Register today for our Generative AI Foundations course. Use code GenAI99 for a discount price of $99!
Skip to content

March Madness

run by Patrick Fisher (brother of statistics.com instructor William Fisher). Students from statistics.com’s Rasch measurement courses will recognize the methodology.

Statistics has always had a vital role in sports, but it has traditionally been of the scorekeeping nature. The variety of attributes and events that can be quantified has expanded, and so has the amateur armchair analysis that picks out interesting tidbits from among the morass of data. Serious quantitative analysis of the same has been around for a couple of decades, but only recently has it been incorporated into player selection, compensation, strategy and tactics in a big way.

I once spoke to a former environmental scientist who was convinced that the betting world in Las Vegas was setting odds based on imperfect information. He felt sure that more rigorous analysis, incorporating more variables, would yield more accurate odds, particularly in horse racing and baseball. He came up with the models, and started to apply them. However, the edge he gained, while definite, was minimal. To truly make money with the improvement in odds accuracy, large sums would have to be wagered, sums that did not come readily to an environmental scientist. He hooked up with the moneyed interests in Las Vegas, and shared the profits with them (and probably slept more fitfully). After some years, the rest of the business caught up with him and what once were crude odds became more accurate and harder to beat. So, he retired and wrote a magazine article about his experiences.