In order to build a winning team, one first has to understand the factors/events that generate victories. This is where a thorough analysis of data is required.
The story we relate below clearly shows that out-of-the-box thinking is required to determine these win-producing events. Traditionally used statistics may not really reveal the true picture and, in fact, may cause wrong decisions to be made.
In his book Moneyball, journalist Michael Lewis details the way the Oakland Athletics, a Major League Baseball (MLB) team based in Oakland, a suburb of San Francisco, rethought the game of baseball by turning traditionally held myths on their head and recruiting players who were not sought after by other teams (and hence not very expensive), but possessed skills that allowed the team to generate more wins.
Also Read Previous articles in the series
Lewis’ book is about the general manager Billy Beane, and how in the late 1990s Beane made rather unconventional (and daring) hiring decisions by relying on the analysis of a breed of baseball statisticians/analysts who called themselves “Sabermetricians.” Sabermetricians believed in the analysis of baseball through objective evidence. They did this by rigorously analysing baseball data without allowing conventional wisdom to bias their analysis.
The analysis by sabermetricians showed that traditionally used statistics such as batting average (same as that in cricket), stolen bases (equivalent to byes in cricket), and runs batted in (equivalent to partnership in cricket), are not good predictors of the run scoring ability of a batsman, and by extension the team. However, analysis of data showed that runs scored were explained more accurately by on-base percentage (roughly equivalent to rotation of strike in cricket), and slugging percentage (equivalent to the strike rate).
All other teams in the late 1990s were using the traditional measures to evaluate the run production efficiency of players, and using them to make recruiting decisions. The highest valued players were the ones who had high batting averages and runs batted in. There was no competition for players who fared well when evaluated based on sabermetric measures such as on-base and slugging percentage.
Billy Beane was, therefore, able to assemble a winning combination at a very low cost. In 2006, the Oakland Athletics were 24th out of 30 MLB teams in terms of salaries and fifth in terms of their winning record.
The Twenty20 cricket game is probably the closest to a baseball game in terms of the length of the game—both in terms of time and the number of pitches. Also, run production is a very important consideration in this short version of the game. Therefore, analysis similar to that pioneered by the sabermetricians in baseball will shed some light on factors that are good predictors of run production and will help teams identify players who are capable of high-run production, based on these metrics. Such analysis will also allow the coaches to rethink the game and devise strategies that are appropriate to a 20-over game.
As in cricket, there are two other facets of the game—pitching and fielding—in baseball where analysis is now being used to identify key plays to evaluate players.
Team before self
While what we have talked about so far relates to revenue, costs, and value maximization, there are a few other interesting angles for a researcher, coach and the team management. This has to do with the behaviour of players, i.e., what skills they choose to exhibit, and whether they play to maximize their value or whether they play to generate a win for the team.
Again, we refer back to American sports and to Michael Lewis. In a recent article (“The No-Stats All Star,” 15 February 2009, The New York Times) Lewis highlighted the use of analysis in professional basketball. He points out that the front office of the Houston Rockets, a successful team in the National Basketball Association (NBA), using analysis similar to that used by sabermetricians, has found that “box scores” that are widely reported and used are not relevant metrics to evaluate players.
Lewis gives the example of a Houston Rockets player, Shane Battier. Battier was traded by many teams before he ended up at Houston. Even at Houston, his statistics, as highlighted by the box scores, are dismal—he hardly ever scores, and does not have many rebounds, steals, blocked shots or assists. However, the Houston general manager Daryl Morey found that teams won when Shane Battier played.
Coming from the Boston Redsox, an MLB team that is now well known for the analytically minded front office, Morey was tasked with rethinking basketball by the owner of the Houston Rockets.
Battier, Morey says, is an unselfish player. This is very rare in the world of basketball where most players want to show their wares and look as attractive as possible. Battier’s contribution to the team is not revealed in the box-scores. Even though limited in talent, Battier knows where exactly to be on the court at any time and he makes it difficult for the opponents to score. Unfortunately, the tactics he employs and his strategies to make it difficult for his opponents to score are not captured by the conventional metrics.
The Houston front office, in the process of rethinking the game, has now collected data on all the small things players do that help teams win. This has resulted in new strategies, and new ways of evaluating players performance and impact on the game.
Cricket and the role of incentives
This is extremely interesting from the point of view of cricket, especially the Twenty20 brand of cricket. Given the short length of the game, strategy matters. Also players do not have much time to grasp the situation and come up with appropriate strategy. Given this, it is very easy for players to adopt strategies that may not be in the team’s best interest.
For instance, players may resort to shots that are risky, but yield a boundary if they work. The question is whether the player is maximizing his self interest or the interest of the team in these circumstances.
Players respond to incentives, and if the incentives are not correctly set, the response is suboptimal as well. If, for instance, a player’s performance is measured by the runs scored and his strike rate, it could result in the player pacing his innings in a way that achieves these dual goals. But this might not be in the interest of the team.
For instance, a player could take 70 balls to score the 50 runs, and then accelerate to score the next 50 runs in 30 balls, thus achieving the twin goals—ensuring at least a 50 while making his strike rate 100.
In Twenty20 cricket, use of analysis can pin-point exactly what are the critical plays and strategies that contribute most to a win. This can then be used as the metrics by which players would be evaluated.
For instance, Shane Warne the captain of Rajasthan Royals is said to have designated a role for each player and drilled it into them. The players are not required to react to the situation based on their reading of the situation, but rather perform their role when called upon by the captain. In this sense, the captain reads the situation and decides what roles would fit the situation.
Players are evaluated on how they perform in their designated roles.
The introduction of the Twenty20 format, coupled with the franchise model, has added a new dimension to cricket. The new format has attracted attention the world over and it could jostle for space with other professional sports in the American and the European markets.
Given the similarity of cricket to baseball, we could predict that sabermetrics-type analysis will become popular in cricket and sound economic analysis, backed by statistical analysis, will drive the decision making at for-profit franchises.
The author is director at Nathan Economic Consulting India Pvt. Ltd.
The series is concluded. Respond to this column at firstname.lastname@example.org