sd5782 

Not sure if this as been asked before, but how close is close enough for you on your old tech wheel sensor mounted bike computer. I have maybe 4 or 5 mounted, and recently have been playing with a phone gps app and comparing. The bike mounted ones save phone battery life, and are handlebar mounted for easy viewing. I don’t like a mounted phone bracket.

On a new build with 35mm tires, the default tire setting on the computer read about 4% high. I input another wheel circumference and got about 1% low. I think I will shoot for around 1/2% with error to the plus side. Kinda funny, compared to the old axle mounted clicker style Schwinn one I had in the late 70s. I would be curious to know how accurate that was before the little plastic star wheel broke.

USAZorro

Most of the computers allow for adjustment due to variations in rim/tire size. Info here.

Miradaman

I seem to recall reading somewhere that a correctly set bike computer is more accurate than a GPS app. I use an app as a backup (and to record my route), but defer to my bike computer for my "official" speed/distance.

SpedFast

Ditto-My phone is now my bike computer but I ran the same route that I used to run using a Garmin GPS and made my adjustments according to that. I get much less fluctuation from the BT sensors and smart phone than I do from my Garmin.

sd5782 

I think the app opened the can of worms . I understand the gps apps may be less accurate if riding with lots of turns as the gps pings are less accurate at corners. I guess my old method of recording a known distance with a car and setting my computers actually turned out to be pretty close. The apps are kinda cool and useful I think for setting up the computer. 4% on the first attempt was too far off for obsessive compulsive C&V guys though.

dedhed

I like to go between section corners on a straight road to check it.

iab

I use the clock on the kitchen wall. If it +/- 10 minutes I can live with it.

GhostRider62

I demand precision for speed on my bike computer when doing experiments (Crr and CdA). (I use a wheel sensor, not GPS)

Bad Lag

I have removed the computer and sensors from my bike. I did that decades ago when I realized I paid more attention to the screen than the scenery, more attention to the numbers than the nuances of my daily ride; caring more about the machine than the man.

Get rid of your computer.

blacknbluebikes 

Another angle on this is to quantify the error into the real world: I do a 30 mile ride, and a 1% error equals about 1500 feet. Does that matter? Maybe, maybe not. Probably not.

nlerner

Isn’t this C&V?

gugie 

On roads I haven't ridden on before I like to know where to turn. With cue sheets, accuracy isn't super critical for that use.

If you're wondering how many miles you've ridden, try the approach that @Andy_K and I use. We round up. Say the route is supposed to be 36 miles. With one significant digit, that's 40.

Steve B.

Computer just using GPS positioning, functions by dropping "breadcrumbs" as it where, usually about 1 second apart. I think that's the most accurate a Garmin can do. That means on some courses that are really twisty (mt. bike single track) you get a short course. I seem to recall way back when I had a Garmin 810, I did a comparison with a calibrated Cateye wireless old style computer, the Garmin was off maybe 1-2 miles in 100. With speed sensors now pretty common, as well as the new Multi-band GNSS, accuracy is much improved. Begs the question accuracy compared to what ?. Did somebody take one of those measuring wheels on the courses you ride ?, is that what you compare to ?. If I'm a mile in 20 that's good enough.

davester

I prefer two significant digits. 36 rounds to 100.

rccardr 

I only use Cateye Strada Wireless cyclometers, and input the actual measured circumference of the wheel. It’s reasonably accurate.

bikingshearer 

Whatever my Garmin 130 shows is plenty close enough for me. It's not like I'm setting records out there.

DiabloScott

My first computer was a Push - there were settings for about 3 different wheel sizes I think. Circa 1984... toe clips, exposed cables, and DT friction shifters, and freewheels.


When electronics got better, you could input your tire circumference to the mm... and there were all kinds of arguments about how to get the perfect measurement considering tire size, air pressure, rider weight, etc... everyone wanted the highest accuracy possible, and it was a PITA.

Then when GPS came along, most people said "yeah, +/- 10% is fine".

I have both a Garmin watch and cyclocomputer... I pick whichever one reads higher to post on Strava.

52telecaster

The only time I worry about it is when touring. Sometimes it's handy when you get lost to know which unmarked gravel road to turn on. Other than that I really don't care.

merziac

I only run GPS for evidence if something goes wrong to back up the cameras, lights and helmet, 24/7, 365, period.

Plenty accurate for me and my meager mileage.

