I’m on my third cyclocomputer since I started riding. It’s a Cateye V2c, and is the best of them all, because it’s wireless and has cadence as well as the usual assortment of current, average and maximum speed readings. It gives me everything I want to know about a ride (Okay, being able to see the percent grades of my favorite local climbs would be nice too) without involving the steep cost of a GPS unit.

However, I do have a minor gripe: accuracy. I’ve noticed that with each computer I’ve owned, I get different distances for the same route.

My benchmark ride is a trip from my house to the Austin Veloway, three laps of the track, and back home. It’s my 15-mile ride. To be precise, it’s 14.9 miles — 2.8 from my house to the Veloway, three laps at 3.1 miles per lap, or 9.3 miles, and 2.8 miles home. But my computer clocks it at 14.7.

Now, you adjust your computer according to the outside circumference of your tire. A 23 mm tire should have a circumference of 2096 mm, and the instruction book that comes with the computer suggests setting it to that number. But tires flex, and since my computer’s sensors are on the rear wheel, which carries the most weight, I thought I would measure by hand.

I positioned the bike on the garage floor so the rear wheel’s valve was perpendicular to the ground. A black Sharpie made a hash mark on the floor. Then I got on the bike and pushed off with my feet for one rotation of the tire. Used the Sharpie to mark the floor again. When I measured the distance between the two marks, and converted from inches to centimeters, then reduced, I got — 2096.

What I thought was odd was that on my previous tires, which were 25 mm, my hand measurements also came out to 2096. So my computer readings were consistent.

Earlier computers clocked different numbers, though. A Cateye Strada measured 15.1 miles for that benchmark Veloway distance.

I’ve alluded to my finickiness when it comes to record keeping. I had to come up with a way to get accurate readings from the computer. The obvious way is compare it to a known distance.

Several years ago, I zeroed the computer at the entrance to the Veloway. As I mentioned, it’s 3.1 miles around. But a circuit of the track clocked exactly three miles on the Cateye. So if I was a tenth of a mile short for three miles, that meant that I could simply add a tenth for every three miles ridden. If the computer showed I rode 18 miles, I could add six-tenths to that distance for an accurate reading.

This jury rigged system actually worked. When I rode the MS150, the Cateye clocked the second day’s distance at 74 miles. Divide by three = 24.6 — so I could add 2.5 miles (round up the .46), and get 76.5 miles. The official route map showed the distance at 77. Getting closer.

It’s been awhile now, and I’ve been adding my tenths. But I decided last week to check the accuracy of my system again. This time a 3.1 mile lap registered 3.05 miles on the Cateye. Half a tenth off from the previous readings. That meant I could now add a tenth for every six miles ridden, instead of every three.

When I compared a couple of recent rides using this method to maps created on Ride With GPS, the new system came pretty close.

I think you’ve gotta be a certain type of personality to go through the machinations I do just to get what might or might not be an accurate measure of distance traveled. A twisted type. That’s me — accurately.

JC says

My dad has the Cateye V2C too. I have a Bontrager trip 5w. What I noticed is that the V2C requires a higher speed to start recording data than the 5w. It is programmed to start recording only when it hits a certain minimum speed. So part of your trip may not get tracked if you dont or cant(like a car in front) get to that minimum speed required.

Jared says

Seems like you could do this fairly quickly by figuring out what the circumference should be given a known length. So for your example where 2096 gave you 14.7 miles and should have been 14.9:

2096x = 14.7

x = .00701

y * .00701 = 14.9

y = 2124.5

Bike Noob says

Math! Math! I never went beyond high school geometry. Thanks for the calculation.

Wow, this is funny you bring this up. I was just wrestling with the same problem: accuracy.

I currently have that Cateye Strada Cadence you have pictured. A couple weekends ago, I was competing in a triathlon. I was actually amazed by how accurate the distance was. As I passed over mile markers painted on the street, the computer was either right on the money or within 1/100 of a mile. At the end, the computer told me my average was 18.4, which I was very satisfied with since it was a hilly course.

