What are the most common scores in Cricket?

Kane Williamson in the 2019 Cricket World Cup Mark Lockett, CC BY-SA 2.0 via Wikimedia Commons

At the start of this year, Kane Williamson scored 238 in New Zealand’s first innings of their second Test match against Pakistan in Christchurch. Remarkably, this was the first time this score was ever achieved by anyone in international cricket. Telling my dad about this statistic made us consider what the distribution of all scores in international cricket would look like.

Using the freely available data at Cricsheet, I set to work mapping out data on international scores. The data is from International Men’s and Women’s matches going back to 2005, across all three formats (Test matches, One Day Internationals and Twenty20 Internationals).

What is the most common score?

As expected, the most common score in international cricket is 0 runs (otherwise known as a duck), with 11.3% of batsmen in the data used getting a duck. Following this, the distribution seems to be as expected: the higher the score, the less common it is.

Nervous Nineties?

In cricket, however, there are milestones particularly coveted by batsmen, which are 50 runs and 100 runs. Do batsmen concentrate harder as they close in on these landmarks, improving their performance, and hence are they less likely to get out in the 90s?

You can see from this graph, which puts scores into groups of 10, that each set of scores is less common than the previous 10, as you would expect. However the percentage of scores between 90-99 (0.80%) is the same as the percentage of scores between 100-109 (0.80%). 90-99 are often referred to as the “Nervous Nineties”, and whilst I can’t accurately decipher nervousness from this data, it does seem that players concentrate more as they approach 100. The percentage of batsmen dismissed as they approach 100 runs seems to be lower than it might be if the 100 mark was meaningless.

How do good batsmen compare to bad ones?

I also wanted to consider distributions to see if they looked particularly different for “good” batsmen. I’ve defined this, rather arbitrarily, as someone averaging over 25 runs per match and having scored over 1000 runs in all formats of the game  - cricket fans this cut-off means Moeen Ali is “good” (5026 runs, 25.9 average), but Ravi Ashwin (3454 runs, 24.5 average) isn’t “good”.

I thought the percentage of low scores might remain similar between “good” and “bad” batsmen, but the “good” batsmen would be better at going onto bigger scores once they had “settled” i.e. reached around 10 runs. But it seems that better players are less likely to get very low scores (5 or less) than worse players.

The next highest unscored score in international cricket is 251. I’m sure we can all agree that with the Ashes coming up this winter in Australia, all eyes will be on England, but most importantly, will someone be able to score this?

I hope you enjoyed this post and for those interested in my analytical methods, the code for analysing can be found on Github. If you have any thoughts, questions or ideas of what to post next then these are very welcome, please email me at martin@butterflydata.co.uk.