Monday, December 20, 2010

Till We Meet in Hell

Well I fumbled the ball with Blackstrap Hell but Valerie scooped it up and scored a touchdown with the Till We Meet In Hell race. The format was simple - run around a 1.15 mile trail loop for 1:30. But the race was handicapped by race times from Bradbury. More on the handicap system later.

The course was located adjacent to the USM-Gorham campus and featured a mix of cross country trail and single track with about 150 feet of elevation difference. The single track had enough circuitousness to scrub some speed but the real speed killers were the several short, steep ravines with the path moving straight down and back up the fall line.

I mistakenly believed that the course would be flattish, like the Bradbury Scuffle and Bruiser course. Why I'm not sure since I've been to USM Gorham enough times to know the terrain. I took all runners who ran both the Bradbury Scuffle and Bruiser and for each runner used these data to interpolate the distance they should run in 1:30. I then computed separate quadratic regressions of this distance on Scuffle, Breaker, and Bruiser times. I then used these functions to estimate the distance that anyone who ran any of these races should run in 1:30. Since most of us had more than one estimate, I used the maximum estimate to avoid sandbagging. Valerie then used the ratio of the entrant's distance/Blaine's distance (since Blain was the entrant with the longest projected distance) to come up with a multiplier for our actual distance. Multiply each person's performance by their own multiplier and voila - we have a handicapped race.

How good were my projections?

Runner Actual Predicted Breaker Adjusted
Blaine 10.9 12.3 11.4
Andy 10.8 11.7 10.8
Ryan 10.5 11 10.2
Jeremy 9.8 10.1 9.4

Not bad, especially given the slowness factor of the TWMIH course (and I don't think Blaine has been training optimally because of a running injury). Had I thought about this before, I would have used the "Breaker Adjustment Factor", which is computed as the average of Breaker Results/Predicted breaker results using Scuffle+Bruiser results.

1 comment:

  1. Even though the distance prediction wasn't right on, I think the ratio was just fine. Although, for you, it might have been fun to use the Breaker to see how close you could come, it worked. If you are bored on break, perhaps you could see how the Breaker would have done!