Garmin 405 not measuring correctly?

Here is an interesting thought in regard to measuring a course. If you use a jones counter or wheel to measure a grass course, it will measure every little elevation change which could be a simple 1" clump of grass. A steel tape will measure right over those clumps as will a GPS. When running the course, a person with an average foot size of 10" and an average stride of approx. 5' would actually only contact approximately 1/6 (17%) of the course surface. Knowing that, I have to ask, Why would we measure all those little elevation changes (clumps) when athlete's stride will straight line over 5/6th (83%) of the them?

Hmmm?????

OK, you may be right. I don't have any experience with Garmin's people. However, let me give a real life example with Polar's S625X. Like you, I feel certain that there are very smart people at Polar. I've looked at some of the research papers, and observed the quality of research they do, to use HR measurements to predict Calorie consumption, HRmax, VO2max, overtraining detection, etc. I'm sure the research even involves some curve fitting with statistical analysis.But, when Polar introduced the S625X, with the S1 footpod, there were some instances, with non-steady state running like fartleks or repetitions, where the device showed significant distance discrepancies. I don't speak of simple calibration issues, but having the device tell me that the distance of a half-marathon was both 20.5 and 20.8 km, for the same event. Which distance shall I calibrate? It seems that there are at least 5 different ways to see a race distance, either in real time, or after the fact with stored data, and each one has it's own software bug that might change one distance, independent of all others.Some trivial programming errors include:1) Storing the speed, rounded the nearest 0.1 kph/mph, essentially giving 3 significant digits, virtually guaranteeing up to 0.5% error, simply due to choice of data representation2) Computing the distance for each interval with a simple integration algorithm (speed x interval), and truncating the intermediate result to the nearest meter (or yard), effectively losing 1 or 2 of the 3 significant digits, for each data point sampled (e.g. every 5 seconds)3) The simple integration does not handle the initial measurement correctly. That is, if you start measuring with a "flying" start (a non-zero speed), you effectively start your event with non-zero distance, possibly as much as 40m. This kind of space warping isn't typical at the sub-light speeds associated with running. This bug applies to laps/splits too, so that each lap time you take can inject 25-40m into the recorded distance for each lap.4) This extra lap distance does not appear in the sample distance calculation described in 2), therefore causing a divergence of the two distance calculation methods, with every lap you record.When Polar introduced the RS800SD, and the corresponding PPT software, they increased the sampling rate to allow 1 second recording. Intuitively this might seem to cut all errors by a factor of 5, but the error described in 2) above, was actually multiplied by 5, with the smaller interval width still being truncated to the nearest meter, compounded by an increased probability of accumulating errors with the same sign, lowering the probability of multiple errors canceling each other.So while Polar may also be full of smart people, the ones doing the research are likely not the ones coding the simple algorithms, or they may be under tight staffing, schedule, and budget constraints, and in programming, it's easy to get error management wrong by an accumulation of decisions that the calculations are close enough.Regards,

Mr. Ray wrote:

You presume there are PhD's level mathematicians and programmers at Garmin. Sure the processors are quite capable, but if the programmers have a 4th-grade level math capability, the processors will perform 4th-grade math. If you factor in fixed bit length registers, and poor error management, could even be a lot worse than 4th-grade math.

dukerdog wrote:

I feel certain there are some very smart people at Garmin. The level of complexity they have to deal with in processing the data they get from the satellites in order to determine absolute position is far beyond that needed to do a simple curve fit, or even more advanced signal processing, to determine distance traveled. People think the satellites just "tell" the device where it is. Of course there is a lot more to it than that.

If we take that further, don't runners move up and down with each step? Shouldn't we measure the arc that the center of gravity follows, rather than the flat or bumpy surface of the ground?

Dogman wrote:

Here is an interesting thought in regard to measuring a course. If you use a jones counter or wheel to measure a grass course, it will measure every little elevation change which could be a simple 1" clump of grass. A steel tape will measure right over those clumps as will a GPS. When running the course, a person with an average foot size of 10" and an average stride of approx. 5' would actually only contact approximately 1/6 (17%) of the course surface. Knowing that, I have to ask, Why would we measure all those little elevation changes (clumps) when athlete's stride will straight line over 5/6th (83%) of the them?

