Caribou Driver

In my 2010 Rachael's Run road race forum posting, I suggested two methods of comparing races. I wanted to see how a race with two moderate to tough hills (Rachael's at the Air Force Academy) compared with a flat course (St Patrick's in downtown Colorado Springs). Both were 5K's. I suggested two methods: . "Method #1: Find results for the top five runners who ran in both races and compare the median times for each race. Method #2: Find the 25th percentile and 75th percentile results for each race and compare. I think method #1 would give a measure of course difficulty and method #2 would give a measure of who showed up. Perhaps I could use the results of method #1 to adjust the results of method #2."

Method I: Runners placing 2nd, 5th, 9th, 10th, and 12th at Rachael's Run also ran St Patrick's. The 3rd on this list (Tony Briggs) actually improved his time by 1:30 and the 4th (Andrew Woolridge) by 11:40. I'm guessing Andrew ran St Patrick's with friends and/or family (people do that at St Patrick's). Fifth on the list (Jonathan Briggs) improved his time by 45 seconds. Breaking a cardinal rule of statistics, I involked the rule "If you don't like the results, change the model". So my plan was to add two more runners to the list. The next person who ran both races (George Greco) testifies in his St Patrick's road race forum posting, that he made a pit stop during that run. Now it is legitimate in statistics to throw out the "outliers", however, it is not legitimate to throw out more than half of your sample. In comparing these races, method I is a total failure.

Method II: I ran into an insurmountable wall here. The St Patrick's Day "overall" results are not to be found. There's overall male results and overall female results but not the combined results. So method II was also a failure.

I did notice some interesting things along the way. At St Patrick's, the females outnumbered the males 932 to 734; that's more than 25% more females than males. Tony Briggs (mentioned above) moved from the 25th percentile at St Patrick's the the 8th percentile at Rachael's.

I did do the method II percentile calculations on the race results I had and got the following:

St Pat's males: 25th 23:29 and 75th 31:34
St Pat's females: 25th 29:07 and 75th 36:52
Rachael's: 25th 29:03 and 75th 41:31

Draw your own conclusions.