Garmin tax on course measurement

Since the GPS is some 200 miles in the sky, a Garmin can't be as accurate as the guy measuring the course. And the guy measuring a course probably doesn't have to take into account where water station tables are going to be placed.

On a recent 5K and marathon, it amounted to a 5 second 'tax' per mile, where I ran longer than the course measured. In Kathy's half marathon report, her pace was 9 seconds different per mile. And I enjoy seeing Kevin overlaying a city map on where his Garmin said he actually did (running through buildings without a scratch!).

So what is an acceptable difference in pace per mile (or additional distance ran) for a race? Thanks for your time.

I don't understand the question. "Acceptable" difference?

The courses are certified on the shortest legal route with a small added distance to account for possible variance. That is what it is. If you run the tangents perfectly as it was measured then you ran the course perfectly.

GPS devices are close for tracking purposes, but they are not 100% accurate. My experience is that they usually indicate I ran farther than I actually did. Over the marathon distance I've regularly seen 26.4-26.7 miles. Considering the distance travelled and how small we are as dots moving over the surface of the planet (with various obstructions) I think that is pretty darned accurate.

Given that, I can't use the GPS for proper pacing. I use the course markings and manually click the laps. The GPS is a 'close approximation' of pace at a given point in time, but is not a race clock.

oz, I am guessing you are asking how much you can trust the Garmin pace in order to hit your target time?

Then, I am with Matt that I use actual time at mile markers. Garmin pace is just to guide my pace. I am trying to limit the time I glance the Garmin, and hope I can merge these two as one.

In marathons, I don't worry about the difference until late stage especially after mile 20. If I have it, I should be able to catch the difference.

In short distances, proper pacing is more important. I never know if I can hold until the end, so I typically don't worry about the actual time. The intensity is high that calculating is too much a burden as well.

Well, then a .2 to .5 difference is what you've seen for your marathons.

Would you then say that if you ran 26.4 miles, given all the conditions, would probably be as close to 26.2 that you would get with all the people and other obstacles around?

I'm curious how close to 26.2 people have run a marathon and what an acceptable overage would be. In other words, "Oz, if you can run a marathon and have ran 26.xx miles, then that's as close to 26.2 as you're going to get. Even if the measurer paints a dotted line for how he measured the course on the street, you're Garmin still going to show you ran 26.xx. That's just the way it is."

Jerry wrote:oz, I am guessing you are asking how much you can trust the Garmin pace in order to hit your target time?

Then, I am with Matt that I use actual time at mile markers. Garmin pace is just to guide my pace. I am trying to limit the time I glance the Garmin, and hope I can merge these two as one.

In marathons, I don't worry about the difference until late stage especially after mile 20. If I have it, I should be able to catch the difference.

In short distances, proper pacing is more important. I never know if I can hold until the end, so I typically don't worry about the actual time. The intensity is high that calculating is too much a burden as well.

When I finished Houston, my 305 showed I ran it at 13:05 pace and 26.41 miles. The official pace was 13:09 on 26.2. That bummed me out, but made me think about what an acceptable difference should be? If I want to run a 12 minute pace officially, then I should be running at a Garmin 11:56 pace on my watch.

If a 3 hour time gets me a BQ, then I want to know during the race if my pace is going to get me there, including the extra amount I'm running. It's easier to spread 4 seconds per mile over 26 than 6.

I do appreciate y'alls thoughts.

ounce wrote:In other words, "Oz, if you can run a marathon and have ran 26.xx miles, then that's as close to 26.2 as you're going to get. Even if the measurer paints a dotted line for how he measured the course on the street, you're Garmin still going to show you ran 26.xx. That's just the way it is."

Right. My point is that the number you are seeking does not exist. GPS is not exact and can fluctuate +/- over the course of 26.2 miles. You could run a marathon and the GPS say you ran less than 26.2 miles. You could run a marathon and have it read 26.2 miles. In general, my Garmin has read long by .2 to .5 miles so I know I can't use it for hitting splits in a marathon. I can use it as a pace guide in real time, knowing that it's probably showing a faster pace than I am actually running.

The literal answer to your question would "26.2 miles", but I know that's not what you are asking...

EDIT: The answer that I would give is to not use the Garmin autolap for pacing. Turn it off and hit the button at the mile markers. That will record your REAL pace per mile and allow you to adjust during the race.

I've had a couple times on road marathons that I hit the pretty 26.27 both times at Boston. Chicago I had it were it was 26.17, now two weekends ago in my trail races since the trees weren't full bloom yet I nailed both distances right on the head for the 1/2 marathon and for the the 50k.

