Why Batting Averages Don’t Make Sense in Cricket.

Dear Cricket,

You are all about a ball and bat. It is a hit and run case. There are 11 odd people from each team that chase a ball-very similar to chasing a child who is on a sugar rush. Pretty exhausting right! However, I never knew I would enjoy delving into the statistics of a game that sounded crazy to me.

This week’s viz focuses on batting averages against left arm pacers of key Indian batsman since world cup 2015. So, without any further ado let us understand what these terms really mean.

Batting average — The traditional metric to determine a batsman’s performance over several matches is the batting average. It is simply defined as the number of runs scored for every time he is dismissed. It is calculated as the number of runs scored by a batsman in, a given time, or a series or over a career divided by the number of times he is dismissed over the same time.

Left arm pacers- If there is a bowler who can make the batsman change tack before a ball is bowled, it must be a left-arm fast bowler bowling over the stumps. When a right-hand batsman is facing a left-armer bowling over the wicket, he needs to open his stance a bit, for otherwise the point the ball is delivered from creates a blind spot for the batsman.

What is with these statistics? Why are we interested in analyzing them?

For starters, we are analyzing these statistics to see how Indian cricketers fare against left arm bowlers. And — a bit more on the nuances.

Firstly, let us understand why batting averages are not that great in terms of conveying information. Any measure of central tendency (like mean, median, mode) effectively tries to summarize all the information using one number and irrespective of how this number turns out, some information is lost. It is no different with a batting average.

Instead of using the batting average, what about using the metric, median runs scored per innings? This is basically (batting average*number of dismissals)/ number of dismissals.

To understand this further and see the applications of this metric, let us first dive into the original visualization and see what stood out and what did not stand out in the viz.

What worked for the visualization:

1. The title of the viz is clear and precise conveying important information.

What did not work for the visualization:

1. The images of the cricketers are not particularly useful and does not convey anything useful in the viz.
2. Some context on what batting average and left arm pacing means would be helpful.
3. Bars and color legends would make the chart easier to understand.

My interpretation:

1. From our previous discussion, we realized that batting averages is not a useful measure to convey information. Kohli has the highest batting average amongst those listed but that could also be possible by making some big scores in some matches. He stands as an outlier in the chart.
2. A better measure of central tendency is median number of runs scored per innings that does not get affected by big scores. Based on this metric, we find that Dhawan has the highest score of 12. This implies, that half the times Dhawan goes to bat, he makes at least 12 runs. Kohli, on the other hand scores poorly on this measure. This implies that Dhawan is the more consistent player. A higher ratio in terms of median runs scored may not necessarily imply that Dhawan is a better batsman, but he is surely more consistent.
3. But all this still does not unequivocally state that things are better for Dhawan than Kohli despite a low batting average and high median score for Dhawan. The role of the team features in perspective as well. For an opening batsman like Kohli, time is usually not much of a constraint and the ability to play a long innings has a great impact on the team’s score. So even if Kohli makes 5 runs or less half the time, the fact that he can hit a big score will weigh significantly in his favor. For someone like Dhawan, who is lower down in the batting order and has little opportunity to play a long innings, a consistent 12 might be valued much higher. Or the speed might trump the number of runs he made.

Hence, the batting average might be a great metric for an opening batsman but might result in loss of information for a batsman placed somewhere else in the batting line-up.

Now, that was quite some cricket information! Do you agree with the above?

Sincerely,

New Cricket Aficionado