when baloney gains credibility

[Tim Y]

http://en.wikipedia.org/wiki/Misuse_of_statistics

http://www.ibiblio.org/rdu/a-liestt.html

http://everything2.com/title/How%252...2520statistics

"A pleasantly subversive little book, Guaranteed to undermine your faith in the almighty statistic."—Atlantic

The last is a great primer to exemplify how it works. This text from 1954, How to Lie with Statistics" (by Darrell Huff) was a suggested reading when we were learning about experimental design and the statistical analysis that zoology studies so often create. A cancer study was the prime example done in Sweden. Cancer clusters were correlated to living in proximity to power lines. Only problem to the study was that the researchers never looked at other populations exposed to the same environment that did NOT get cancer of any kind so the study was exposed as “framed specifically to draw a pre-determined conclusion and had no adequate control.”

Most of the “stats” we see in sports are of the raw variety never having been subjected to the analysis, framing and repeated scrutiny that a scientific study would require before it’s finding would be accepted by the general scientific community.

When our best quarterback won the Heisman Trophy in 1968 (Gary Beban), it was said that HIS only deficiency was HIS pass completion average. Those who had watched him all year knew that he was saddled with some of the most MEDIOCRE receivers in the league who dropped ball after ball put right in their hands, YET that was HIS statistic and not one where the causation could be attributed.

Mike Pizzolla reported a cohort at the track who slavishly utilized trainer statistics and went on and one about how the trainer was so many percentage points when a horse was dropping and that his filly winning percentage was so and so, TOTALLY disregarding the out of form, totally in decline filly that was right in front of him today’s race. How can the secondary statistics of a trainer have ANYTHING to do with the INDIVIDUAL right there is front of you? ….was his conclusion. Wake up! it is the horse not the connections that is there in the race.

Study the “flash in the pan” trainers like Cam Gambolotti or Louie Rousel. Were their stats reflective of their competence the year they both had GOOD Triple Crown horses, or were they skewed by the relative colts they had in their barns Spend A Buck and Risen Star? Would old Grover Buddy Delp’s graded stakes statistics be better or worse in 1979-80 as compared with 1984-85? The good horse he had in the barn, THE INDIVIDUAL, made his association with him look FAR better than his talents projected.

Attributing cause and effect where that connection is both spurious and diluted is not logical. Imagine a trainer with two excellent mudders in a barn of 30 horses. A stat on the TRAINER’S ability with mudders would be diluted by the total number of wins of ALL of those horses and thus LOST or averaged out as to be of no use at all while LOOKING at the individuals, the horse’s right in front of you, would locate the author of that mudder ability and not just who managed them.

The individuals in front of you, the ones that are in or out of form, those that match up to today’s pace structure, the ones that are there competing today are the ONLY ones to look at with any degree of confidence. PERIOD