When I race I don't look at the pace or distance on my watch. I log it to have the data for later to look at but on race day I only watch the time on my watch and then I have my pace band that shows what time I need for each mile and for every 5k to help see if I'm on pace or not on pace. To me the bottom line is that the distance on the watch doesn't matter all that matter is getting from the start line to the finish line as fast as possible. I have to put my trust in the RD that the course has been measured correctly.

I'm sure other will post all kinds of data the acceptable distances and what the error of margin is but to me it doesn't mean anything. Even in training I log the miles that the watch sees but I can careless my training and racing is now done by time and effort. Yes I do training runs were I watch the pace on my watch in Tempo runs/Fartleks and some others because I want to know the intensity I running at and I feel the Garmin gives me a close enough reading on what the pace is, bottom line again is that I know when I'm pushing the pace and when I'm not.

That is why I decided a long time ago to gage my training/racing on effort and getting to know what effort feels like to my body rather being a slave to what the data on a Garmin says because it could be +/- and the same goes with HR data with me. I will never truly know my MAX HR sure can I get close but even with that it is only a good guess and there for the numbers could be +/-.

So what I'm getting at in that long unnecessary post is that I run my ass off on race day from start to finish and trust the RD on the distance and when it comes pace, HR, etc. etc. I trust my own internally GPS that is inside my body.

That's an easy question to answer then, I have 8 records in Garmin Connect: 26.34-26.71.

Interestingly for Houston -

2007: 26.34
2009: 26.41
2010: 26.34, I did run the shortest in my PR race 2:58:35.

As you mentioned, Kevin Fitz is the 365 resident expert on GPS, and if he tunes in over the weekend from his "boondoggle" in Rwanda I'm sure he wll have a lot to say.

Until he does, in addition to "building bounce" I think the easiest way to decribe it is that, even if you run perfect tangents, GPS measures the distance run by adding together a slew of tiny straight lines as a proxy for the curved lines of the tangents. The sum of those straight lines will generally be longer than the actual tangent.

From past discussions, the other relevant piece of information is where the satellite is located versus where you are running. So, identical paths in TX would yield a different Garmin result than in MA.

But, to give you a ball park extimate, I would say you can peg GPS error to about 1.0% to 1.5%. That's about 0.3 to 0.4 miles over the course of a marathon, which is about what most people report after running a measured course.

How many seconds that is will be a factor of how fast you run. But to validate it, I measured the error against permanently painted mile markers on the first four miles of the Boston course in Hopkinton and Ashland - twice, so that I would know how much to adjust the displayed Garmin pace by on race day. Now, I certainly have seen mile markers that have been "off" even in bigger races, but I think they know *pretty well* where the mile markers should be on that course. It came to 4 to 7 seconds while running a 7:00 pace, depending on the straighness of that particular mile, or...wait for it...1% to 1.5%.

Two extra info:

1. I typically round down the pace. For example, 3:00 marathon's pace is 6:52. I do 6:50. That gives me 52 seconds to offset the error if I don't use the potties. +/-5 seconds used to my boundary. Anything over, I considered a failure to good pace. But I started learning to use body feel.

2. Don't over use the tangent. Besides keeping a good rhythm, the side of the road is often a slope, which is not good for feet. Its difficult to run tangent in mega races, but what if one likes the atmosphere?

I have been pretty involved with GPS technology for years, mostly with the handheld devices (I am a volunteer administrator for geocaching.com-a site involving a GPS game). It is common for the accuracy of a consumer device to be off anywhere from 5-40 feet at any given time, with the average around 5-10 feet in good conditions and 15-25 or so in bad. GPS also measures in lines, so it doesn't handle curves well and does averaging on those. In urban areas, lost signal or bounced signal can cause its own problems with accuracy, causing it to be off by hundreds of feet at times. Complicating it further for running is that you don't know if the device thinks you are 10 feet ahead or behind of yourself or the the right or left at any given time and that is going to always be in flux a bit.

In my opinion, my Garmin 405 has accuracy to about the same level as my handheld and is normally within about 8-10 feet. I'm judging that from the couple times that I used it to find a geocache and am estimating. It also was spot on when I ran the Sandhills marathon in perfect conditions the middle of nowhere with full sky view and satellite reception and in an area with a high level of ground augmentation, plus the marathon was almost entirely one straight line. My handheld there tended to have little or no error and the marathon I ran measured exactly 26.2. That would be highly uncommon under any other circumstances and surprised me even then. But it did tell me that the device doesn't have worse accuracy normally than my handheld.

