measuring xc courses?

J.R. wrote:

Steel tapes are useful for setting up calibration courses, as these are straight line measurements.

I see no reason why they would be any good for measuring cross country courses though, as such courses are not in straight lines. Apparently some people use them but I have no comment about that.

Also I've not used a surveyor's wheel. Are they completely metal or are the tires made of rubber? Have you given one a push down a long hill to see how far it will go on it's own?

I agree about not using the steel tape method on undulating surfaces and with multiple and varied curves.

There are many types of surveyor's wheels - most have a thin rubber or nitrile coating over a metal or poly wheel.

I have not rolled any wheels down hills, but I can get then to go straight across the shop floor all day long.

Ah, but the question is, if you only briefly looked at all 8 pages, do you have any idea what we are saying?I say, in cross-country we wear spikes, to prevent slipping -- why not put them on the click wheel too?

Ray L wrote:

Having briefly looked at all 8 pages I am still amazed that anyone would want to accurately measure a cross country course. Do you people have any idea what you are saying????????? Learn the history.

rekrunner wrote:

Ah, but the question is, if you only briefly looked at all 8 pages, do you have any idea what we are saying?

I say, in cross-country we wear spikes, to prevent slipping -- why not put them on the click wheel too?

Ray L wrote:

Having briefly looked at all 8 pages I am still amazed that anyone would want to accurately measure a cross country course. Do you people have any idea what you are saying????????? Learn the history.

Same reason Jones Counters have to recalibrate all of the time because of different surfaces. Just imagine how off the wheel would be if it had 1/8" spikes that worked perfectly on golf course fairways, but couldn't penetrate the cart paths or other harder stretches of ground.

The "Click" wheel should never be used for course measurement. And a person can learn to apply appropriate pressure on the handle of a continuos gear push wheel to eliminate slipping.

rekrunner wrote:

The additional cost of a footpod is usually around $100-$120.

In additional to what though, my shoes?

1) I feel that such a device would need would first of all need 100 percent accurate stride rate measurement.

2) There should be a very accurate calculation of pace (stride length) from the rate. So far, there might be some that come close on the rate, but none that accurately extrapolate this to pace.

3) Users should be able to precisely calibrate the pace based on rate, for at least two different speeds. For example I should be able to input my pace at 165 and 178 spm.

These could show a very stable base estimation of pace. Beyond them, the device could start fine tuning with acceleration and such.

Here's an idea for the calibration of a device that calculates pace.

1) Set the device for "calibration #1";

2) Get a running start, say 10 meters, and start the watch at the start line;

3) Stop the watch at the finish, input the time and distance (1/4 mile or mile etc);

4) Repeat for calibration #2.

The watch would automatically count the number of steps for the distance each time, and you're inputing the paces, so everything else can be worked out from these. Then when you run, provided the surfaces and footwear are similar, any other paces can be easily calculated from your current stride rate, and this could be done continuously. I think this would be quite a useful device.

I meant to say that in calibration #1, the runner runs at a moderately easy pace, and in calibration #2 runs at a faster pace.

These would optimally circumvent the usual running paces.

Of course, then the calibrations could be redone any time.

Surveyor Steve wrote:

And a person can learn to apply appropriate pressure on the handle of a continuos gear push wheel to eliminate slipping.

I guess I haven't learned that yet because when I used my continuous drive wheel to measure 100m calibration courses on pavement and grass, I get about 1.5m difference between the two surfaces.

Have you ever actually tested your measuring wheel on a cal course that has been previously measured with a method known to be accurate (like a steel tape)? Until a few days ago I never had, and I was very surprised at how much difference there was between grass and pavement.

I could see the soft fairway ground I was on deforming under the wheel, so I doubt pushing harder, and causing more deformation, is going to improve things, but I'll give it a try.

Are you sure that the grass route you are measuring is as flat as the pavement on your calibration course? Also, do you walk slowly, quickly, or jog when checking your wheel? You may be able to see your tracks in the grass and find that the error is in the side to side deviation from the true line. It is very easy to be off if one is not skilled in something that seems so simple. Just liked shooting hoops or running, if you want to get better, you need to practice and refine your skills.

When you mention the soft fairway ground deforming - was the ground actually deforming or the wheel pushing into the grass? This again is an area that seems to be overlooked by the JR Count crowd.