Hmmm?????

Interesting thoughts.

Sorry, dukerdog, but the Garmin 405 doesn't do that. How ould the Garmin know that I was running a curve rather than just running straight lines. How does it know when calculating a curve fit is appropriate and when it isn't? If it did this, it would calculate a curve erroneously when I really did a 90 degree turn, would it not? If it did this, it would give large erros based on its placement error alone. I generated a track map of my run yesterday and it had me off in the grass by 10's of meters on stretches of my run. How would it know not to calculate a curve on any set of 20 points?

Just look at a track map in Goolge Earth generated on a track. It tells you everything you need to know.

Plus, dukerdog, your confidence that the Garmin 405 does this well doesn't account for the evidence - 100m per mile errors on judging distance on a track verses almost no error judgine distances in straight lines and/or gentle infrequent curves. It only updates every 5 seconds or so and at a pretty good clip around a tight continuous curve, that would give crappy curve data even if it did do this. I am still unsure how the Garmin would know if it should or should calculate this magical 20 points as if you were running in a straight line or a curve sine every single pinpointed spot can be 10's of meters away from where you actually are. I am not saying it can't sort of gentle out a hars set of plotted points, but that 5 second updating data makes this pretty crappy on a track and that it will misapply this strategy in situations where is isn't needed (such as a series of points plotted when you are running straight but there is left to right error in fixing you on the globe making it appear you wavered back and forth).

I have another general question - does the 405 take its first reading from a satellite right when you push the start button or is there a regular 5 second cycle and it may be as much as 4+ seconds before it takes its first reading?

Actually it's not so clear that they don't play with the data. I recall that the older version of the ForeRunner - the 201 - used to have a user-configurable "smoothing" factor. It clearly was playing with the data after the fact, not just going point to point.

There does not seem to be such a configurable item in the 205/305 or later but that might just mean that they chose whatever they thought was best.

I don't know one way or the other.

Regardless, while I consider the ForeRunner a useful tool, I would never attempt to actually measure a course with one unless I had no other alternative. Perhaps when satellite technology is better, and ForeRunners have WAAS capability, etc, they might become accurate enough for certification. For now I would always trust a wheel over my 205 anyday, on any course, grass or roads.

i have a 405 and use it around a loop with lots of trees and turns. Everytime it is right on with the markers that have been measured on the path. Also the overall distance of the loop is always the same and what the sign at the park says. I have had nothing but good things to say about my garmin. It's very accurate

No, for me it wouldn't be a big deal.

"We" are speaking of more than just race day. Training was mentioned prior to your previous post.

Okie-dokie.

Pay attention, stick with the context. I referenced recording and posting every mile split.

You are now using revisionist (mis)interpretation to desperately cling to a pedestrian practice. In other words: you're wrong.

All my comments are prefaced with being greatly pleased with my 405. I am not knocking it at all. I am just interested in how it works and the various factors that cause it to be more or less accurate. I would be interested in it even if I didn't have one. It is a pretty amazing device given that 20 years ago we didn't even have decent home computers.

It seems to do fine with gentle curves. The track maps reveal alot about what is going on on the track. It does seem to try to smooth the turn, but it makes mistakes. The the biggest error on the track is left-to-right errors compounding themselves the longer you stay in a relatively sharp turn. It actually seems to do better with quick 90 degree turns than a turn you stay in for a long time. I think the error is compounded by a collusion of two factors - the sharpness of a turn and how long you stay in it. An immediate 90 degree turn and it just rounds off the tangent and little harm down. A long, but very gradual turn and the radius error caused by a misplacement inside or outside where you actually are isn't that great. But a realtively steady curve that you spend a long time running through seems to throw its calculations way off owing to a combination of not putting you in the correct "lane" so to speak and the imperfectness of its rounding attempts.

