Monday, January 21, 2008

FLASHBACK: March 12, 2007 - Publius and boobies

In the last of the series of retro-Publius, the individual continues to express his love of science (among other things)

Date: March 12, 2007 2:01:22 PM EDT
Jennifer Cooley, Cindy Cooley, Nelson May, Jill Keller, Cantey Land, Meghan Marsano, David Hogeboom, Bruce Wagoner, Phil Chamberlin, Jin Woo Jun, Sean Welsh, May King, Jody Dennis, Tim Long, Joseph Asaro, Franco Ordonez, Keith Mrochek, Josh Saak, Ben Calhoun, Monica Ferkowicz, Bevin Jett, Mary Franceschi, Mary Jo Arrington, Renee Simcoe-Martin, Joseph Madera, Liza Beth, Lisa Palermo, Jocelyn Sikora, Ian Seeney, Pete Kaplan
Subject: HOORAY FOR BOOBIES! (and regression analysis)

Best album title - ever. Even The Todd (from Scrubs) would agree

“I have to go, there's a breast reduction on the fourth floor... I'm gonna go try and stop it.”

The MAP Sprint Triathlon on Saturday mostly went well! This is my fifth year doing it, but the first three were in uptown. Last year I finished at 1:17. The 500 yard swim did not go all that well. My left eye cup lost suction and then I had to roughly pass three people near the end (two of which turned out to be Mona’s friends…) I finished at 10:05, five seconds worse than last year, placing 226th out of 332 male finishers. The big difference this year was my sweet new Trek Madone, named Holman. We flew at 20.5 mph over 20K to cross at 36:31, a huge improvement from 40:15 last year. I continued the downward plumet placing 109 out of 331 male finishers. The 5K was a reasonably strong run, finishing at 22:29, another couple of minutes faster than last year. For the run that put me at 93rd out of 327 male finishers. Final time 1:11:21.

Before the results were released, it looked like I had a reasonable shot of breaking the top male 100 for the first time. That didn’t happen, but 109th (0.33) is my strongest finish ever for a full NCTS race. Overall 126/477 (0.26).

I am not going to get all worked up about it, because that race is really meant for me to just start thinking “triathlons” and my goal is very different from the spill-your-guts-out approach of a sprint since I need to go EIGHT times the distance for each in August.

My main interest here is the relationship between bib number and resultant swim time / resultant total time. Bib numbers were assigned linearly from fastest submitted time to slowest. (This is probably not true for the first say 25 elite athletes, whose numbers are often assigned on previous standings or results from last year). After watching all the cluster fucks / passing / scratchings, it’s clear a lot of people aren’t in tune with how fast it takes them to swim 500 yards. Or is it? Attached are some summaries of my analysis.

I did a regressional analysis for the two independent variables. In each the dependent variable is bib number. The charts show a basic plotting first of actual swim time versus bib number (top) and actual total time versus bib number (bottom). Even Todd can see there is evidence for linear relationships. The regression results for SWIM are shown, with a R-squared of 0.42 and some fairly suspect p-values. The next tab shows TOTAL results – the R-squared here is 0.38 which shows there is (obviously) a little less of a tighter relationship. This makes complete sense, given my relatively poor pool swimming ability (bib # 419) versus the dozens of people I passed on the bike and run. You can see I beat two people in the “top 25,” outlasted 13 people in the top 50, and 35 in the top 100. Only 2 people with a higher bib finished better. The first tab with results shows predicted SWIM and TOTAL times based on the proposed intercept and x variable. What would be more relevant per se are a suite of t-tests on predicted and actual swim times, but I don’t have those. And how about comparisons of the two dependent variables? CONCLUSION? People are sorta clueless, but maybe they know something.

This is a cursory analysis fitting for a mediocre tenth grade science project; but alas I do have an actual baseline report deadline this afternoon.


p.s. speaking of the Discovery channel, there was a very interesting piece on army ants yesterday. It reminded me of a McGyver episode where he made a protective suit out of a hose to defeat a massive army ant invasion.*

*okay I was ten at the time and it was a scary episode

No comments: