by Dr J, Volume 21 - Issue 4 - Summer 2003-04 Sport Health
In these days of the Internet, pay TV and an excellent standard of newspaper journalism, I read far fewer books than I used to. Recently I bought a book that I couldn?t put down, reading it from cover to cover in a day. It was called Moneyball, written by Michael Lewis, and its subject of baseball may make it less of an essential read for those Australians who are much more accustomed to cricket. I happen to like baseball but the main reason why I loved the story in Moneyball is that I am a self-confessed stats geek. The reality of professional sport is that the people who control it ? whether they are players, coaches or managers ? practice far more as artists than scientists. Moneyball suggests that this balance may slowly begin to move further towards science.
The protagonist in Moneyball is Billy Beane, who is the general manager of the Oakland Athletics. Beane is not a scientist. He is an administrator who wished that he had a university degree, except that in his youth he was far more preoccupied with trying to become a top professional baseballer. However, Billy Beane is a disciple of science, in particular statistics. The coaches and scouts who ?found? Billy Beane when he was a young player predicted he would become a super-elite batter, based on his obvious characteristics of blistering speed, a strong arm, the strength to hit the ball out of the park, and perhaps also his good looks and charisma.
However, Beane felt that he was a fraud as he never reached the expectations of the scouts who had the gut feel he could make it all the way to the Hall of Fame. No matter who how good he looked and how much his batting style promised to come good, statistically Beane couldn?t get to first base at a rate any more than a fringe Major League player. Hence that is all he ever was as a player ? a fringe Major Leaguer. He quit at the height of his career as a fringe player, after being traded from team to team, to join team management. As general manager of the Oakland As he holds the equivalent position of CEO in an Australian sporting team, yet at the As he wields more power than the manager (coach) as he insists that the team base its philosophy on what is statistically sound rather than gut feel.
Moneyball is subtitled The Art of Winning and Unfair Game, and it is worth knowing that MLB is an unfair competition from the salary cap viewpoint. Basically there is not an effective salary cap and hence some teams spend over three times the amount on player salaries than other teams. The As have been the second lowest spending team over the past five years, basically because they are only the second most prominent team (to the San Francisco Giants) in a smallish market and their owners insist on running the franchise at a profit. In general in MLB, there is a high correlation between amount spent on player salaries and team performance, but the As provide an amazing exception. They have made the playoffs ever year for the last five and have built one of the highest winning percentages in MLB over that time.
Baseball is a sport that attracts stats geeks like myself, as its participants are more obsessed with statistics than any other sport on the planet. Where the Oakland As differ from other teams is that they are obsessed with applied statistics rather than just descriptive statistics. Billy Beane wants to know which statistics correlate with winning games rather than which ones look good.
The prime example is the difference between batting average and on-base percentage. The traditional worth of a batter is batting average, which is a percentage measure of how often the batter can successfully reach first base after safely hitting the ball. On-base percentage is measure of both successful hits and also getting to first base on a ?walk?. The As have done regression analyses (correlating winning games with various statistics) and found that -- surprise surprise -- it doesn?t really matter how you get to first base, as long as you get there.
Traditionally it was thought that getting to first on a walk was a soft option, with the superior batters always getting to first on a hit. The As found that there was actually a trend in the other direction, that on-base percentage had a higher correlation with winning than batting average, perhaps because players who drew a ?walk?, in doing so, tired out the arm of the opposing pitcher. From various associated statistical runs, Beane and his team of applied statisticians worked out that the most important, and undervalued, characteristic of a batter was the ability to not swing at pitches outside the strike zone. Ironically this was the one ability that failed Beane himself as a player, and he as a manager would never have signed himself as a player because of this flaw, and despite his speed and power.
Where the book (and philosophy) is controversial is in Beane?s management of characteristics he considers are overvalued, particularly that of running speed in batters. Obviously, all other factors being equal, it is beneficial to be fast, both in terms of getting around the bases and taking catches in the field. However, the regression analyses that the As have performed show that speed correlates only minimally with actually winning games, yet it massively pushes up the value of a player. On-base percentage obviously has a built-in component which takes speed into account, as a faster player will more often get to first base. What the Oakland As believe is that too often, speed can camouflage the far more important characteristic of whether a player swings at pitches he should be leaving.
Billy Beane would love to have the budget to purchase fast players as well as careful swingers, but with a limited budget, he always trades for the latter. In fact, he believes that speed is so overvalued in baseball (that is, it almost never decides the result of a game) that he generally won?t let his coaches instruct players to steal bases.