For running, what this means is that your GPS is not going to be as accurate as the official course measurement and the longer the race, the more that will become apparent. But trying to calculate that into something meaningful such as a degree of error in the pace etc is going to be difficult since the accuracy can vary at any given time based on too many different circumstances. I would trust it more over the short term, such as for current lap pace, than over the long term, such as total distance run on a long run, which is going to get increasingly off as the run gets longer. I also trust it more in rural areas with a clear view of the sky or in open flat lands with ground augmentation (I am assuming Garmin's forerunners use that based on my Sandhills experience, but I don't know 100% for sure) Ultimately though, I like to just assume that I might really be going a second or two slower than what it says and I don't get too terribly worried about it.

Edit/Update: I just checked and the Garmin 405 and above use the ground augmentation (WAAS) and the older forerunners do not, so the 405 and above are normally going to be bit more accurate.

I do what Matt does and turn off my auto lap and manually hit the lap button at the mile markers in races. I have a much better gauge of my pacing by doing that than if I was to use auto lap. In fact, when I lead pace groups, I instruct them at the start to either turn off their auto lap or ignore it because it will go off short of the mile markers more than likely and they won't have accurate data.

My marathons have ranged from 26.35 to 26.48 and are probably that close just by chance or because I usually run flat courses. I usually just count on race day adrenaline to give me the 2-5 seconds per mile that the Garmin tends to overestimate the distance traveled.

If you are manually hitting the lap button at each mile marker, as I do during a race, remember you must have the field set to 'Average Time per Lap', not 'Average Pace per lap'. Reason being, if you don't run the tangents well, you may very well run 1.02 miles on a given lap. If it takes you 9:00 to run that lap, your time is 9:00, but your watch will say your pace is 8:49. Thats because it thinks it took you 9:00 to run 1.02 miles. Yet as far as the course goes, you only ran 1 mile.

Tim C wrote:If you are manually hitting the lap button at each mile marker, as I do during a race, remember you must have the field set to 'Average Time per Lap', not 'Average Pace per lap'. Reason being, if you don't run the tangents well, you may very well run 1.02 miles on a given lap. If it takes you 9:00 to run that lap, your time is 9:00, but your watch will say your pace is 8:49. Thats because it thinks it took you 9:00 to run 1.02 miles. Yet as far as the course goes, you only ran 1 mile.

That makes sense - glad u r smartr than me.

Tim C wrote:If you are manually hitting the lap button at each mile marker, as I do during a race, remember you must have the field set to 'Average Time per Lap', not 'Average Pace per lap'. Reason being, if you don't run the tangents well, you may very well run 1.02 miles on a given lap. If it takes you 9:00 to run that lap, your time is 9:00, but your watch will say your pace is 8:49. Thats because it thinks it took you 9:00 to run 1.02 miles. Yet as far as the course goes, you only ran 1 mile.

Good advice...BUT... Mile Markers are not certified! I have found them to be as off, or more than the GPS!
>I use the Garmin Pace, but manually lap at the mile markers during bigger races were I am confident that the mile markers are accurate (wheeled).

My old 305, and my 610 (after a few software updates) is always within ~1%, usually with ~.25%. Large buildings and heavy tree canopies will decrease the accuracy. I have also found the faster I run, the more accurate distance seems to be. I am guessing this is because every data point has error, and the faster I run, the less data points there are.

Tim C wrote:If you are manually hitting the lap button at each mile marker, as I do during a race, remember you must have the field set to 'Average Time per Lap', not 'Average Pace per lap'. Reason being, if you don't run the tangents well, you may very well run 1.02 miles on a given lap. If it takes you 9:00 to run that lap, your time is 9:00, but your watch will say your pace is 8:49. Thats because it thinks it took you 9:00 to run 1.02 miles. Yet as far as the course goes, you only ran 1 mile.

If I don't hit the lap button at each marker, then Average Pace Lap is the one to use, right. I don't want to hit the lap button because it's one less thing to keep up withdraw

My thanks also to Carleen for the certain details of the watch's function.

ounce wrote:

Tim C wrote:If you are manually hitting the lap button at each mile marker, as I do during a race, remember you must have the field set to 'Average Time per Lap', not 'Average Pace per lap'. Reason being, if you don't run the tangents well, you may very well run 1.02 miles on a given lap. If it takes you 9:00 to run that lap, your time is 9:00, but your watch will say your pace is 8:49. Thats because it thinks it took you 9:00 to run 1.02 miles. Yet as far as the course goes, you only ran 1 mile.

