---

On a number of occasions at this Football Think Tank Web site, I have expressed my admiration for the Bill James/Moneyball approach to baseball analysis. The 10/16/06 Forbes magazine reports that there is a Web site called Protrade.com that applies the Moneyball approach to football and basketball as well as baseball and extends that way of analyzing sports. Hear! Hear!

Moneyball is a fabulous book by Michael Lewis about the Oakland Athletics baseball team’s low payroll-per-win cost and how they achieved it. I am currently reading Lewis’ latest book, Blind Side, about how the value of offensive left tackles skyrocketed after free agency came to football because of belated recognition of their crucial role in protecting right-handed quarterbacks.

Primarily, Protrade is a sort of stock market where Web visitors can buy and sell shares of various pro players in football, basketball, and baseball. That idea does not much interest me. It sounds like all-star team balloting. Why should anyone care what a bunch of ignoramuses think a player is worth?

But Protrade’s “About us” Web site also says they have “assembled a unique team of MIT statisticians, economists, and leading sports figures to develop a patent-pending valuation engine to quantify on-field performance and assess an athlete's contribution to winning.”

I am not sure why only MIT statisticians are involved. One of the founders graduated from UCLA; the other from MIT.

And I would expect economists could comment about who is over- or under-priced in terms of pay versus performance.

As far as “leading sports figures” is concerned, I kind of thought they were the problem when it came to intelligent analysis of sports strategies and tactics. They are certainly the enemy in Moneyball. They are what’s known in the technology business as “legacy problem.” That is, they have a vested interest in doing things the bad old ways.

According to Forbes, Protrade has a massive database of five years of football plays (all NFL I assume). That’s 500,000 plays. Great! That’s the starting point for intelligent analysis.

Forbes also says they have algorithms that calculate scoring and winning probabilities for players and teams. Great again! I put one of Pete Palmer’s win probability tables inside the front cover of my Football Clock Management book. Fundamentally, win probability is the basis for all football decisions. That is, you always pick the decision that has the highest win probability. What this Football Think Tank Web site is about is highly-paid coaches at the college and pro levels who, either out of unacceptable ignorance or deliberately, choose the decision that has less than the highest win probability available at the moment.

Protrade says they use binary logistic regression equations and Markov models to calculate scoring and win probabilities and to value players. Protrade has licensed this stuff to ESPN. Interesting. They never showed any interest in either Pete Palmer’s work or mine which are somewhat similar to what Protrade is doing. I brought both Palmer’s and my work to the attention of ESPN and other sports broadcasting organizations nine years ago in 1997.

Forbes says the CNNSI.com Web site also is now using Protrade analysis.

I am not sure I understand all Protrade is doing or want to. But bits and pieces of it sound extremely healthy for my purpose of getting football coaches and owners to move out of the dark ages into the enlightened era when it comes to decision making in strategies and tactics.

To contribute an idea or comment to this Football Think Tank web site, either email to johnreed@johntreed.com or fax to 925-820-1259 or snail mail at 342 Bryan Drive, Alamo, CA 94507.

Best wishes,

John T. Reed

Order Form | Coaching home page | Secure shopping cart | Football Think Tank