The Math’s of Moneyball?

Discussion in 'Science & Society' started by ScaryMonster, Apr 26, 2012.

  1. ScaryMonster I’m the whispered word. Valued Senior Member

    Messages:
    1,074
    Moneyball is a film starring Brad Pitt and Philip Seymour Hoffman and Jonah HIll. It’s based on the true story of Oakland A’s Baseball team manager, Billy Beane (Pitt), who along with Ivy League Economics graduate and Uber Maths Nerd Peter Brand (Hill) used player stats to save the Oakland A’s baseball team in 2002.

    Beane had to deal with enormous budget restraints ( Oakland A’s budget was $4o million compared with the Yankee’s $125 million), he showed that the statistics were more effective than experience ie. the stats beat the club selector’s know-how.

    [​IMG]

    This equation is called the Pythagorean Expectation.

    So what’s going on? This equation, doesn’t give you that much information. We will fix ‘runs allowed’. OK. If ‘runs scored’ is high Win is high; it ‘runs scored’ is low Win is low. Um, this means scoring runs is good. Yeah! I knew that without the maths.

    The 2002 A’s scored a total of 800 runs, and allowed a total of 654 runs, for a Pythagorean Expectation of:

    [​IMG]
    This compares to the team’s actual win percentage of 103/162, which is around 0.636.
     
    Last edited: Apr 26, 2012
  2. ScaryMonster I’m the whispered word. Valued Senior Member

    Messages:
    1,074
    Peter Brand applies this formula in order to estimate the number of runs the team needs to score, along with the maximum number of runs it can allow, in order to secure a playoff spot. In one scene, he tells Billy Beane that he thinks the A’s will need to win at least 99 games to guarantee a playoff spot. In a 162 game season, this equates to a win percentage of around 0.611. In order to ensure that the Pythagorean Expectation is at least this large, we set:

    [​IMG]

    With a bit of Algebra, it’s the same as:

    [​IMG]

    In order for this to happen, the team needs to score at least 814 runs, and can allow no more than 645 runs. This gives a runs allowed to runs scored ratio of 645/814, or around 0.793 < 0.798 (though, if I were being anal, I would point out that with 814 runs scored, the team could allow as many as 649 runs and still have a runs scored to runs allowed to runs scored ratio that is less than 0.798).

    The thinking behind it was there and accurate, but like some people in the movie said, the 2002 A’s weren’t fundamentally sound as far as skills go. I wouldn’t trust a 25 year old from Yale with an economic degree. I would go with the conventional wisdom of, a good fundamentally sound player, should be on my team as opposed to someone who i can get for cheap.

    But still the Red Sox won in 2004, putting to rest the so called “Curse of the Bambino,” supposedly using this method.
     
    Last edited: Apr 27, 2012
  3. Cavalier Knight of the Opinion Registered Senior Member

    Messages:
    157
    I do not understand the opposition to sabermetrics (and it mas mostly fled the field...in baseball). These are big money games that rely on the genrally consistent performance of athletes, and so eminently understandable mathematically. In fact, baseball has always had its statistics, it just had the wrong (ad in, unpredictive) ones.

    I'm waiting for the trend to take over football. It was shown a while ago that it's always mathematically worth it to go for it on fourth down, rather than trying for a field goal, but coaches basically thumbed their noses at that. The game is waiting for a mathematically inclined coach to come and clean house.

    IMO, there is too much money on the line to leave these games to the superstitions of non-technical minds.
     

Share This Page