I've only seen the footpod come as an option with running watches/heartrate monitors, like Polar and Suunto. Maybe Timex, and I think even Nike has one too.I didn't want to "steal" the thread with talks about footpods, but current ones today are more sophisticated than simple step-counters, multiplied by some stride length. The accelerometers take something like 1000 acceleration measurements per second in two orthogonal directions, and attempt to reconstruct in realtime the entire path of the foot in a 2D plane in your running direction. Part of the algorithm (dead-reckoning?) assumes that the footpod is flat when your foot is on the ground, and it's looking for accelerations characteristic to a walking and running gait. I saw a PhD paper on it once, but surely didn't read the 100+ pages about it.It's not directly affected by running speed, as such, but indirectly, inasmuch as faster speeds may change how your foot lands.Of course they can detect and measure your cadence.It's fairly accurate walking or running, but there are some implementation issues. It's hard to say exactly, but I think Polar's first implementation used the same algorithm for rejecting "bad" HR samples (e.g. from transmission interference or errors), to reject "bad" speed samples. This causes a delay in detecting sharp changes in speed, such as during an interval workout, or from a stopping start of a workout/race.But as I said, there is a calibration factor, and it can change significantly between roads and trails, and between racing shoes versus cushioned shoes.I liked the gimmick at first, but eventually found I already knew the distance of most of my running routes, and when it really matters, I go to the track, (or some pre-measured circuits).After a certain point, I just stopped changing the batteries.

J.R. wrote:

In additional to what though, my shoes?

1) I feel that such a device would need would first of all need 100 percent accurate stride rate measurement.

2) There should be a very accurate calculation of pace (stride length) from the rate. So far, there might be some that come close on the rate, but none that accurately extrapolate this to pace.

3) Users should be able to precisely calibrate the pace based on rate, for at least two different speeds. For example I should be able to input my pace at 165 and 178 spm.

These could show a very stable base estimation of pace. Beyond them, the device could start fine tuning with acceleration and such.

rekrunner wrote:

I've only seen the footpod come as an option with running watches/heartrate monitors, like Polar and Suunto. Maybe Timex, and I think even Nike has one too.

See, but then you have to add the cost of those monitors. So it's $120... plus another $180? That's quite pricey.

I didn't want to "steal" the thread with talks about footpods, but current ones today are more sophisticated than simple step-counters, multiplied by some stride length. The accelerometers take something like 1000 acceleration measurements per second in two orthogonal directions, and attempt to reconstruct in realtime the entire path of the foot in a 2D plane in your running direction.

I see them as being a measuring device, feel they are pertinent to the thread, and appreciate you contributing about them.

However, all their accleration gimmickry sounds like scientific mumbo jumbo to me, which is why they are easily fooled.

First of all, they do not accurately measure the number of strides, per distance. Therefore everything after that is based on faulty logic. First they need to accurately measure the number of strides, and calculate the number of strides to a pace. The pace should be able to be calculated at any time, especially when the runner is well warmed up. Then, and only then could some various acceleration adjustments be made.

I did some considerable testing with stride rates vs pace a few decades ago and found a very specific relationship, for myself, and for other runners, that varies with the runner. This is why calibration is important.

There is a lot of falsehood around about rate vs length, for example that all runners should run at 180 spm all the time, regardless of the pace or their physical structure, but that is totally wrong, as all these things make a difference. Perhaps this topic is getting too far from "measuring xc courses" though.

Yeah, I guess I assumed that most runners buy a watch anyway. But you're right, a watch with a footpod is about the $300 price range or more. I bought a Suunto T3C with footpod, and that was more like $250.

If you're the kind who likes to read scientific papers, you might want to give this a look -- the footpod is no simple pedometer:

http://plan.geomatics.ucalgary.ca/papers/RossSterlingThesisUofADec03.pdf

There is no doubt that it accurately measures cadence. What is unique is the calculation of speed, distance, and pace. I'm not quite sure what you mean by "strides, per distance", but barring calibration issues, the footpod accurately measures the distance of each individual stride.

In the introduction, he addresses "stride length and stride velocity modeling" and its strengths and weaknesses.

Did your "considerable analysis" go so far as estimate a magnitude of expected error after calibration, and what sorts of things can affect accuracy?

He talks a little about about GPS devices too, and where footpods would be more applicable (heavy forests and areas with tall buildings). My experience is if you want steady pace feedback, the footpod is more stable than GPS devices. But I'm not that familiar with GPS devices (don't have any experience), and I know they have improved recently, and I certainly don't want to hijack the thread again with a GPS/footpod war.

Footpods have some inherent problems, but Polar's S625X implementation added a lot of avoidable errors in the firmware of the watch, and the accompanying PC software. Another 100+ page paper detailing how well that works, and doesn't can be found here:

http://users.telenet.be/wy/sport/S625X/S625X%20Accurate%20Speed-Distance%20Measuremnt.htm

It was a first implementation, so I know things can go wrong the first time, when you don't know what you're doing. I think here, it was that Polar is more an HRM company, than a software development company.

The 1.5 meters (in 100m, 1.5% difference on grass seems surprisingly large to me. Of course, I am thinks of pressing down on the grass with a bike and not with some device that is relatively light.

The side-to-side variation does not seem like it could be very large because the wheel cannot deflect at a very large angle (see the discussion of a bike with large trail and the advantages thereof).

When I measured on grass the readings were less than 100m. If I was moving side to side or the ground wasn't flat the readings would be more than 100m.

