Post Position Regression Analysis

[Cangamble]

The first book I ever ordered in my life was at a University book store. It was Winning at the Races by William Quirin. He was definitely ahead of his time.
In a later chapter he wrote on Post Position regression analysis. I actually did manual computations by going back on a years worth of Woodbine charts.
I then did a chart with post position on the horizontal line and lengths beaten on the vertical line (I can't remember the math jargon after 27 years).
I found that at a mile and a 1/16th posts 1-4 were pretty much equal, but post 5 was worth minus two lengths in comparison with posts 1-4, post 6 was minus 2 and 1/2 lengths, post 7 was minus 3 points, posts 8-10 were minus 3 and a half points, but then 11 post was minus 5, 12 was minus 5 and a half, and 13 and 14 posts were minus 6.
There was an outside bias back then when it came to sprints. For 6 and 6 and a half furlongs, the one post was minus 2 and post 2 was minus 1.

Anyways, I'm too lazy to try to compute thes anymore. Does anyone do this these days?

[Overlay]

The only posts I pay special attention to are the ones that Quirin found to be positive independent variables (post position 1 in dirt sprints of 6 furlongs or less, and in two-turn dirt routes of 1 to 1-1/8 miles, and also post positions 1-3 in two-turn turf routes, all at tracks where the main dirt oval is at least one mile in circumference.)

[SMOO]

The only posts I pay special attention to are the ones that Quirin found to be positive independent variables (post position 1 in dirt sprints of 6 furlongs or less, and in two-turn dirt routes of 1 to 1-1/8 miles, and also post positions 1-3 in two-turn turf routes, all at tracks where the main dirt oval is at least one mile in circumference.)

Thanks for the info. Any bias for turf sprints?

[Overlay]

For general post position data for turf sprints, I'd have to start from scratch. When Mike Nunamaker (for example) updated Winning at the Races with his 1994 study, Modern Impact Values, he required that he be able to find at least 100 races in a given subcategory of races (broken down by factors such as age, sex, and running surface) before he felt he had enough data on which to base valid conclusions for that subcategory. He used 12,815 races for his entire study, and out of that quantity, he was unable to find the required 100 races for either turf sprints in general, or for turf routes for two-year-olds, in order to publish findings on them. (I also don't recall Quirin specifically addressing turf sprints in Winning at the Races, either in the chapter of the book on post position, or in the one on turf racing (although I don't have my copy in front of me at the moment).) (If he did discuss them, I think I can say with confidence that (unlike dirt sprints, dirt routes, and turf routes) he didn't find any post positions in turf sprints that passed his statistical tests for being independent variables (either positive or negative). (I'll re-check that as soon as I can.) So perhaps turf sprints are more common now, but at that time, it appears they comprised less than 1% of races being run (assuming that Nunamaker's cross-section was representative of races in general). I imagine that other board members who maintain more current and voluminous data bases would have up-to-date information on post position as it relates to turf sprints.

[Overlay]

I've re-checked Winning at the Races. On pages 245-251, none of Quirin's general or track-specific findings with regard to post position in turf races addressed turf sprints. His main 2,270 turf-race post-position sample dealt solely with two-turn routes of 1 to 1-1/8 miles. He did comment that 90% of horses making their turf debuts did so in route races of one mile or longer, rather than sprints. His only statistically significant finding (positive or negative) regarding post position in turf races was the positive one I cited for post positions 1-3 in two-turn turf routes of 1 to 1-1/8 miles, with an impact value of 1.34.