If I don't hit the lap button at each marker, then Average Pace Lap is the one to use, right. I don't want to hit the lap button because it's one less thing to keep up withdraw

My thanks also to Carleen for the certain details of the watch's function.

Ounce, if you miss one marker, hit the lap twice at the next marker. For Example:

Average Lap = Lap Times / Lap
And
Average Lap = [10:00+9:59+10:01]/3 = 10:00
And
Average Lap = [19:58+0:01+10:01]/3 = 10:00

Both of you, oz. and Gobbles, are on a slippery slope I've slid down before. I recommend you get off it.

You may use any pace metric you want on your GPS training device to monitor your pace during a race. But realize, as Carleen, Tim C and others have already said, TIME is the ONLY metric that can give you an accurate reading of you pace on course. That's because GPS-estimated distance is in every pace metric you can display. In other words ... there is no way to tell your training device that each time you hit the Lap button, the distance should increment by one mile. You may use your GPS training device give you a rough idea of your pace during a race -- but that's all. Because there are so many factors that may bias the GPS-estimated distance long [and sometimes even short]** there is no way to know if you're actually going as fast as your device says or 10 seconds/mile slower ... except to look at your TIME when you cross a mile marker.

Personally, I keep autolap on during a race but manually mark laps if autolap triggered more than a few seconds before a marker or has not triggered yet at a marker. I then remove errant lap markers in SportTracks when I'm analyzing my performance. For monitoring my progress during a race, I pay little attention to Lap pace. It gives me an indication only. The real feedback comes when I cross a mile marker. You can take the TIME to the bank gentlemen. I suggest you graduate to the TIME metric too.

That said, Garmin places no tax nor levies any fee on course distances. Same can be said of Polar, Timex and other GPS enabled training devices. It is natural for these devices to estimate somewhat long. So if you insist on using Pace readings as a guide, assume your actual pace is somewhat slower than displayed [5-10 sec/mile] and check it in your head against your TIME at major mile markers on course.

Welcome to the NEXT LEVEL

**<< I tire of explaining this over and over and so will not.>>

KBFitz wrote:Both of you, oz. and Gobbles, are on a slippery slope I've slid down before. I recommend you get off it.

You may use any pace metric you want on your GPS training device to monitor your pace during a race. But realize, as Carleen, Tim C and others have already said, TIME is the ONLY metric that can give you an accurate reading of you pace on course. That's because GPS-estimated distance is in every pace metric you can display. In other words ... there is no way to tell your training device that each time you hit the Lap button, the distance should increment by one mile. You may use your GPS training device give you a rough idea of your pace during a race -- but that's all. Because there are so many factors that may bias the GPS-estimated distance long [and sometimes even short]** there is no way to know if you're actually going as fast as your device says or 10 seconds/mile slower ... except to look at your TIME when you cross a mile marker.

Personally, I keep autolap on during a race but manually mark laps if autolap triggered more than a few seconds before a marker or has not triggered yet at a marker. I then remove errant lap markers in SportTracks when I'm analyzing my performance. For monitoring my progress during a race, I pay little attention to Lap pace. It gives me an indication only. The real feedback comes when I cross a mile marker. You can take the TIME to the bank gentlemen. I suggest you graduate to the TIME metric too.

That said, Garmin places no tax nor levies any fee on course distances. Same can be said of Polar, Timex and other GPS enabled training devices. It is natural for these devices to estimate somewhat long. So if you insist on using Pace readings as a guide, assume your actual pace is somewhat slower than displayed [5-10 sec/mile] and check it in your head against your TIME at major mile markers on course.

Welcome to the NEXT LEVEL

**<< I tire of explaining this over and over and so will not.>>

That's why I wear a pace band with the cum time on it at each mile and that is also my major watch setting. I do watch instant and ave pace per mile so that I have an idea because I'm not the best at setting a steady pace without it. I use it as a regulating tool not an absolute time monitor.

Kevin,
I think you read into my comments too much.

I look at Garmin lap pace, but I know of the inherent error.
Like you, I use auto lap, sometimes manually lapping if the auto lap is too far off.
If the course has certified intermediate distance (5K or 10K) I manually lap (so I can look back).
All that said, time is the largest display on my watch - I create a detailed split plan prior to the race. I know where I need to be at what elapsed time. I'm also thoroughly annoyed by how inaccurate some mile markers and course clocks are at some races...

