We have all been there: 67km into a 65km race, with nary a sight nor sound of the finish. TIM BRINK set out to see how and why event organisers get this so wrong, or right. And who need the Altitude Adjustment Klap.
It all started with a semi-heated discussion in the Full Sus offices. A recent Absa Cape Epic finisher was still fuming, a fortnight later, that the elevation projected in the event literature had been considerably off. He is one of the three types of GPS-watchers we see in events; the climber, the how-far-to-go-er and the tikkel-tokkel. Our visitor was the first – he plans his day by knowing exactly what is coming from a climbing perspective, adjusting his input and output by judging how hard he needs to push to make X-hundred meters of ascent in the remaining Y kilometres. We tend to find this A-type towards the sharp end of the field, with the middle- to backmarkers focussing more on the distance covered and to go. The rarest of the three, the third mentioned, simply doesn’t care, riding for riding’s sake and enjoying his surroundings over his electronica.
The reality for event organisers today is that just about every entrant will have a swish, ‘accurate’ GPS unit, and will be convinced that their reading by the end of the ride is the correct one, no matter what. Unfortunately, it isn’t that easy.
Let’s start by looking at the different ways it would be possible to measure these routes. The most accurate
would be to take one of those measuring wheels we occasionally see men with hard hats walking along
highways with. Calibrated by the SABS, these are dead-accurate, but logistically, I don’t think we can expect an event organiser to go quite this far. In any case, the distance side of things is the easier of the two to get right with GPS technology; it is the altitude gain and loss that is more challenging. GPS units measure this in one of two ways, or in a combination of both: barometrically or through the mapping on the device. The former uses a little air-pressure gizmo in your device (you will see little holes somewhere underneath it, the latter relies on the accuracy of your GPS to know where you are, and tells you how high you are on the map. The better devices use a combination of both, with an algorithm creating a best-case scenario. Some
firms use a barometric base, with a GPS correction, others do it the other way round (some even use different combinations within their range, which will explain why model A differs from model B).
Why the complication? Why not just use the simple barometer, and get it done with? Barometers are finicky, and affected by temperature and air-pressure change. So if you start measuring at 10 degrees, and finish at 25 degrees, you will have a skewed measurement. The algorithms factor in these changes. Similarly, imagine measuring a route on a Highveld summer afternoon, as the afternoon thunderstorm approaches, and the air pressure drops by 10% halfway through the exercise. Or, a teeny-tiny piece of mud blocking part of the air hole under your unit.
In a similar vein, pure map-based readings are problematic. Take, for example, Joberg2C’s beautiful ride along the edge of the escarpment. If the GPS accuracy is at the best-possible one metre, all good and well. If it is cloudy, or the satellites are playing silly buggers, the accuracy could drop to as much as five metres… meaning that for portions of your dead-flat ride, you ‘drop’ a few hundred meters down the cliff face, before rising again, and dropping again. John Swanepoel’s Swartberg 100 Grand Fondo should be simple to measure – 171km, all on big open roads. And yet, some riders measured 2 200m of ascent on the day, compared to the advertised 3 000m with a whole lot in between. “I headed out three times to
measure the route. I had three different units, with different GPS/barometric algorithms; a Garmin Oregon
navigational unit, a Garmin 910XT and a Suunto, and I drove the route at pretty much race-pace. Then I took the mean from the nine readings, and that becomes gospel.” Even such a simple route challenges; it crosses the Swartberg twice, and runs through a few different climatic regions. “It can be 45 degrees in Prince Albert, and 15 at the top of the pass. So, pure barometric becomes inaccurate, and pure GPS is ‘bouncy’ in the deeper kloofs like Meiringspoort and descending the Pass.”
This measuring thing is a mug’s game. “We take proper GPS units on our trial rides,” says Hendrico Burger, route guru at the Absa Cape Epic. “The bike ones are fine, but really we get more accuracy from the more industrial units. We set them to record the smallest intervals possible, and put in new batteries regularly to make sure we can measure the whole route as accurately as we can.” Therein lies another factor: sample intervals. Very few bike GPS units will be able to sample this regularly for a full Epic or Joberg2C stage, as manufacturers keep batteries small and light. So, your gizmo will choose the most appropriate intervals as you go along, to ensure you don’t run out of juice and memory. In some cases, this could be as much as 30 seconds between recordings. And a lot can happen in 30 seconds: you could comfortably dip 10 metres, and climb back to your previous height, which your unit simply wouldn’t see. Do that 20 times in 100+ km, and you can see how marked differences in altitude gain can creep in. “We have far more angry riders at the back end of an Epic stage, than at the front, simply because the sampling is so spread out.”
Craig Wapnick and his Joberg2C team run a similar system to the Epic – lots of units on the trial rides, download all the stats and mix in a whole lot of Google Earth plotting. “We have found, over the years, that
you can get pretty close to perfection on Google Earth, actually. We ride and record, with Suunto and Garmin and Wahoo, and we take an average out of all of that, but we also back it up with a whole lot of homework on the computer.” Joberg2C is a relatively simple operation to measure. It is linear, and much of it is out on open roads. “Our biggest challenges come when we ride into forested areas,” says Hendrico. “That, and next to cliff faces, where the satellite signals either get blocked or bounced around.” Much of the Wines2Wales routes, which Hendrico also lays out, is measured in these circumstances, but the accuracy is fairly on point, as the area is well-ridden. “We did some research last year, and it turns out that Strava is fast
becoming an accurate option for elevation gain,” says Wapnick. “Especially in areas where there are many
riders, and their database is well-populated. They have so much info to work off, in so many varied conditions, that their averaging is obviously going to be more accurate than us heading out with a handful of units.”
Back to where we started. Maybe we stop giving the organisers a hard time; they are on a hiding to nothing, especially in more remote areas, and on the routes we love to ride, with mountains and forests and all those lekker things. What they can possibly do is tell us what they measured their routes with, when they do quote numbers. “1 250m ascent, measured on a Garmin 810, at 25 degrees,” will give a better indication, maybe, but also gives them a plausible explanation for finish-line discrepancies.