Does the pitch matter in T20s?
Over at Cricinfo’s It Figures blog, Ananth suggested that 170 was a score Twenty20 captains could target when batting first to have a ~90% chance at winning. He mentioned that he ignored pitch conditions while conducting the analysis. This struck me as a very extravagant assumption to make — I thought the pitch would have a significant effect on the innings score. Surely there are some pitches where 140 is a decent effort, and others where chasing 180 is a walk in the park?
I decided to do some analysis to determine whether the pitch really has an effect. I used a simple method to see if there was really such a thing as a “180 pitch” and a “140 pitch”. Specifically, I checked to see whether the 1st innings run rate & the 2nd innings run rate in the same match were correlated. That is, if the first team scores at 9 runs/over, does the second team also tend to score quickly? If the two run rates are significantly correlated, that suggests that the pitch does affect the score that a team posts (assuming the pitch doesn’t change over the course of the match). However, my conclusions surprised me.
A scatterplot showing the 1st & 2nd Innings Run Rates from all T20 International Matches. Each match is plotted as a point (1st innings, 2nd innings)
Looking at the scatterplot, and best-fit line, there’s only a very weak positive correlation (r²=0.09), and the slope of the line is small as well. As a statistical aside, the slope is significantly different from zero with p=0.02, so there is indeed some effect, but it is small. This suggests that the wham-bam game of Twenty20 has nearly taken the pitch out of the equation. With batsmen riding their luck & going for broke, and edges running for four, the pitch doesn’t seem to matter much.
There are a couple questions you could (and should!) raise about this analysis:
- For the above analysis, I used every Twenty20 international played. What if we restrict the analysis to only matches between the top 8 sides, excluding the minnows like Bangladesh & Zimbabwe? I ran those numbers as well, and the results look remarkably similar. The slope is nearly identical, and the correlation is just as weak.
- I was perhaps a bit presumptuous when I said no correlation between first & second innings scores implies the pitch has no effect. A tighter interpretation would say that a change in the first innings score doesn’t change the second innings score much. There could be other reasons for this. One could be that all the pitches aren’t really different after all! The cricket boards and the TV folks want a good batting pitch for the matches, so perhaps the wickets in SA, England, India, and Australia aren’t as different as they are in Tests.
- Another factor I didn’t control for is that the 2nd innings team might naturally score at a higher rate than the first innings team, because they’re chasing a known target. They may come up short, but they usually go down swinging, rather than tamely playing out the 20 overs. The data I have suggest no significant difference in 1st innings & 2nd innings run rates, but who knows? The original regression suggests that the 2nd innings score increases slightly as the first innings score increases — this could be due to the effect I just mentioned, rather than due to the pitch.
The takeaway message for me is that perhaps it doesn’t really matter so much where the Twenty20 match is played. That should be good news for the IPL as it heads to South Africa for season two. Do post a comment if you have some thoughts about where I could take this analysis from here.