However, after the results were posted, I did the math myself based on the “official” time and distance and was confused to find out that my pace was actually 18.0. This isn’t too far off, but it means that the timer inaccurate. Therefore all speed related measurements would be inaccurate as a result.

Anyone ever encounter this??

GPSes are generally a lot LESS accurate, and RideWithGPS and the like are good, but still not very accurate, definitely not accurate enough for a benchmark — your cateye, once calibrated, is almost certainly more accurate than either a GPS OR a RideWithGPS.

You could do the same ride repeatedly with a GPS and get different numbers each time.

That’s not to say a GPS isn’t useful. I use one all the time, but I always take the numbers with a grain of salt. If a GPS loses satellite connections briefly, or if there’s some atmospheric interference, or tall buildings, or the satellites aren’t ideally positioned, all these factors lead to errors. It’s compounded, when you’re moving.

What if it’s a problem with significant numbers? If you’re odometer is only presenting in tenths of a mile, then you may not be a tenth of a mile off. You may only be off by a couple hundredths of a mile (15.04 rounds down to 15 and 15.05 rounds up?). If that’s the case then you’re only off by less than 100 feet.

@ken – a friend of mine ran a sub-three-hour marathon based on the mile markers and his pace, but the official results (based on the ankle timer) said he was over by like 50 seconds. It would have been a bummer if he had been trying to qualify for another bigger race…

I’ve got a Bell F20. Last year I replaced the batteries and had to reset it by using the numbers (wheel diameter) listed in the directions. This year, I accidentally reset it, and decided to calibrate it by measuring the wheel travel instead. On a certain route I travel, the mileage came up slightly different.

Then I used the GPS on my phone (Samsung Galaxy S), and came up with something completely different.

And then I started experimenting with the different GPS apps – My Tracks (Google), MapMyRide, Endomoto, and Outdoor Navigation all gave slightly different results. The GPS map on MapMyRide actually showed me deviating slightly off the road at various points.

I’m convinced that the the best way is to measure the tire circumference, either with a tape measure, or by marking points on the floor as you wheel it across. But I think another factor is having the sensor as close to the rim as possible. Why? Because I have 2 bikes, both with Bell cyclocomputers, and I calibrated both of them the same way – marking the wheel travel distance across the garage floor.

Turns out that my roadie (1983 Cannondale) with the 27″ wheels *still* gets a slightly different reading on the same route as my MTB (1995 GT hardtail with 26 inch wheels and slicks). The Cannondale has the sensor up near the rim, while the GT has it near the center of the spokes. I believe (not sure how to prove it) that the longer travel distance by having the sensor closer to the actual wheel circumference helps to dampen small, inherent measuring inaccuracies and rounding errors.

And again, for casual riding and exercise – which is what I do – a tenth or two is not going to make a difference, and as Nate mentions above, some of it may be due to rounding errors. But if you’re training, then there has got to be a way to get much more accurate readings.

I’m right there with you in the Note Takers Club.

I hate it when my Garmin gives me two different distances for the same ride, or when I gain/lose overall elevation despite finishing where I started, or how the temperature seems to max out at 102.2 degrees (five times – same high temp. What are the odds?), or the very rare moments when the entire thing decides I’ve moved 500 feet off road and carries that error perfectly for a mile or two before putting me back on the asphalt. I am not a biology expert, but I strongly suspect my caloric meter is at best a wild guest. Sigh

Regarding your difference in distances when riding around the veloway: Are you consistently measuring distance around the inside circumference or the outside circumference of the track? Or do you randomly move from inside to outside perimeter and back and forth? If you look at track meets, you will notice that the runners all start at staggered points with those on the inside lane starting farther back and those on the outside starting farther forward. If you are going to be consistently accurate, then your measurements should be taken from a consistent lane position.

When I measure, I try to stay as close as possible to the center white line.