The effect of venue on scoring rate in ODIs

How much does a particular venue affect the scoring rate in the limited over Cricket? It is well known, for example, that scoring rate is generally higher at Indian venues, as compared to say those in England. In this blog post, we quantify the effect that a venue has on the scoring rate in ODIs.

Ideally, we would have done an exhaustive analysis for all the international Cricket venues. However, even if we take the data from all the ODIs played in the last 10 years, we may not have a sample size big enough for each venue to quantify the effect conclusively. Instead of looking at each venue individually, we instead look at the ODIs played in each country and then analyze the scoring rate for the venues in individual countries. This gives us a sample size big enough for each country to draw conclusions that are statistically significant.

We look at the data of all the ODIs played from 2002 onwards. We analyze only the first innings data, since there is an inherent selection bias with the using the data from the second innings: teams generally calibrate their scoring rate in the second innings based on the target set in the first innings. Therefore, the numbers from the second innings cannot be used as a reliable indicator for the overall effect of the venue on the scoring rate.

Overall, there are 18 countries which have hosted ODIs in the past 10 years. To quantify the effect on the scoring rate, there are three average statistics that we can use: strike rate, batting average and runs per inning. We choose strike rate to quantify the venue effect, because the number of balls for each country is big enough to prevent any problem due to small sample size. The strike rate of each country is then divided by the overall strike rate during this period to get the venue effect. The numbers are summarized in the table below.

The Venue Effect Factor (VEF for short) tells us by how much the scoring rate at a given venue is more than or less than the overall average scoring rate. For example, The VEF for India is 1.082, which means that on an average teams score at a rate which is 8.2% more than the average when playing in India.

It should come as no surprise that two sub-continent teams, India and Pakistan have the highest VEF among all the venues. The other two sub-continent teams, Bangladesh and Sri Lanka, however, are at the other end of the spectrum. West Indies and England fall right around the average, whereas New Zealand, Australia and South Africa, all three of which are considered to be bowler friendly, end up being slightly above average.

Where can the VEF statistic find use? One is in calibrating the statistics of individual players. The batting statistics of players playing in the subcontinent are somewhat inflated because they play most of their matches on batting friendly tracks. The VEF can be used to scale their statistics accordingly (for example, the runs scored can be scaled using the VEF) so that we get a venue-neutral statistic when comparing players from different countries.

The second place where VEF can be used is in target resetting in rain affected games. Presently, in the Duckworth-Lewis method, no adjustment is done for individual venues. However, as the above analysis shows, scoring rates do vary across different venues, and so it makes sense to incorporate an adjustment factor based on VEF when resetting the target in rain affected games.