It can help win the 1 or 10 million $$$ lawsuit when the army of lawyers tries to pick it apart.

steelbikeguy

I've got some wired Cateye computers on some bikes (Mity 3, Mity 8, Velo 7), and various Avocets on most of the vintage bikes. The Cateyes are pretty good, really. Looking at the manual for the Mity 8, I notice that the calibration number is the wheel circumference in mm. For a 27 x 1" tire, the calibration number is 215. That allows an accuracy of about 0.5%, which is pretty good!

Of course, this is limited by how accurately you can measure your wheel circumference, so you might only be getting 1% accuracy (or thereabouts). The circumference will also vary a bit with the tire pressure, so you might lose another small bit of accuracy as the tire loses pressure or the temperature changes. This is probably good enough for most navigation requirements.

Of course, the next question is: what is your standard for measuring distances, and is it any more accurate? Civil engineers and surveyors know how to achieve higher accuracy, but I haven't seen too many road signs marking distances in tenths of a mile or better.

Steve in Peoria

sd5782 

I guess the impetus for the accuracy question is a one of personal truth. If I tell my wife or someone I went for a 15 mile ride, but it was actually only 14.5, it would just rub me the wrong way. I guess I have a ways to go before I decide to round it up to 20. I’ve heard fishing stories like that too.

RCMoeur 

I still remember lusting after this one pretty girl in 1982...

...wait, I mean her Pacer 2000 bicycle computer, which had 6 buttons and amazing functions.

(OK, she was pretty cute too).

Some times, I'm still think that one of the biggest improvements in riding technology wasn't indexed shifting, but electronic bike speedometers. Those of us that had to estimate mileage via Multitos or cable-driven speedometers back in the early 80s can commiserate.

In the early 1990s, I standardized my bike fleet on an Avocet cyclometer. One head that could move to many different mounts on each bike. just quickly changing the calibration number each time from a handy list. When properly calibrated, accurate to within well less than 1% error.

But how do I calibrate? I can do a rough estimate using a rollout on the driveway (several revolutions, butt on the bike). But to get the final number, I use a calibration course that is 2.151 miles long from my mailbox to the stop line at a nearby collector street intersection.

But how do I know that mileage? By using my motor vehicles, which have distance measuring instruments connected to the ABS or transmission sensors - super-precise odometers & speedometers that are calibrated over several miles of known-good mileposts around the state. When properly calibrated and synchronized with good surveyed mileposts, it'll tick over .000 right at each milepost. In fact, I have to adjust periodically for tire wear. And I verified with more than one vehicle. And my wife likes them, because she can see from the right seat exactly how fast I am speeding driving in big numerals.

So now we have the wonder of GPS and 'smart' phones (I use RWGPS as our club has a group membership). How does it compare? Usually dead on, or differing by 0.1 mile depending on rounding, even with startup & stopping at intersections & destinations. Except if there is an extended on-foot break in the middle of the ride such as volunteer work, breakfast, etc. Then it's sometimes different, with the RWGPS invariably higher. Even though RWGPS pinky-swears that it is auto-paused and doesn't include that obviously-pedestrian-speed mileage in your riding... they often do include it in the final tally, even if they don't ding you for it on the average moving speed.

Am I overthinking this? Yes, and for almost 3 decades. And my carefully-hoarded pile of Avocet 45s (and other compatible ones) is ever-diminishing, so someday I may go all phone-based. So all I can say is: your mileage may vary.

RCMoeur 

Relying on the precision of a gang of long-dead cranky guys in the 1800s with a magnetic compass, optical transit, and metal surveyor's chain.

And an a lot of places, they were pretty darn close - sometimes only yards, feet, or inches off after setting a township line. In other cases, though...


Still, as a civil PE who doesn't have his RLS, I shouldn't give them too hard a time, considering the difficulties of cadastral surveying back in the day.

RCMoeur 

I try to be honest. OK, I try.

The first recitation of the ride describes the mileage as "almost 15 miles".
The subsequent retellings become "15-mile ride", "almost 20-mile ride", "epic ride far beyond a mere 10 miles", and so on.
Depending on rounding and precision... somewhat technically not entirely incorrect.

steelbikeguy

As an engineer, I learned to always carefully check for the accuracy of any measuring equipment that I use. There's also the rule of "significant digits"... never use more digits when taking a measurement than the measuring device actually is good for. If my Cateye is only good for 1% accuracy, it would be useless to say that my ride was 110.5 miles long... it could have been 109 or 111.

There is also the idea of "good enough" accuracy for a given task. More accuracy costs time and money, usually.
Besides... no one really cares how far you (or any one of us) rode.
I prefer to measure rides in the number of donuts enjoyed...




Steve in Peoria