Moneyball for Cricket

Major League Baseball (MLB) has an inherent disparity between the haves and have-nots since there is no hard and fast salary cap. Therefore, teams in bigger TV markets like the New York Yankees have a lot more money at their disposal to obtain the players they desire that would give them the edge in winning yet another championship over smaller market teams.  Moneyball, if you already didn’t know, was a book written by American author Michael Lewis in 2003, describing a new approach undertaken by the General Manager of the Oakland A’s baseball team (a small market team), Billy Beane, in identifying the value of players and assembling the team that still could be highly competitive with the big boys. To give you an idea of how wide the gap is, the Yankees have a payroll upwards of $200 million and the Oakland A’s are in the bottom ten in terms of 2011 payroll sizes at $66 million.

The Billy Beane approach involved a rigorous statistical analysis based evaluation of player values (Batting average on balls in play (BABIP), on-base plus slugging percentage (OPS) etc.), known as “Sabermetrics” instead of the conventional/traditional means used in baseball, namely, runs batted in (RBI), batting average (BA) etc. Sabermetrics is tool that provides an objective, empirical data based approach that evaluates “in-game activity” as a way to understand a player’s value. Once sabermetrics approach found success in Oakland, many major league teams have adopted it, so much so that the Billy Beane story has been made in to major Hollywood motion picture starring Brad Pitt as Beane.

The Moneyball approach, even at a cursory glance, can be seen as a tool to not only aid smaller market teams compete with bigger market teams with deeper pockets, but also in salary-cap enforced leagues where there may be limitation in terms of resources (players). It may not be suitable per se for country-based sport such as test match cricket or ODI cricket since, theoretically, the resources at that country’s disposal are limitless. However, with the advent of T20 cricket, which has sprouted franchise-based cricket around the world, there certainly is a need for an objective tool to evaluate player values to assist owners to make wise, educated player investments.

Let us take an example of an IPL franchise. Given a choice every franchise would love to have a Sachin Tendulkar and a MS Dhoni and a Virender Sehwag and a Dale Steyn and a Chris Gayle, for marketing as well as cricketing reasons, but they just can’t afford to do it, as it becomes cost-prohibitive. After spending on so many top level players, there will not be much left to sign quality role players to the roster. One may say that the IPL already has a salary cap in place and a situation similar to MLB (of haves and have-nots) will not arise, and therefore there is no need for such an approach to identify less expensive, high impact players. I disagree. It makes more sense to have one or two marquee players but staff the rest of the roster with high impact-low cost players (similar to a Paul Valthaty), as determined by an objective and rigorous statistical tool so that there is high return on investment.

The conventional cricketing evaluation involves the use of batting averages, strike rates (for batsmen) and economy, dot balls, average and strike rate (for bowlers). Are these good enough to identify the high impact players?

The outcome of a T20 match is sometimes determined by what happens in the space of 6-12 balls. That is 2.5-5% of the entire duration (240 balls) of a match. Thus, player analyses need to be broken down to include situational statistics. Is a batsman that averages 40 runs necessarily a better bet than a batsman who averages 30? It may seem like a very easy question to answer but many different factors may have to be considered:

  • Whether the batsman is a top order bat or a middle-lower order bat
  • The strike rate at which he achieves those runs
  • Whether those averages were obtained in pressure situations such as those towards the end of an innings
  • How does the batsman fare while his team was chasing a target
  • Whether he accumulates his runs mostly via singles, two’s or boundary hits
  • Number of dot balls on average the batsman plays out
  • Percentage of runs scored by the batsmen to the team score with respect to percentage of deliveries consumed

These are at least some of the factors that should be considered before even a simple question of comparing the averages of two batsmen can be done.

Similarly, some basic “situational” parameters can be developed for bowlers.

  • Economy rates of the bowler in the powerplay overs, slog overs (last five) and the middle 9 overs
  • Number of dot balls on average in the full quota of 4 overs
  • Strike rates in the powerplay overs, slog overs (last five) and the middle 9 overs
  • Averages against top order and middle order and lower order batsmen
  • Strike rates against top order and middle order and lower order batsmen

A player’s value should also incorporate their fielding ability, skill and dexterity to field within the ring to take sharp catches and effect swift runouts, as well as range in the outfield. Fielder efficiency is an area that has never received much attention even in a stat-obsessed sport like cricket. (Peter Della Penna raised this question, as did Rob Steen). A fielder may take plenty of catches but if he drops equal number of them, his effectiveness is severely diminished. Similarly, the ability to prevent boundary shots or limit the number of runs from a shot seemingly destined for a four are priceless abilities to have in T20 format where the margins of victories are ever so slim. A player with outstanding athletic ability and an accurate arm is worth his weight in gold in the T20 format. I had written about a possible way to evaluate fielder’s run stopping efficiency in my old blog. Here is the gist of it:

A GPS chip could be installed in a player’s accessory such as hat, shoes etc. (Gadgets like Run Keeper have been available for a while that use GPS to keep track of how much you have run). Wagon wheel software already exists in Cricket that tracks the scoring shots of batsmen. The idea (just as an example) is to incorporate that in to 10 m diameter circle around any particular fielder. The 5 m on either side of the fielder is his range. The efficiency of the fielder to stop anything within 5 m of his fielding position will be quantified. So, you have a computer screen on the analyst screen in the production room. It will display a 5 m radius circle around each of these fielders. This way, it will be easier to see the number of runs a particular fielder saved during the course of a match. The 5 m range will be graded in to 1m wide regions (basically, 5 concentric circles). Each circle away from the fielder would indicate a level of difficulty such as from 0-1 m (easy) and 4-5 m (difficult). So, a composite number (+/-) could be derived based on the different levels of difficulty fielding effort produced by the fielder.

We are seeing several top players cut their international career short (retiring from tests) to prolong their moneymaking ability in the more lucrative T20 format and there is nothing wrong with it, morally. We may soon be heading towards an era of specialized T20 players. We are already seeing the increase in the number of players who are basically T20 mercenaries such as Kieron Pollard and Chris Gayle amongst others. It is then prudent for franchises to move towards an objective player evaluation tool.

(Thanks to Devanshu Mehta for his feedback on the post. Do read his take on bringing Moneyball to IPL)

This entry was posted in IPL. Bookmark the permalink.