Just wait, I'm about to make you look like the coolest kid in school, in comparison!

In general, I agree with the others who said that all the other variables that affect speed (wind resistance, type of hills, energy level, scenery, etc.) mean that predicting speed/distance based only one variable is rather futile. The first two, wind resistance and hill types, are the ones that can largely be ignored for hiking, but are important factors in cycling; on a bike, you can cover rolling hills pretty close to flat-land speeds because you don't use energy on the downhills and can store it for the uphills, but with hiking, you use energy both up and down.

However, the data hound in me couldn't resist doing some statistical analysis. If I plotted a large number of data points based only on daily climbing and average speed, would the the other variables be washed away in the noise, allowing a trendline to appear?

And, much to my surprise, a very distinct trendline *did* appear!



This is data from two separate tours, covering about 60 days of riding. The X-axis is that day's cumulative climbing (as reported by my altimeter-equipped cyclocomputer) divided by that day's miles. So, sort of an "average grade", that ignores all the downhills. The Y-axis is then just my total average speed for that day.

Google Spreadsheets doesn't show the trendline, but an equation OpenOffice came up with is:

AVERAGE_SPEED = -2.7 ln(CLIMBING) + 22.8

where CLIMBING means feet-per-mile.

My flat-land, no wind cruising speed is around 17mph, so perhaps a more general form of this equation would be:

AVERAGE_SPEED = -2.7 ln(CLIMBING) + NORMAL_CRUISING_SPEED + 5.8

Please use the term NDSL (Neil's Dorky Speed Law) when popularizing this concept!

Neil