I've been thinking lately - never a good idea but better than taking zoppies when sleep is slow to come and more effective than counting sheep...
I have often been asked why I don't pre-ride the routes to measure the distances between waypoints and "adjust" them in the roadbooks. I will attempt to explain why such an exercise would be pointless given the combination of technologies we use.
The net result would simply exchange one set of systematic errors for another...
There is, however a navigation technique that works around these systematic errors which is described below.
Take a deep breath, relax and read this two or three times to let it all sink in....
There are two essential quantities in Time Speed Distance Regularity motorsport, time and distance, we know that - but why do the roadbook distances never seem totally right?
In my previous professional career, I was a Healthcare Scientist. One of the people who look after "the machines that go Ping!" to quote someone else, to calibrate them, hook them up to patient-people and make relevant measurements of their physiologies. Instrumentation, calibration and measurement was one of the big things. So, you see I have a bit of a history with instrumentation and calibration.
Part of my thinking about the "odd distances" problem has been to compare and contrast what we do on bikes, in the woods and mountains under the heading of TSD and what the good people in cars do when they TSD.
The outcomes of my ponderings has led me to the understanding that, for us at least, using a combination of technologies (GPS distance measurement, conventional cartography when preparing the roadbooks - Google Earth and flat-map projections of distinctly three dimensional terrain and wheel-rotation odometers) there is an "event horizon" of distance measurement accuracy (not precision - see below) that we cannot overcome. If you can persevere, I will explain.
Measurements of Time.
In the days before internet and GPS, the measurement of time was always done by "timepieces" that had to be synchronised for event timekeepers to use. No big deal, but watches had to be wound, electric and then electronic clocks had to be continuously powered.
These days it is easy as everyone uses GPS time. So accurate and precise, the Relativity effects of satellite clocks in orbit around the Earth have to be adjusted for. Everyone's phone, tablet and computer run to the same time standard. Who wears a watch these days rather than swiping up on their phone to see the time?
We don't need to worry about the measurement of "time" in our events as GPS time solves all those problems for us!
Accuracy versus Precision.
In conversational speech, these two words have almost identical meanings, but for the scientific and TSD communities they have related but distinctly different meanings.
A measurement or set of measurements can have accuracy, precision, both accuracy and precision or neither......
Let me explain. Say we are attempting to measure a "quantity". The quantity can be anything. We are talking about distance here, so let's think about the measurement of a running track. 100 metres.
Accuracy/Exactness
When marking out the track, the runners will want the track to be exactly One Hundred Metres... not ninety nine, not one hundred and one. Accurately, 100 metres. The target of the distance measurement is the length of the running track and the people who mark out the track want to measure it's length accurately. They will want a tape measure (or some other form of distance measurement that will reliably measure the length of the running track.
There is another "quality" of the measurement they will be interested in though
Precision/Repeatability
How close to 100m will those who are setting the course want to be? How Precise? Running tracks are usually laid out with painted lines for lane separators, start and finish lines. How close to 100 metres do they need to be? Well an error of 1 metre is obviously not good enough (1 part in 100). The painted lines would be 100mm wide so if the start were measured from the start of the start line and the finish at the far side of the finish line - and the measurement was accurate, the error would be 1 part in 1000 (100m +/-10cm) How about if the course setters wanted to be more precise yet and they attempted to measure from the start of the start line to the start of the finish line but the edges of the painted lined were indistinct or wavy and there was a 1cm doubt about where the lines actually started, the precision would be 1 part in 10,000 (100m +/-1cm).
If these people were going to measure to a precision on 1/10,000 they would want a measuring method that would reliably give the same result, not a rubber tape measure calibrated in centimetres that would stretch when pulled, not a metal tape that would shorten or lengthen depending on the temperature of the day. They would also need a measuring system that could be replicated anywhere in the world. This is "repeatability" - making the same measurement of the same thing in the same way and getting the same result, or at least a very small range of results.
We have "accuracy", synonymous with "exactness", we have precision as an expression of the fineness of the detail that can be measured and "repeatability" the likelihood that a given measurement on one occasion will be the same on another occasion when measured under the same circumstances.
When thinking about all this it is easy to find yourself going down one of the "rabbit holes" spoken about so often these days...
The measurement of distance.
I have written before that I have been told (and have no reason to doubt) that the car people expect to win a TSD event with zero time penalties (every time control point hit to the second) over a course that exceeds one hundred miles in length. Now there are some very good reasons why they are able to do this - having a co-driver to call the turns/manage the speeds/having the location of the CPs identified in the roadbooks/having the ability to claim "time allowances" for when they are delayed en route, for example, but....
