Megastat time! With the results from all 4 days in, it’s time to compare and contrast, reflect and debate, and generally guesstimate at how absurdly out of touch these metrics are with the pains and hard graft on the road itself. (Hint: I am devising an equation)
Before we start, I should again stress that these are based solely on the available GPS data and don’t reflect the ride. Most importantly, they don’t account for any stopping durations, and so things like lunch/breaks will necessarily affect data. However, if we assume similar breaks per day then there is reasonable accuracy in a relative comparison. Apart from the last two, the graphs are not corrected, so use the raw GPS ‘crow flies’ data in kilometres per hour, ignoring rest breaks. These four graphs also use subsets of data which have been fairly roughly selected to correspond to hourly patterns, and so again are not hugely reliable.
First up, a linear representation of speed measures from each of the 4 days. Throughout I have (I hope) kept the same colour scheme of blue, red, green, orange respectively. Remember to click through on the images to see the full size version. Something immediately noticeable is if you imagine smoothing the graph out even further, i.e. draw a straight line of best fit through each day, the first two days slow down as the day goes on, while the second two days speed up as they move into the afternoon. Something about the food?
Notably, there is actually a highly consistent average speed throughout the four days. Comparing the mean of each day against the mean of the means for each day, you only have a variance of between -.82mph and +0.49mph. Essentially this means that when you look at the average speed of the entire ride, day 1 was 1.55% below average, day 2 was 7.73% above average, day 3 was 13.01% below average, and day 4 was 6.83% above average. I don’t think this matches the qualitative feedback from the riders!!
Onto the second image. This is the same as the first, but with each day overlayed onto one another rather than spread across the x-axis. While there are obviously data forcibly removed from this view, I think it gives a good indication of consistency again - the vast majority of points are within the 6-12km/h band, and if we apply some correction to that (i.e. increasing speeds by 25% and guessing that you could halve the impact of the down-points to take account for rest breaks - say, a drop from 12km/h to 6km/h may only be a drop from 12 to 9 - then you end up with quite a consistent ride pattern.
This third graph is exactly the same as the previous, only it’s had some semi-correction applied by multiplying all values by 1.25. This is an arbitrary correction based on the knowledge that the actual ride was around/over 250 miles, whereas the GPS/crow-flies mileage came in at 200. We’re still in km/h here though.

For those born before Decimal Day, this graph is the same as above but converted into miles per hour. Remember it’s only semi-corrected: there are lots of other factors unaccounted for which are reducing the absolute speed calculated here, but as I wasn’t on the ride, I’ll wait for the girls to fill us in before we post a fully-correct version!

Posted: May 28th, 2008 under the ride.
Comments: 2