[Back to XC-2004 Page] [Back to TullyRunners Home Page]


Speed Comparison - Footlocker 2004 vs Footlocker 2003

(Bill Meylan - Dec 24, 2004)


A number of inquiries have asked how I determined the speed ratings for the Footlocker Finals ... It was apparent to several people that I used very different adjustments when comparing the Footlocker 2003 and Footlocker 2004 results ... They are correct - and that's because the Balboa course was not running at the same speed level in 2004 as in 2003 - the Balboa course was much slower in 2004.

This is a great example of where the "Time-Tested" theory of race course evaluation can fail miserably ... The "Time-Tested" theory is basically observations over a period of years with a typical conclusion such as "race course X" is 15 seconds faster the "race course Y" ... Now these conclusions may be valid under many circumstances, but they can also be wrong and very misleading when applied with no supporting data ... It was obvious to most observers that the Balboa race times in 2004 were slower than 2003.

Finding the difference in the Balboa speeds (2003 vs. 2004) is very easy ... Simply plot the race data for both years (finish place , race time) on the same graph (the finish places need to be scaled because there were 32 runners in 2003 and 40 runners in 2004) ... Graphs of both boys and girls races are shown below (the first two graphs) ... Now making only one assumption (that the overall quality of all runners was approximately the same in 2003 and 2004), you can directly read the speed difference as the time difference between the two "best-fit" straight lines (the red lines).

For the Boys, Balboa was 22 seconds slower in 2004 ... For the Girls, Balboa was 24 seconds slower ... This means you need to be subtract 22 seconds from the Boys 2004 times to compare them to the times in 2003 ...  Likewise, subtract 24 seconds from the Girls times in 2004.

According to comments on the DyeStat Forum board, Balboa may have been slower in 2004 because the course was changed slightly due to a fallen tree and rain softened several sections of the course ... I don't know the exact reason, but it was definitely slower ... and statistics can tell you exactly how much.

As noted above, the "Time-Tested" theory requires supporting data when it's applied ... An example of supporting data is shown in the bottom graph which compares Footlocker Boys in 2002 and 2003 ... the data points are very close and actually criss-cross several times indicating the Balboa speeds in 2002 and 2003 were virtually identical, so the final times can be compared directly with no adjustment (yes, the lowest data point is Chris Solinsky).

And for those who accuse me of making-up numbers arbitrarily - You obviously have no idea of what I actually do and the amount of time I spend doing it (even though I've tried to explain it in some detail throughout my website) ... I derive only numbers that have a statistical basis where the results are reproducible by people (other than myself) who understand basic statistical correlation (and that's all it is).