We have a paved path loop around a lake here with meandering turns in and out of side coves that the local USATF dude measured with a Jones Counter at 7.39. The city put in GPS markers that suggest it is 7.35. My GPS always reads 7.41 on any steady run at any pace - 7:20 down to 5:20. However, if I do a fartlek on this loop with lots of pace changes, it always reads longer - up to 7.5+. I am wondering if the smoothing feature is confounded by pace changes since you cover more distance between point readings at some paces than other.

Brian, just because you don't understand how something works doesn't mean it doesn't work, it only means that you don't understand it.

I responded to the point about using a Garmin in a race. Nothing more.

You "referenced" this:

"If you fall apart, especially in a marathon, you usually know it without even needing to look at a watch."

Watching the mile splits early is intended to allow you to avoid falling apart in the first place.

Oh so their PRs were "miracles" eh? Big surprises? Get real. You have just demonstrated you have no idea what you are talking about.

Jones was out to smash the field that day. He had already run 2:08 the previous year, 1984. In 1985, he was looking to take his world record back from Lopes. The 2:07 was no miracle (except to the uninformed) for a 2:08 guy who was in better shape and already capable of a 1:01 half.

Rodgers' PR was all of a 28 second improvement after what he himself described as the toughest training cycle of his life. Yeah, big miracle.

Average_Joe wrote:

I responded to the point about using a Garmin in a race. Nothing more.

The point is that how you do in a race is a reflection on how you train. The two are inseparable, like it or not!

You "referenced" this:

"If you fall apart, especially in a marathon, you usually know it without even needing to look at a watch."

Watching the mile splits early is intended to allow you to avoid falling apart in the first place.

Go back further. I originally questioned the utility of recording and reporting every split in a road race.

Oh so their PRs were "miracles" eh? Big surprises? Get real. You have just demonstrated you have no idea what you are talking about.

Right, and what is BR's famous quote about running a great race? Do you really know anything of SJ's first two marathons in Chicago? Don't tell me there is anything about the attitudes exemplified in each that supports your anally cautious approach. It's deluded to believe that a Garmin-humper such as you races the marathon anything like either of them did in their respective primes. Wake the fvck up, don't be a clown. To paraphrase Lloyd Bentsen: you, sir, are no Bill Rodgers or Steve Jones.

do the math wrote:

The point is that how you do in a race is a reflection on how you train. The two are inseparable, like it or not!

Wrong.

Do you really know anything of SJ's first two marathons in Chicago?

Clearly more than you do. Enough to know that the two I mentioned - 1984 world record and the 1985 PR - are actually his second and third Chicago marathons, not his first and second. He dropped out of his first one at 16 miles.

I watched the 1985 one when the coverage was first aired. I've watched both the 84 and 85 races, start to finish more times over the years than I can count.

Don't tell me there is anything about the attitudes exemplified in each that supports your anally cautious approach.

Are you even reading these things before you respond to them? You're like the presidential candidates. You are asked one question and answer a completely different one.

I said there were no miracles on race day. I said Jones and Rodgers PRs were not miracles but clearly times that training and previous marathons indicated they could do.

You said that was "wrong." So back it up. How was Jones 1 minute PR a "miracle"? How was Rodgers 28 second PR a miracle?

I started this tangent making a simple point about the utility of the unit to allow a runner to know his pace when mile markers are in doubt. If close adherence to even pacing (and slowing down when the early pace is too fast) is good enough for the fucking world record holder, then it's good enough for me.

Um, you are the one claiming that "miracles don't happen" and now you're going against that. At least be consistent.

Who doesn't know that? He started in 1983 and, despite stepping in a pothole, ran the same tactic in that one and in 1984. Nothing about either of those runs was metered. He's said so himself. I guess you presume to know more about his running that he does, but I shouldn't be surprised by such arrogance out of you at this point.

Is there supposed to be a relevant point in this? Congrats on being such an avid spectator, I guess.

I don't pay attention to "the presidential candidates," so whatever. Do you always misdirect when you have nothing relevant to say?