There are also some very good systematic instrumentation and calibration reasons for their precision and accuracy that we in the wild and woolly wilderness will not be able to even begin to replicate.
There are arguably three different ways to measure distance travelled in a moving vehicle, and two related ways of measuring "distance to be travelled" in a roadbook.
The three ways of measuring travel distance in a moving vehicle are:
From GPS data
From the front wheel speedometer gearbox, measuring wheel rotations
and
From extra equipment attached to a roadwheel counting wheel rotations (perhaps these second two are variations of the same thing?).
Wheel driven systems
To take the speedo drive first, there is a little gearbox attached to the wheel spindle and every time the wheel rotates, the gearbox rotates the speedo cable a given amount. So many wheel rotations equal one rotation of the speedo drive cable, so many rotations of that equals one mile or kilometer and so distance travelled along the road is measured. This is converted into a speed indication by the speedometer meter head but we don't need to go into exactly how.
Of course, if the wheel rolling radius/diameter were to change (different tyre pressure, tyre wear, change of tyre) the calibration of the odometer (distance meter) would change. Also, by law, vehicle speedometers have to meet certain calibration standards, but whether that is achieved in the distance measurement function or the meter head calibration isn't important. What is important is that the speedo/odometer is nearly always going to read wrongly!
The car TSD people finesse their "along the road" distance measuring instrumentation by attaching some sort or another of wheel-rotation detection hardware. It might be a mechanical device, electric/electronic or an optically triggered system, but it provides a signal that can be counted, and speed as we know is distance per unit time (and time is already sewn up...).
In their eternal search for accuracy, precision and repeatability, the good car people seek to control as many variables as possible. They will be obsessional about tyre and road temperatures, tyre pressures, tread depth, rolling radius, road surface and any number of other parameters. They will probably have a set of wheels and tyres that are only fitted for TSD events. They will have carefully calibrated their instruments to road distances that have been accurately and precisely marked out on a test road to control for as many variables as they can think of.
In the effort to separate increasingly competitive car TSD participants, event organisers have had to engage in a technological arms race with participant competitors for increasing accuracy and precision when laying out, measuring and writing roadbooks. Event organisers, for car events at least, are required to check and recheck the distance information they publish with technology at least as equal in accuracy and precision as the technology used by the participants. Measuring "along the road" distances...
Both these wheel-driven systems measure the distance travelled along a road. This is an important point that we will return to shortly...
GPS-driven systems
GeoPositioningSatellite systems are the ultimate "eye in the sky". From whichever Earth orbit your GPS Satellites are in, they look down on you and follow your movements across the face of the planet.
Your pocket device detects the data broadcast from the satellites and works out where you are. You move, it detects where you have moved to. It works out the difference in those two positions and it says you have moved so far. Catch is, it assumes you have moved in the horizontal direction only. Unless you have a very sophisticated system (not smartphones) GPS doesn't know how far up or down you have moved. Ok so if you have a mapping system that knows what topography you have covered, the mapping system might be able to work out your road distance from your horizontal distance, but I don't think Rally Navigator is there yet.
"What are you talking about....?" Some of you have already got this, but if you haven't::::
Maps are two dimensional projections of three dimensional surfaces. The mapping outlines are drawn as if seen by an observer at very high altitude looking down.
The map representation of a horizontal (level) road is a true representation of its actual length. A 1for1 representation on the flat page (scaled of course to fit).
As the steepness of the road increases, the two-dimensional representation of the road in three dimensions shortens the projection of the road (by the cosine of the angle of the road inclination in degrees).
If you take this to the absurd limit, if you had a vertical cliff that was 1000metres high and you laid a road down the cliff face (if ever!), the road would be 1000m long, but the projection of this onto the flat, two dimensional map would be just a thin line. Get the idea?
The car TSD people have their events on mostly flat, everyday tarmac roads that on the whole present no great challenge in completing.
The events we are concerned with cover roads that at every possible opportunity are as far from flat as they can be.... bwaaaahaaaahaaaa!
But how great are the differences between "on-road" distances and GPS flat surface projections?
Well, take a hypothetical road rising at an angle of 25 degrees from the horizontal. A steep road but not impassable by a cross-country bike. Say the length of this road was 10km (to keep the arithmetic easy). The map projection of the length of this 10km road would be 9.06km. In round numbers a 10% error. In those 10 road kilometers you would have risen 4,220m in altitude, somewhat unlikely, but if you went up 5km at that steepness then came down the same amount, the result would be exactly the same. You have apparently lost 1km when measured by GPS!