I walked slowly.

The fairway grass was very short and thick. It's like a mat, so it's hard to determine if I was pushing the grass down or pushing the ground down. They were one in the same.

But now it's time for you to answer my question. Have you ever compared your wheel readings to a measurement known to be more accurate? If you haven't how can you be so sure your wheeling is accurate?

Dukerdog-

Bingo

You asked the $64,000 question!

Something I have been trying to say for 9 pages!!!!

Could Wejo please make this thread go away now?

I think it's time that we all focus on more productive projects.

26mi235 wrote:

I lived in LA at the time and ran for Culver City AC and sort of knew the folks doing the marathon course.

I knew some of the people on the CCAC.

Way out of my price range! :)

I scanned through most of the paper, thanks. It kept crashing at the end.

I didn't notice anywhere that there was much of a focus on stride frequency, though a good job was done matching the accellerators with stride length. I am impressed with that.

> barring calibration issues, the footpod accurately measures the distance of each individual stride.[/quote]

Which brings up the point, IF it does measure stride length accurately, then why is it off by an average 3 percent difference, i.e. a 6 percent variance? That is 24 second at a 6:40 pace. I have run 17 milers and been within 1 second of my planned pace every mile.

The only way that it could be off that much, provided stride length measurement is accurate, is by NOT considering stride frequency. Otherwise the calculation of pace would be quite precise.

Also, the article pointed out that "no calibration is needed", certainly a major mistake.

I think the error could be greatly reduced by connecting the length predictions with actual stride rate, and allowing the user to calibrate the device.

> Did your "considerable analysis" go so far as estimate a magnitude of expected error after calibration, and what sorts of things can affect accuracy?

My analysis consisted of a notebook of hand calculations. I did consider statistical probability at the time.

Yes, I included calculations barefoot vs with shoes, on the flat and up and down hills.

Aren't gps devices more accurate than 3 percent?

Hopefully they will keep improving the devices.

Our local USATF guy hosted an 8K race recently. The Garmin runners were getting 4.80 miles which means the course was probably 4.69 miles. I went out with my XC wheel the next day and got 4.68. I can see some margin of error, but we are talking over 6%. He is usually within 3% on the roads, but is blaming the error on this course on the lack of standardization of XC measurements. There is a local 3:0x:xx marathon that run this race in 27:0x and was calling it 5ish. Shame.

dukerdog wrote:

When I measured on grass the readings were less than 100m. If I was moving side to side or the ground wasn't flat the readings would be more than 100m.

I walked slowly.

The fairway grass was very short and thick. It's like a mat, so it's hard to determine if I was pushing the grass down or pushing the ground down. They were one in the same.

But now it's time for you to answer my question. Have you ever compared your wheel readings to a measurement known to be more accurate? If you haven't how can you be so sure your wheeling is accurate?

What brand wheel are you using? I find it hard to believe that you would be short especially when you mention deforming the grass surface while measuring - slippage is the only possibility. Was the ground damp when you did your test?

100m steel tapes are always used to set the straights and guide the line painters. The wheels are always checked before staggers/exchange zones are marked. Mistakes waste time and money, and as crazy as it seems the demand for short tracks hasn't taken off yet!

You're welcome. I'm pleased someone is interested in a subject that was somewhat a pet peeve of mine a few years back. I guess you can say I did considerable analysis on the footpod a while back.Are GPS watches with similar features any cheaper?What this guy did as a Masters thesis, and what ended up in the market aren't necessarily the same. 3% error sounds like uncalibrated measures. Both of my footpods have calibration factors, which should usually bring the accuracy to within 1%.One important assumption is that the foot is flat during ground contact. This defines a frame of reference to solve all the other equations. If you toe-strike, run up or down hill, position the pod poorly on the shoe, this causes systematic, or temporary deviations from the calibrated factor. The second paper I linked suggests that a linear calibration curve, as a function of speed, gives better results than a single global factor. This is similar to what you said before, about calibrating at two speeds. (Thinking about that for a minute, there looks like a recursive dependency, since the device requires the calibration factor to determine the speed). I think the speed factor is more a function of foot position on the ground, rather than some inherent direct speed bias.While the overall distance and pace of GPS devices are well within 3%, any graphs of GPS paces I've ever seen, show the pace oscillating all over the place, like a seismograph during an earth quake. Instantaneous pace graphs with footpod devices are much smoother.My opinion is that the price/accuracy of footpods versus GPS are quite comparable, but each have their strengths and weaknesses. I can wear my footpod in tunnels, but I can not import my data into Google Earth.Finally, Polar did introduce a second generation footpod (actually they all come from Dynastream), and improved some of the software algorithms, increased the sampling rate, and generally reducing their own programming errors. But as far as I know, no one wrote an exhaustive follow-up to see how many bugs remain in the RS800, versus the S625X.