Batter strategies and their impacts in the IPL

Spend time or hit out? How have different strategies affected teams’ winning chances?
Data science
Cricket

Introduction

The role of batters in cricket is to make as many runs as possible for their team, and sometimes, especially in test cricket, to also play as many balls as possible. In T20 cricket, due to the short format, batters can take more risks than in the other longer formats, and hence an increasing emphasis is on scoring as many runs as possible in as little balls as possible. Of course, this doesn’t automatically imply that batters always take risks as it can mean that the team runs out of specialized batters with many overs still left to play. Batters take more risks than in ODIs or tests, thereby allowing quicker scoring, while also spacing out these risks appropriately, to be able to playing out the allotted 20 overs. T20 cricket has also led to very specialized roles for batters and a lot of emphasis is given to ensuring that the best hitters get enough balls to showcase their skills.

In the last 14 years since IPL began in 2008, batters in the debatably world’s premier T20 tournament have employed various strategies to make these quick runs. Depending on whether the batter is an opener, a middle-order batter or a finisher, different teams and batters have developed diverse templates. In this project, I investigated individual batter templates and strategies, and their impact on the team’s winning consequences.

The trade-off between strike-rate and risk

The trade-off between risk and strike-rate takes the center stage in individual batter strategies. A batter willing to take more risk has access to a wider range of shots for each ball, often diverging employing shots that diverge from the prescribed shots in the coaching manual. Some batters like Andre Russell score quick runs while taking more risks. This increases their chance of getting out each ball, and thereby they end up playing fewer balls. Such finishers usually come later in the innings when there aren’t a lot of balls left. Some teams like the Kolkata Knight Riders have often sent out “pinch hitters” like Sunil Narine earlier in the innings, often as an opener. On the other hand, batters like AB de Villiers and Jos Buttler switch gears between low and moderate risk. But even at the highest risk, the emphasis is slightly more on surviving to play another ball than is for batters like Andre Russell. More conventional batters like Mike Hussey, Tendulkar and Kallis rely heavily on a risk-minimal approach unless the situation strongly demands otherwise.

Strike-rates vs balls faced Strike-rates vs balls faced

The figure above illustrates some aspects of this trade-off. Keeping tail-enders aside, notice that batters are spread on the diagonal from the top left to the bottom right. Classical players like Mike Hussey, Kallis, Tendulkar and Gaikwad are in the top left quarter - they play more balls than most others and hit at a lower strike-rate than most others. As we move through the large group of unnamed batters marked in blue, we find another group of players like Maxwell, Pollard and Hardik Pandya who play fewer balls than most others and hit at a higher rate than most others. Some notable exceptions are batters who seem to escape this trade-off, or perhaps push it to its limits - KL Rahul, Jos Buttler, AB de Villiers, Warner, Gayle, Russell and the like.

Openers play more balls than the other batter types while also exhibiting a wide range of strike-rates. For example, Gaikwad’s average strike-rate is just short of 100 and Jos Buttler’s is around 135. Gaikwad, however, typically plays a few more balls than Jos Buttler. How do these different strategies impact the match outcome?

Strategy differences between winning and losing

Separating the innings played by each batter into two groups based on whether their team won or lost the match, we can measure how different the batters strategies are in the two groups.

Instead of measuring the difference between mean strike-rate in wins and mean strike-rate in losses for each batter, we use percentile scores. The mean percentile score from strike-rates in wins (losses) tells us overall in all innings, how often the batter scores slower than they typically score at in wins (losses). The difference between these two values (wins minus losses) quantifies the typical increase (if positive) or decrease (if negative) of strike-rates from losses to wins. A similar logic applies for balls faced - the difference between mean percentile scores in wins and losses quantifies the typical increase (if positive) or decrease (if negative) in the number of balls faced from losses to wins.

Differences in strike-rates and balls faced between wins and losses Differences in strike-rates and balls faced between wins and losses

How is the calculation done? Percentile scores transform data into values between 0 and 1, making comparisons across datasets with different ranges easier. The percentile score of a particular value is the fraction of data that is smaller than the value. We do this for each batter separately for both strike-rates and balls faced. The percentile score of a particular strike-rate (in some particular innings) of a batter tells us how often the batter has scored slower than in that innings. So for the batter’s IPL career-best strike-rate, the percentile score would be 1 (signifying that the batter plays slower in all other innings), and for the batter’s IPL career-worst strike rate, the percentile score would be 0.

Interactive Plots

Here is an interactive version that presents these features of strike-rates, balls faced and their differences between wins and losses. You can select a subset of batting positions on the right to limit display to those positions. Hovering over a point would reveal the batter information, along with their average performance in wins and losses. When you hover, two small plots pop up on the right edge that reveal how strike-rates and balls faced differ between wins and losses in all individual innings of the batter. After hovering, you can click to label the batter even when you move the mouse away from the point.

Below is an interactive plot of how different batters did (strike-rates and balls faced) in their IPL innings when their team won and lost, compared to the overall distribution of how batters do in general.