Rodgers advises to go "berserk" and Jones positive split his WR. They got more out of themselves than some "oh no! that split was 20 seconds too fast!" weekend warrior ever would sticking to an overly-cautious approach.

You said "miracle," I did not. I used them as examples to counter your belief that your split-sweating approach or "blowing me" are the only two plausible ways to race a marathon.

On a fucking aided downhill course. Sorry, the mentality of seeking out an aided course is nothing like what is behind risking a WR attempt. Even for even splits, going through the half at what Geb did requires more balls than some guy seeking out gravitational aid just to break 2:40. You're the one saying that the pace was too fast. What about Steve Jones's WR run splits resembles your beliefs and what you advocate? I started this tangent making a simple point that such utility is negligible if you train right. I believe that Jones's and Rodgers's approach leaves no doubt that they got the most out of their training and themselves and that nothing about what you do (sweating minutiae) reflects that. Jones had NO IDEA that he was going to get the WR, that's irrefutable evidence that he was not checking splits.

I am perfectly consistent. You just didn't understand the point. Miracles do not happen. Really bad days do. Runners who are great in training and terrible in racers do. So how you do in a race is quite often NOT a reflection on how you train. For some runners, it rarely is.

I never claimed they were metered. For the FOURTH time now (please pay attention now), I said they were not miracles. They were not.

I know that. That still doesn't make the time any sort of miracle. Nor does it make it the ideal path to the fastest time. It just makes it unevenly paced.

Yes I did say, "miracle." And in direct response to that point, you said, "wrong."

That is not my belief nor did I ever make any such claim. I don't know what you think you were countering but there was none of that.

But then again, you are clearly responding to the posts in your head rather than the ones on the board.

I used the Pocono race because it was an example of a race with a couple of badly placed mile markers. To directly counter the point I was responding to. The terrain was immaterial.

It was too fast. Had it been correct and had I held it for many miles, it would have caused me to blow up. That's simply hard won experience. Far too many marathons where I threw out the plan and "went for it". When I started to run marathons more evently, I started to run a lot faster, with the same training. Go figure.

And the fact that Steve Jones didn't run even splits one year at Chicago does proves nothing except that Steve Jones didn't run even splits. The fact that Haile Gebresellasie felt the need to evenly pace his records does every bit as much to counter it. You should let Haile know how "defeatist" he is for worrying about pace.

The year he got the world record he didn't go out too fast. He held back until around 20 miles and THEN took off. His pacing was not ridiculous at all.

The year that he MISSED the world record by one second he went out much, much too fast and cost himself the bonus by one second. He ran PR pace for the half and then died.

But the point you keep ignoring is that either way, neither time was a "miracle" or anything like it. Jones had already proven he was capable of running with that crowd at long distances.

I understand all too well and typically it is. Weather and illness are typically not significant factors. If you screw up then you didn't prepare well enough. Most people prefer to not believe this, though, and they chalk up their failures to all sorts of mostly negligible externals. It happens in any endeavor in all walks of life, why should running be any different? Miracles, as you define them, may not happen but then disasters do not 'just happen,' either. Most poor races are a direct result of people shooting themselves in the foot before they even cross the starting line, and often also after they cross it, whether they intend to or not.

My only claim was that the way they ran is nothing like your conservative approach. Why you're dragging melodramatic language like "miracles" into this is unclear.

Nor does it make it a disaster nor the path to underperforming. What is the difference between a "miracle" (since you dragged the term into the discussion) and a breakthrough?

It was the exact implication of how you stated it, like it or not. Do you fail to understand how you implied that? Your quote, since you cannot even remember what you said:

"I've encountered the alternative many, many times, in many, many marathons. I have years of going out too hard and blowing me telling me that it was the right decision."

See, the choice of "the alternative" over "an alternative" sets the meaning to the reader.

You're saying that you ran Pocono by accident or at least that you had no idea that it was an aided course when you stepped to the starting line? I'm saying the very choice of that race, along with other details, paints a picture.

