Why you can't buy the IPL title

It is no secret that there is big money in the Indian Premier League. The BCCI estimated that the IPL contributed $182 million to the Indian economy. Ben Stokes was earning $280k per week for his play with Rising Pune Supergiant, which compares favourably with the top salaries in the Premier League

But some teams are worth more than others. Can the most valuable franchises leverage their superior financial heft to increase their chances of winning the ultimate glory - the IPL title?

The most recent champions, the Mumbai Indians, are currently valued as one of the top two franchises and I recently tweeted a chart which showed a relatively compelling relationship between total salaries and success in the 2017 IPL season

Both teams who contested the final had high overall team salaries. Rohit Sharma was keen to celebrate Mumbai's strength in depth whilst Pune benefited from some sensational performances from their aforementioned overseas star player throughout the league stage

Salary comparison of IPL teams 2017 including highest paid player

But the outlier in the chart tells a different story. The Royal Challengers Bangalore had by far the highest wage bill in the league and some of the highest paid stars in the league in Kohli, Gayle, AB, and Shane Watson (they also paid Tymal Mills an astronomical amount for all the good that did them)

Are the 2017 Royal Challengers Bangalore a common occurrence and a tale of caution not to overspend on a team of superstars, or are they the outlier in a league where money talks and high salaries are necessary to build a strong team?


Making sense of IPL salary data

The first issue when trying to answer this question is data. I wanted to expand my analysis to include all IPL seasons but salary information for the early seasons is hard to acquire

This is mainly because the IPL has not historically had a coherent set of league rules, so much as the teams make up a completely new set of rules each year that attempt to cope with any scandals or financial problems that they have at the time. A salary cap exists but there is no consistent mechanism which determines how it might change year-to-year and most teams don't adhere to it anyway

The sale prices for players at auction are public information and so the salaries are easy to work out. But this is the exception rather than the rule as the salaries of retained players and players signed outside the auctions did not need to be publicly disclosed by teams in the early seasons of the IPL

Distribution of player salaries for IPL teams in 2017. Rising Pune Supergiant have a high concentration of salary in heir top players. Delhi Daredevils have a more even spread; strength in depth

This made analysis difficult and I was forced to make assumptions. I limited the analysis to the top 15 players in each squad. This covers the vast majority of team salaries anyway. When teams did not have 15 players with publicly available salaries I assumed that all other players were paid $50k (the Mumbai Indians had just five in 2008)

This last assumption does limit us from being able to make any conclusions about whether strength in depth is important to achieve success in the IPL. However, it is unlikely that strength in depth is an easily purchased commodity. Rahul Tripathi earned just $15,000 (10 lakh) for his season and yet produced some stunning batting performances. My suspicion is that the many young domestic stars of the IPL, often playig for their loacl team, make correlating depth with salaries very difficult


The link between salary and wins

The next step was to trawl through the data looking for a decent correlation between player salaries and team success. The strongest to be found was between the amount spent on a teams' top three players (versus league average) and Net Run Rate

Weak correlation between the salaries of the top 3 players in a team and the overall team success. Data includes all 10 IPL seasons

There are a hundred reasons not to trust this analysis. The correlation between salary and success is obviously very weak. This was just one of a multitude of correlations that could have been picked, increasing the chances of a false positive. The sample size for each season is just 14 games and T20 is a highly unpredictable format

Chart showing how to identify the highest correlation and a false positive

However, it is unsurprising that these variables in particular turned out to be (weakly) correlated. It is common in sports that differential statistics (such as Net Run Rate) tend to be more predictive than points or winning percentages. And there is clearly a relationship between the amount a player is paid and their ability to win matches for the team. Perhaps with reliable salary data and a greater number of matches played each season, this correlation could have been much stronger


A cautionary tale

We only have to look to the Royal Challengers Bangalore to see how the fortunes of a team can change in the IPL, regardless of the salaries being paid...

With the highest paid team in the league in 2016, RCB rode their superstars to the IPL final. Their top four batsmen all had the ability to destroy bowling attacks and make even the largest run totals look trivial. When one superstar failed there were always three others to fall back on

This season they were the statistical anomaly, the outlier on all my charts. The key characters are almost unchanged from 2016: Kohli, Ab de Villiers, Chris Gayle, and Shane Watson were all back for another try. But the outcome was different. They lost 10 of their 14 matches and posted embarrassingly low scores

Results in the IPL are fickle. You can buy the best players but you can't guarantee that they will succeed