Ground control to Major Tom...

If you follow this blog, and I doubt you do, you know that I have spent the past seven or so months building a new bike. The bike is a classically styled machine, and none of the computers I have in my vast parts bin looked right on it. Sure, I could go full L'Eroica, dispense with the computer, navigate by the stars and wear only wool cycling clothes from now on, but screw that. I started riding in an era when bicycles had computers and they are an integral component of my riding life. I set a number of distance goals each year, which is part of what motivates me to get up on a 35 degree Saturday morning and go for a ride. I'm not going to give that up just because I got a new retro looking bike. So what I wanted was to find something small and unobtrusive that wouldn't look completely stupid on the bike.

As with any purchasing decision of a bike computer these days, one of the first questions to come up is to GPS or not to GPS? I'm not a serious data weeny and have resisted going to GPS computers as several of my peers have over the years, mostly because they're so large and horrible looking. Yes, I want to collect data and know how many miles I've traveled so I can set goals and effectively use a cue sheet. But while I have thought it would be cool to have some of the nifty navigation features several of my friends have, I've been content to do some prep work (cue sheets) and I don't worry about getting lost because I always have Google Maps on my phone if I need to figure out where I am in a pinch.

Teeny, tiny Garmin Edge 25

I don't do Strava, or any social networking for that matter, so up until now there just hasn't been a compelling reason to get a fancy schmancy GPS bike computer. I don't much care about heart rate, cadence or power. I don't want to read my email or news or stock quotes while I ride. It would be neat to let the Managing Partner track my whereabouts so she can come pick up my corpse when it comes to that, and get turn-by-turn instructions by uploading predefined courses; but I've never had those features before, so up until now they haven't persuaded me to become satellite dependent.

Still, Garmin has a super tiny GPS model, the Edge 25, which I found very appealing. It checked all my boxes: small, relatively inexpensive and simple to operate. Some of the Cateye computers are as small, but the wireless units require a sensor on the fork. Have you seen these forks? I'm not about to ruin those lines by zip-tying a tumor on them. So after reading and watching several reviews, I decided to go ahead and get the Edge for the new bike figuring what they heck - it's smaller than the super integrated ANT+ computer I used on my recently deceased carbon fiber mid-life crisis bike, and I think it's even cheaper. So I pulled the trigger and put a Major Tom computer on my 1950s style bicycle.

And generally I'm pretty happy with it. But...

Before putting the Garmin on the new bike, I thought it would be fun to mount it on my beloved aluminum racing bike and compare the performance to the Planet Bike Protegé 9 computer, which I have used and loved for years. The Protegé 9 reported distance very close to that reported by my Bontrager Node2 and the Cateye Astrale that preceded that. 

I have had several computers on my bikes over the years and they have all produced similar results. Whether wired or wireless, they've all calculated distance by sensing the rotation of the wheel and multiplying by the circumference. All have measured distance and speed pretty closely on a variety of courses. Until now. What I found is that the Garmin measures distance just about 4% shorter than every other computer I've had (or the Managing Partner has had) for the past 15 years or so. Or roughly 4 fewer miles per 100.

One of these things is not like the other

There is some grumbling on the Forums about the accuracy of GPS generally and the Edge 25 specifically. Most of the griping seems to come from mountain bikers, and I can see why. In general what I gather is that GPS "samples" location coordinates periodically and draws straight lines between the points. Note in the image below how each of the corners has been cut off because a coordinate was taken before entering the turn and then again only after the turn was complete. This shortens the overall distance traveled which would account for the disparity between GPS and an actual count of the number of wheel rotations. If this is a significant contributor to the disparity between "physical" calculation and GPS, winding routes would be particularly affected, explaining why mountain bikers are the most vociferous gripers.

Turn, turn, turn

Generally a 4% variance wouldn't be a big deal, really, since I'm riding for pleasure and not ego. But in a 3,000 mile year, this disparity will cost me about 120 miles. Again, not an earth shattering loss, but still. That's 12 hours of my life, give or take. I have yet to measure against cue sheets supplied by organized events to see if this has an impact on my ability to navigate with them, though I've often seen disparities with cue sheets over the years. So even that won't be too big a disappointment. Hell, for all I know the Garmin is right and all my previous computers were over-reporting my distances. I'm not going to sweat it too much. I like the form factor and its core functions so I'll stick with it until something better comes along. And in my mind I'll tell myself after each ride that I actually went farther than the confounded machine says!