So was it correct or was it too fast? Or both? You see, you say that your watch confirmed that the pace was 'correct' to your way of thinking. Your body apparently knew it, too, even if you didn't trust it to.

No, it doesn't. Geb has always struck me as being much, much more cautious of a racer than Jones was. Geb errs on the side of safety, that is his reputation. He knows that, this is why he employs rabbits and hand-selects the fields he faces in his WR attempts and has been able to finish his track WRs with huge final laps. I don't need to "let Haile know" that, he seems self-aware enough without input from the likes of me. What an obtuse point. Jones, however, did not err on the side of safety. That would confirm that Jones is more likely to have gotten the most out of his body on that day.

And when he crossed the line in 1984 and looked at the clock, he had no idea that the time was a WR.

"I was tired, but I put my head down and sprinted. I didn't know what the record was." [Sandrock 254]

Jones didn't pace, he raced. He blew by the rabbit in Chicago in '85 at two miles. I guarantee you he has no regrets about how he ran and that it might have cost him a WR and any bonus.

"I wasn't even disappointed that I didn't get the record. I did set a personal record. I came to win." [Sandrock 259]

Is a "miracle" (or "blowing me") "the alternative" to "metered" running?

[quote]Mr. Ray wrote:

If we take that further, don't runners move up and down with each step? Shouldn't we measure the arc that the center of gravity follows, rather than the flat or bumpy surface of the ground?

[quote]

A measuring wheel should always be calibrated against a steel tape, as this will take care of any error introduced by those small bumps. But if the course being measured involves different types of terrain (grass, dirt, gravel, asphalt) then a different calibration/conversion needs to be used for each surface.

The errors introduced by these small bumps aren’t that large as one should always use a large 3' diameter measuring wheel if the surface isn’t very smooth. Never measure a bumpy surface using a small 1 or 2 foot diameter wheel, not even if you calibrate the wheel. Measuring wheels slip a lot on cinder/small gravel paths and on grass surfaces (somewhere between 20-80 feet of slippage per mile). That's why you calibrate the wheel to a steel tape on the surface you are measuring.

If you really want accuracy, then use a high quality (manufactured domestically, more expensive) 200-300’ flexible plastic tape that doesn’t stretch very much. Also calibrate it to a steel tape. I’ve seen plastic tapes off by as much as 2 inches per 100 feet. That seems small, but that’s a 27 foot error for a 5K distance.

A convenient generalization. Impossible to prove or refute. I could just as easily say "most poor races are a direct result of poor strategy".

Bull. I said there were no miracles on race day for Jones or Rodgers, that they achieved their times through talent and hard work. You said "wrong".

A breakthrough would be finally putting down a marathon commensurate with what either your training or training or racing has indicated you could but that you've been unable to do.

A miracle would be a marathon performance orders of magnitude beyond what any training run or other race could have predicted. If Jones has also run 1:01:40 for his SECOND half marathon in 85, THAT would have been a miracle.

I implied nothing of the sort. You inferred.

That was talking about outcomes, not strategies. It was an immediate response to your quote about knowing "any alternative outcome". Now who doesn't remember what they said?

No. I am saying that what I knew about Pocono and why I chose it is completely irrelevant to the fact that it was a course with some badly placed mile markers.

Which demonstrates only your ignorance about me and my motives, and is (again) completely irrelevant to the worth of a Garmin.

Had that been the actual mile marker, it would have been too fast. Since the mile marker was incorrect, the time was correct.

Because it's been wrong before. Had I not had the Garmin, I probably would have trusted it anyway. But I see no virtue in guessing when it is so easy to know. Ignorance is neither "ballsy" nor "tough", merely stupid.

Technically most out of his body would be the fastest time. It's not as if there was anyone left to race him after mile 2. Even Thackery couldn't keep up with him. The fact that he died so badly, running a 5 minute positive split is evidence enough that he did not get the fastest time he was capable of.

The former is an outcome. The latter is a strategy.

(The term "blowing me" was an editing error and should have read "blowing up". Thus is it does not suit as a synonym for "miracle". It is more like the opposite)