Capture for the ever-growing To Do List: during the drive back from the Bachman Valley Half Marathon, statistician Don Libes and I got to talking about quantitative running relationships. There is much "folk wisdom" and lore about training — ramping up weekly mileage, doing hillwork and speedwork, tapering before a big race, eating certain foods, etc., etc. Almost none of this has been properly assessed. Do intensive training regimes actually destroy more people by causing injury than they help? What are the correlations between key parameters? How can someone optimize performance, depending on what goal(s) they want to achieve? What are the trade-offs and the statistically significant rules-of-thumb that people should bear in mind when they try to improve?
Data exist. The Montgomery County Road Runners and similar groups have, for many years, sponsored group training programs. Participants and coaches could be surveyed, before and after. Results in races could be correlated. Drop-outs could be interviewed to find out what happened to them. As far as Don and I can tell, none of this has been done. Instead, it's all anecdotal, and we all know the saying about "The plural of anecdote is ...".
Step One: identify key parameters, outcomes, and candidate connections to investigate and model. Among the obvious ones to explore and define:
(cf. NeedForSpeed (2002-08-10), DependentVariables (2003-04-09), DecelerationParameter (2003-12-28), RootMeanSquareDance (2004-04-24), StartSlower (2006-11-02), Weight Management (2009-04-13), Year of Running - 2009 - Further Observations (2010-02-01), 2012-03-04 - B and A Marathon (2012-03-15), ...) - ^z - 2014-09-30