent agreement - what Rick is referring to as the hypotenuse, ghalfaire is referring to as the vector. Same thing.

[zeg]

On another topic: Rick and ghalfaire, you guys are in violent agreement - what Rick is referring to as the hypotenuse, ghalfaire is referring to as the vector. Same thing.

Yes. What I'd add, though, is that the numerical errors due to computing the square roots that ghalfaire is talking about are, for any practical purpose, zero. The error on a distance between two points will be larger than the error in each position individually, but this is just the result of "subtracting" two imprecise values and would exist even if you used infinite precision math. If the position uncertainties for your location and for the target are equal, the uncertainty in the distance between them will be that error times 1.4, or about 40% larger. So if you knew the positions to, say, 1 yard, you'd know the distance to about 1.4 yards, and this is independent of the distance between them, to tie this back to the start of the tangential discussion. In practice, the uncertainties are not likely to be equal since the uncertainty of your current location is the result of GPS measurement errors while the uncertainty of your target is mostly due to mapping errors. In this case, the uncertainty will be slightly larger than the larger uncertainty, and at least 40% larger than the smaller of them. Still, it will be independent of the distance.

[nevets88]

Don't mean to hijack the thread, but this is kind of relevant. Golfshot looks like the most popular golf gps iPhone app. I downloaded an app called Golf GPS, which is free, but needs an e-mail registration so I gave it my throw away e-mail address. Why is Golf GPS free?

[ControlJunkie]

Originally Posted by sacm3bill

Rick was responding to this comment: "Also saw that the teebox distances were off quite a bit from the scorecard on several holes". I assumed, as Rick apprently did, that the comment was referring to the "distance to the green" number that all GPSs have. That value is certainly going to be different than the actual length of the hole, unless the hole is perfectly straight *and* the tee markers are set up in the exact location that was used to get the scorecard number.  So yes, an advanced GPS like yours or a uPro for example can include bends, but I took the original comment Rick was responding to to be referring to the default straight line distance that most GPSs give you.

I gotcha, my bad, like I said before, I'm not really all that familiar with others. I had always blindly assumed it was the same with others as it is on mine, but thanks for clearing that up!

[Dragonsrock]

I have both a laser rnagefinder and GPS, I find each has their advantages.  The accuracy of the GPS is normally pretty good, assuming good satellite coverage, the real issue is that the front and back of the green are rather arbitrary positions and the mid point of these or the centre fo the green is not that relaible, also the flag can vary significantly on the greens at my course, you can have about 5 yards error just with a centre flag position.  For front and back of greens, the gps is fine and here, in Australia, we do not have the WAAS(?) system and so error is greater to start with.

The Laser is definitely more accurate, but not enough to affect my game.  I remapped my home course and at least I know what the front and back positions are for the green and therefore that source of error is eliminated.

[ghalfaire]

Since the mapping errors are one of the significant errors in the Golf GPS I have often wondered how they do the mapping.  I have seen the Sky Golf ads with the fellow with a back pack receiver walking the course but somehow doubt that technique represents the majority of their database.  That would take an awful lot of man hours to for the thousands of courses that are in the database to have someone walking each and every course.  Cost prohibitive in my opinion.  So does anyone out there know for sure just how the course maps are developed?

[RayG]

Some are done using those backpacks. Those are "generally" the Sky Caddie/Golf Logix contractors doing a thorough site layout. A few of the "Free" or near free apps for iPhones/Blackberry etc... are done by the end users (for the most part). Using Google Maps from their website, you can follow their instructions to input tee boxes and middle/front/back of the green. A few also allow editing so that other users can input carry yardages, distance to a hazard, etc... It's almost like a "Wikipedia" for golf yardages. The first guy might just do a quick input if he's playing in the next day or two. The next guy might have time to add the intermediate yardages or even fix some of the errors. It's very simple to do, and for the most part, fairly accurate. It might get a little shaky if the course is on one of those Google Map sections where the resolution isn't that good. In that case it's almost impossible to get good numbers. I've done a couple on the Swing by Swing and BBGPSGolf sites.

[ghalfaire]

Just as an aside to this discussion for the first time I played with a fellow that had the same GPS as I do, a SG3.  We compared readings while playing and there was never more than 2 yards difference between the two units.  Most of the comparisons were either dead on or 1 yard difference.   I have often compared with other Sky Caddies and other manufactures' units and never had this small difference.  In fact one of our regulars has an SG5 and we have seen differences of 6-7 yards between the units (on this same day and course) and this is consistent with some other manufactures' units like Garmin.   Not sure how to explain this yet but interesting and thought it might be worth mentioning.

[RayG]

To be honest- if you need to worry about a 1 yd discrepancy in yardage between various devises, you wouldn't need them. You'd be on tour and your caddie would be taking care of these things!:-D Knowing that a trap needs 137yds to carry it to a tight pin, doesn't mean that most people will be able to say "okay- I'll hit it 138...". a "normal" amateur would take their 140 club and hope for the best (but should take their 150 and swing easy).