Ounce's "tax" does not exist - it is most often what I call "Garmin Garbage" (most) and inefficient running paths.

KBFitz wrote:Both of you, oz. and Gobbles, are on a slippery slope I've slid down before. I recommend you get off it.

You may use any pace metric you want on your GPS training device to monitor your pace during a race. But realize, as Carleen, Tim C and others have already said, TIME is the ONLY metric that can give you an accurate reading of you pace on course. That's because GPS-estimated distance is in every pace metric you can display. In other words ... there is no way to tell your training device that each time you hit the Lap button, the distance should increment by one mile. You may use your GPS training device give you a rough idea of your pace during a race -- but that's all. Because there are so many factors that may bias the GPS-estimated distance long [and sometimes even short]** there is no way to know if you're actually going as fast as your device says or 10 seconds/mile slower ... except to look at your TIME when you cross a mile marker.

Personally, I keep autolap on during a race but manually mark laps if autolap triggered more than a few seconds before a marker or has not triggered yet at a marker. I then remove errant lap markers in SportTracks when I'm analyzing my performance. For monitoring my progress during a race, I pay little attention to Lap pace. It gives me an indication only. The real feedback comes when I cross a mile marker. You can take the TIME to the bank gentlemen. I suggest you graduate to the TIME metric too.

That said, Garmin places no tax nor levies any fee on course distances. Same can be said of Polar, Timex and other GPS enabled training devices. It is natural for these devices to estimate somewhat long. So if you insist on using Pace readings as a guide, assume your actual pace is somewhat slower than displayed [5-10 sec/mile] and check it in your head against your TIME at major mile markers on course.

Welcome to the NEXT LEVEL

**<< I tire of explaining this over and over and so will not.>>

BAM! Spot on and a lesson I learned a LONG time ago!

Thanks, Kevin. It makes sense, only I've never had good luck with pace bands because I'll usually incur a smeared pace due to sweat. I'll have to find a better one.

And I now agree that Garmin doesn't have a tax, as it's just my not running the course as measured.

So, I guess if I want to run 10 minute splits, including 5 seconds per mile for my inefficiency, then at each mile, it needs to show 10:00, 20:00, 30:00, 40:00, 50:00, 1:00:00, etc?

ounce wrote:Thanks, Kevin. It makes sense, only I've never had good luck with pace bands because I'll usually incur a smeared pace due to sweat. I'll have to find a better one.

And I now agree that Garmin doesn't have a tax, as it's just my not running the course as measured.

So, I guess if I want to run 10 minute splits, including 5 seconds per mile for my inefficiency, then at each mile, it needs to show 10:00, 20:00, 30:00, 40:00, 50:00, 1:00:00, etc?

If you run 10 minute mile, why do you need a pace band? Not even Chris does.

My experience is I can always round the pace so that it is easy enough to calculate myself during a race. If you feel difficulty at early stage, you are running too fast. At late stage like after mile 20, I calculate how much time I have left to decide the split I need to hit each mile.

Give me a target time, I can show anyone how to do it.

Jerry wrote:

ounce wrote:Thanks, Kevin. It makes sense, only I've never had good luck with pace bands because I'll usually incur a smeared pace due to sweat. I'll have to find a better one.

And I now agree that Garmin doesn't have a tax, as it's just my not running the course as measured.

So, I guess if I want to run 10 minute splits, including 5 seconds per mile for my inefficiency, then at each mile, it needs to show 10:00, 20:00, 30:00, 40:00, 50:00, 1:00:00, etc?

If you run 10 minute mile, why do you need a pace band? Not even Chris does.

My experience is I can always round the pace so that it is easy enough to calculate myself during a race. If you feel difficulty at early stage, you are running too fast. At late stage like after mile 20, I calculate how much time I have left to decide the split I need to hit each mile.

Give me a target time, I can show anyone how to do it.

It was to make sure I had the concept down in my mind. And I didn't want to go over your head too terribly much.

ounce wrote:Thanks, Kevin. It makes sense, only I've never had good luck with pace bands because I'll usually incur a smeared pace due to sweat. I'll have to find a better one.

And I now agree that Garmin doesn't have a tax, as it's just my not running the course as measured.

So, I guess if I want to run 10 minute splits, including 5 seconds per mile for my inefficiency, then at each mile, it needs to show 10:00, 20:00, 30:00, 40:00, 50:00, 1:00:00, etc?

As for the smearing from sweat, use clear packing tape. I will take a photo when I get home of mine to show you.

As for pace I do two bands one that shows pace and time for each mile and another with 5k splits with half and finish.