The bottom-line with the Oakland As is unfortunately that they have been living in the statistician?s version of hell in that, despite their enormous win-loss percentage over the past five seasons, they haven?t managed to win a World Series. They have watched the Florida Marlins, who have statistically been a far inferior team to the As in terms of win-loss percentage, win two recent World Series because they have peaked at the right time.
AFL fans in Adelaide, by comparison, would care little for the fact that the Port Adelaide Power have a better historical win-loss percentage than the Adelaide Crows ? it is all about the two Premierships that the Crows have won. Bare statistics tell us that Port Adelaide are an outstanding AFL team, that the All Blacks similarly excel at rugby and that Greg Norman was a great golfer, whereas the popular view is that all of the above are chokers.
It is important to remember that, despite the pain of losing in games of critical importance, teams who regularly win to give themselves opportunity in the big games enjoy more success than the chronic losers who never get to play in a big game. The nature of sport is that prizes are generally awarded to the winner of a one-off game or small series, which is statistically often not going to be the best team of the year.
There are similar opportunities to statistically analyse just about any sport you can care to name and Moneyball gives me inspiration that science will have more of a part to play in all sports in the future. In rugby union, for example, I am sure that statistically, goalkicking success percentage would correlate highly with game winning, because penalties are so common and penalty goals are worth three points.
Although signing up Wendall Sailor from league was a marketing coup, I can?t believe that he would be worth more to a union side than fellow winger Hazem El-Masri, who is such a sharpshooter that he occasionally wins rugby league matches from his own boot. In union he would do it far more often, just as Sailor is statistically less likely to score the matchwinning try in a union game than a league match.
With regard to goalkicking in rugby, I am still waiting to see someone statistically prove the obvious that left-footed goalkickers are more accurate than right-footers in the right-hand side of the field and vice versa, which should mean that most teams select both left-footed and right-footed goalkickers in their teams.
In the AFL, there is an obsession with forwards kicking goals and hence many of the most popular players over the years have been the highest scoring spearheads. However, the two most famous of the modern era, Tony Lockett and Gary Ablett, don?t have a Premiership medallion between them. In 1993, Ablett kicked ten or more goals in a match on three occasions where Geelong actually lost. Another less famous, but equally charismatic full-forward of recent years, Matthew Richardson, has been by far Richmond?s top goalkicker of the last ten years, yet he has been out injured for the season in the Tigers? only visits to the Preliminary Finals in this time.
Statistically, the most important characteristic of a forward line is the ability to convert inside-50s to goals, rather than how many goals the full-forward kicks. Alistair Lynch is a less revered full-forward than any of the three previously mentioned, yet he has three Premiership medallions, and perhaps some of the secret is his ability consistently to get the ball to fall at his feet in front when he contests a mark, giving the crumbers in his side the ability to score from his contests.
In golf there is the saying ?Drive for show and putt for dough?. There is a short-game coach called Dave Pelz who also writes excellent books and, based on statistics he believes that golfers should have at least four wedges in their bag, since wedges lower scoring far more than long irons. He has determined, using similar methods to the Oakland As, that clubs like three and four irons are over-rated, as not even the world?s best players can get consistently down in two from a long-iron. Scoring low in golf is all about your percentage in getting down in two whenever you are within 60m of the hole, which means that the wedges are the most important clubs.
In the NRL, almost everyone rates teams based on how many State-of-Origin players they have on their roster, which should mean that the Broncos would win every year. Tim Sheens has said that it is better for teams to have the second best player in each State on their roster, as the Origin players are so often fatigued and beaten up by the time the NRL finals come around.
Even amongst Origin players, they don?t equally contribute to winning percentage of their teams. Props are usually the biggest, toughest players but the top props probably manage 7m per hit up against 5m per hit up made by the so-called terrible props.
Winning is probably far less related to metres gained and far more related to quick play-the-balls, which give the backs more chance to create line breaks, which in turn are far more related to winning matches. Yes, it is better to gain more metres on your hit-ups, but most teams gain more metres on one kick-chase at the end of the tackle count than during five hit-ups. This means that, if you have a limited salary cap in the NRL, to win you want preferably to spend it on good tactical field-kickers than on props.
If you want even more books for your Christmas reading list, and hope that science will one day show that medical staff can contribute to teams winning matches, it is worth reading either of the two books published by retired professional NFL doctors in the USA. Rob Huizenga?s book is the better read from the entertainment viewpoint, whereas Pierce Scranton?s is more of a guide on how actually to practice sports medicine at the coalface. Which one you read all depends on whether you are about entertainment or winning.