[GolfFans]

I see all your pros and cons of GPS Golf equipment.  Went to the App Store to look for my own one.  Then I found something that is out of the discussion you guys have.  I immediately bought this App - GolfSites.  This one has a Golf shot path replay thing.  That means it can show you on the satellite photo of the hole you played where you hit your ball.  This seems to be something no other GPS equipment cover.

If anyone is interested, I will be playing this Thursday and can write something up for you all.

[ghalfaire]

Originally Posted by ghalfaire

Just as an aside to this discussion for the first time I played with a fellow that had the same GPS as I do, a SG3.  We compared readings while playing and there was never more than 2 yards difference between the two units.  Most of the comparisons were either dead on or 1 yard difference.   I have often compared with other Sky Caddies and other manufactures' units and never had this small difference.  In fact one of our regulars has an SG5 and we have seen differences of 6-7 yards between the units (on this same day and course) and this is consistent with some other manufactures' units like Garmin.   Not sure how to explain this yet but interesting and thought it might be worth mentioning.

RayG I really wasn't worried about a 1 yard difference.  I was trying to point out that with identical systems/software that the repeatability of the design seems to be very good (not the same thing as accuracy however).  There were some earlier posts on what part of the error is due to the digital system calculation of distance. But based on this datum I would have to say I was overly concerned about the systems calculation errors.  So it was just my way of eating some crow and the fact I do find it interesting that identical models did so well while comparison with other models from from the same manufacture don't seem to compare nearly as well.  I really don't have a good explanation for that.   With my game I don't worry about accuracy to better than +/- several yards.  I am seldom going for the pin and just want to be putting next shot.

[zeg]

Originally Posted by ghalfaire

RayG I really wasn't worried about a 1 yard difference.  I was trying to point out that with identical systems/software that the repeatability of the design seems to be very good (not the same thing as accuracy however).  There were some earlier posts on what part of the error is due to the digital system calculation of distance. But based on this datum I would have to say I was overly concerned about the systems calculation errors.  So it was just my way of eating some crow and the fact I do find it interesting that identical models did so well while comparison with other models from from the same manufacture don't seem to compare nearly as well.  I really don't have a good explanation for that.   With my game I don't worry about accuracy to better than +/- several yards.  I am seldom going for the pin and just want to be putting next shot.

A systematic error like that sounds to me like a disagreement in the mapping.  There is room in the system for positioning errors, but I would expect those to be a lot smaller than that except under unique situations (very poor reception where a model that's slightly more sensitive could have a big advantage, when they're just turned on, etc).  A way to test this that might shed light would be to use the shot measuring tool between two distant points.  That should give their positioning accuracy a test independent of the maps behind them (the local accuracy, anyway---this wouldn't reveal anything about a shift affecting both positions).

I really wish that golf GPSes were a bit more general---the ones I've seen just give you a spot on their map, but there's no reason they couldn't spit out your latitude/longitude as a diagnostic.  Of course, I guess then you don't buy a second GPS for general use...

[GolfFans]

I went on the course this morning with this App GolfSites. First I need to add my club set to this App. This is not apparent until the point you enter your score.

I had downloaded my course (Brookside) the night before so I can try run this App first. Brookside is in a valley and don't have good signal for my iPhone so I want to make sure I have the course.

Turns out even when there's no signal, the GPS runs just fine, maps all there. Potentially great because if I go on vacation, I don't have to worry about roaming data charges on my phone bill.

This App uses a strange way to enter score, it's like the shape of the course. So when I hit a ball, I press Tee button, then it will ask me which club I use, I picked the driver. Then my second shot in hole 1 is on the left rough, I press the left rough button and then it ask me which club i use. So everytime I hit a ball, I press 2 buttons. It's quite easy as you get the hang of it. First few times a bit distracting and forgotten to press buttons. But after a while it's simple. The press of buttons marks my location apparently.

At first I am used to the scroll wheel like GolfShot to enter score.  But I don't like this type of entering because I have to think too much and that takes my mind off my game.  GolfSites way is a "no brainer" for a non tech person like me.

After the round, I saved my round and when I go to review it. Wow! I see my ball path outlined in green and blue. Maybe not all is accurate, as they did not show shots in some cases, I suppose it's too small to be viewed on the iPhone. But overall, it's a great way to review the game.

The rest of the features are pretty much like all other features I have seen on the App store.

Believe me, I checked a lot of them.

By the way, I charged my iPhone 4 before I go on the course and end up using 50% of the battery. I do use the sleep button quite a bit too.

[Paradox]

Originally Posted by ghalfaire

Since the mapping errors are one of the significant errors in the Golf GPS I have often wondered how they do the mapping.  I have seen the Sky Golf ads with the fellow with a back pack receiver walking the course but somehow doubt that technique represents the majority of their database.  That would take an awful lot of man hours to for the thousands of courses that are in the database to have someone walking each and every course.  Cost prohibitive in my opinion.  So does anyone out there know for sure just how the course maps are developed?

I have seen a Sky Golf fellow walking my home course...he was actually walking the same hole I was playing on.  He went to all the "features" from tee to green and walked the entire ring of the green, as well.  I'm not sure what else he did after that and it was a couple years ago, as well.  Course hasn't changed since then but they may come out and remap every two years or so.