Statistical irrelevance

[slowmotion]

Slightly off topic but can a speeding cyclist trigger a speed camera to flash as they go by.

Groups of furious and reckless riders stand a better chance of tripping them than a lone poseur.

 

[classic33]

This wouldn't be cyclechat if people didn't pick nits!

Thought that was the job of the Nit Nurse?

 

[classic33]

Slightly off topic but can a speeding cyclist trigger a speed camera to flash as they go by.

Done it on the A629, down from J24, M62.

 

[lazyfatgit]

Perhaps, but I tend to be very consistent in what lines I take on the roads in my commutes, because I'm used to the route, and so I pick lines that avoid the bumps (since I know every one of them).

Maybe it's the dips into the potholes causing the variance. I was hoping you'd discovered some sort of bicycle powered time travel device.

 

[Shut Up Legs]

Maybe it's the dips into the potholes causing the variance. I was hoping you'd discovered some sort of bicycle powered time travel device.

Oh, I have! I get on the bike, start pedaling, and somehow when I get to work it's 1 hour later. Bizarre but true.

 

[Ajax Bay]

Went for a long ride today.
When I planned it on "Ride With GPS" earlier in the week it came up as 105.3 miles.
When I actually rode it today, my Cateye Mity 8 registered it as 105.7 miles.
That's a statistical variance of 0.3799%.
(For anybody under the age of 30 reading this, a Cateye Mity 8 has a wire that connects the computer thingy on the handlebars with the sensor gizmo down on the front fork. No Garmin GPS twattery for me

[QUOTE 4960653, member: 9609"]what figure do you input for the catseye? My catseye only accepts cm[/QUOTE]
OP - I find the RwGPS distance figures are very accurate. I ride with a Garmin 500 (which uses ""GPS twattery") and the distance of the RwGPS route is invariably within 0.2% of the figure my Garmin says I've done (average ride distance this year = 184km and 1907m of climb). If I have stopped during a ride for a long period (ie more than an hour) the Garmin sometimes records some distance as having been done (I do not use autopause) as the GPS readings tell it the bike is moving round a bit (while it's locked). Alongside my Garmin 500, my Cateye Velo 2 wired cyclocomputer gives me a different figure, does not go 'walkies' outside the cafe, and by design is 'autopause'(!). I use 212cm on mine: wheel 622-15 and tyre 622-25 at 75psi (actually measures 26.4mm width (but it's the height that matters - which is more difficult to measure accurately).
But this is not the source of your 'error'/"statistical variance".
To get your Cateye Mity 8 to give you a more accurate reading, you need to choose the correct 'circumference' setting. This will depend on your wheel size and tyre size, and as @User9609 has said, many Cateye models, including yours, can often only be set to the nearest centimetre (so immediately a possible 2.3% out even with the closest setting). Here's the Cateye Mity8 manual - I suggest you reset yours, and then, by experiment, find the setting (up or down from the circumference the table (excerpt below) suggests) which is closest to what RwGPS says the distance is. Or borrow a GPS twattery device, briefly, to calibrate your Cateye over a straightforward 20km ride.

700 x 25C 211
700 x 28C 214
700 x 30C 217
700 x 32C 216
700C Tubular 213
700 x 35C 217

 

[Tim Hall]

Surely a 1cm precision over a 212mm circumference is a 0.5% error (or thereabouts) rather than 2 - 3%?

 

[Ajax Bay]

Surely a 1cm precision over a 212mm circumference is a 0.5% error (or thereabouts) rather than 2 - 3%?

But if your best two options (to match the actual circumference) are say, 212cm or 213cm, then one/both of those options will be a maximum of 0.5cm 'wrong', which is 2.4% (of 212cm). That was the basis of my maths.

 

[Tim Hall]

But if your best two options (to match the actual circumference) are say, 212cm or 213cm, then one/both of those options will be a maximum of 0.5cm 'wrong', which is 2.4% (of 212cm). That was the basis of my maths.

0.5/212*100=0.23%. I think you're a factor of 10 out.

 

[srw]

0.5/212*100=0.23%. I think you're a factor of 10 out.

*ahem*

212mm

(But yes, you're right. There was a discussion on the subject over on the RideLondon thread. I don't think it's meaningfully possible to measure distance accurately to within a tolerance of more than about 1%, whether with GPS or wheel computer - the GPS signal accuracy isn't enough to compensate for the identified inaccuracies of the wheel computer.)

 

[Domus]

Statistically, 6 out of 7 dwarves aren't Happy.

 

[MiK1138]

This wouldn't be cyclechat if people didn't pick nits!

Ah but how many nits?, and at what rate?

 

[Gravity Aided]

All of this reminds me of the fine book The Third Policeman, by Flann O'Brien, or Brian O'Nuillian. In which many metaphilosophical aspects of cycling are discussed in a couple of haunting chapters that will change your entire attitude toward bicycles.

 

[classic33]

All of this reminds me of the fine book The Third Policeman, by Flann O'Brien, or Brian O'Nuillian. In which many metaphilosophical aspects of cycling are discussed in a couple of haunting chapters that will change your entire attitude toward bicycles.

Or both!!

 

[Alan O]

All of this reminds me of the fine book The Third Policeman, by Flann O'Brien, or Brian O'Nuillian. In which many metaphilosophical aspects of cycling are discussed in a couple of haunting chapters that will change your entire attitude toward bicycles.

Indeed. For any statistical measurements to be useful, one should take into account the relative proportions of atoms of bicycle and atoms of cyclist merging between the two of them, and that would need to be carefully ascertained at both the beginning and the end of the ride.

(Citation: BMJ - http://www.bmj.com/rapid-response/2011/11/03/de-selby-theory-requires-repeated-weighing)