Maps, mapping apps, Rally Navigator and GPS
As I have already mentioned, the cartographic projection of three dimensional topography onto a two dimensional map shortens road distances when the going gets steep.
If you are plotting a nearly flat surface/road/feature onto a map the difference between actual dimensions and the cartographic representation will be small and probably negligible. Start mapping steep country and the differences quickly become appreciable.
Paper maps, mapping apps (BRMB, Gaia, Trailforks and everything else), Rally Navigator (for writing roadbooks) and GPS for measuring travelled distances all use this "cartographic projection".
When writing roadbooks, Rally Navigator will give precise (see explanation above) distance measurements between waypoints based on the "breadcrumb trail" of the clicked route to be taken.
The precision is to the nearest 10m, ie 10 in 1000, 1% precision. This fits well with the Richta GPS timing system that has a similar precision.
But, we have already seen the accuracy of distance measurement in hilly country can be out by 10% or more...
The indication of precision suggests an accuracy that simply isn't possible.
It has been suggested in the past that I ride the routes to measure and correct errors of distance measurement... I do pre-ride the routes to check the accuracy of the direction instructions given in the roadbook but it would be futile to try to more accurately measure distances between waypoints, say for example by using a wheel-driven odometer for the following reasons:-
The majority of riders use GPS driven odometers (Rally Roadbook Reader and others). If I were to correct the roadbook distances between waypoints by on-road, wheel driven distance measurement, the distances for GPS odometers would be just as wrong as for wheel-driven odometers used on GPS measured distances.
Speedometer-drive odometers are notoriously inaccurate, their inaccuracy can equal the difference between road and GPS distance measurement in hilly country. They are incapable of user-calibration and their precision is limited to tenths of a kilometer (100m in 1000m).
Precision, wheel triggered odometers, as used by the car TSD people do not exist for motorcycles. Therefore if any on-road distance verification were to be attempted, the instrumentation with which to do this would have inadequate accuracy and precision.
If I were to attempt to verify the "truth" of map projection distances with GPS odometers, given both systems use the cartographic projection system, the results would be the same by both methods - achieving nothing.
Event participants have the facility to calibrate their GPS odometers (in Rally Roadbook Reader and others for instance) on a flat road of known distance and should understand the systematic error that will be introduced when they ride any appreciable incline.
This is the "event horizon" limit to distance measurement accuracy I mentioned at the start of this article. An insurmountable limit to the accuracy of distance measurement imposed by the combination of technologies and methods we use in the circumstances we use them.
None of these limitations figure in the car/road TSD world. The car people apply themselves with obsessional attention to odometer calibration detail.
We are developing a unique version of TSD over steep, rough terrain and in the process discovering the limits of the conventional methods and developing solutions unique to our application.
The "Wilderness TSD" solution...
Having gone to some lengths to explain why distance measurement between waypoints in steep country is bound to be less accurate than in flat/horizontal terrain, and therefore accurate (that word again) speed control has a different meaning on steep roads compared to flat roads, there needs to be an understanding of how to best approach the set navigation and speed tests.
As far as speed and target time achievements go, the precision of expecting to pass the Time Control Points "to the nearest second" can be dismissed. The exact location of the CPs will not be given in the roadbook to prevent attempts at gaming the system. Participants can reasonably expect there to be a CP in the vicinity of a speed change, but it may be anything upto 400m before to 400m after the speed change. This on legs that may be in excess of 10km long. Expect 100m in 10000m ie 1 in 100, and of course it will be the same for everyone.
Advice: ride the roadbook, not where you think the CPs are!
Distances between Waypoints.
Distances between waypoints, especially over steep country will be an approximation of the actual distances. If it says 10km you have to know the road distance will be at least 10km... If you are using a GPS odometer and the roadbook is using the cartographic projection, your odo reading will closely match the roadbook distance. If you are using a wheel driven odometer, don't expect a close match...
Navigate from waypoint to waypoint, resetting your odometer to match the roadbook distances as you go at every "waypoint of significance".
Calibrate your odometer before you arrive at the event.
Understand why the roadbook distances can never exactly match what your odometer says - it probably is your odometer that "misreads"!
Finally, while the aim of very good car TSD participants is to return a "clean sheet" of time penalties, motorcycle Wilderness TSD participants can never expect to return anything like a clean sheet, the winner is the rider with the lowest score. But at least there will never be a dead-heat result!
Comments