Monday, May 28, 2007

Sunday, May 27, 2007

Sunday, May 20, 2007

Sunday, May 13, 2007

Sunday, May 06, 2007

Wednesday, May 02, 2007

One-day Cricket: Part 1

Recently I made a comment on a post on Australian success at the World Cup by Amit Varma in his India Uncut blog about how much I dislike the way they have changed the way the game is played. It is definetly not sour grapes. One thing I really like about Aussies is their grit. They play as a team, as any team game should be played. Last I heard that cricket is a team game. But I still can't watch the bluedgeoning work carried out by their batsmen or for that matter by Greame Smith or Jaques Kallis. Setting this aspect aside, which is very much subjective, I decided to look deeper into the statistics.


The recent debacle of Indian cricket team at the ICC Cricket Worldcup(WC) prompted me to dig deep into the performance of the team over the years and I have been trying to put together my analysis, if you can call it one. This is Part 1 of this exercise.

First, I would like to acknowledge the enormous amount of data available at Cricinfo websites' statsguru feature, from where I got all the data for this analysis presented below. While trying to analyze the record for India, I thought it would be good to have comparative data of two other teams, one of which has shown consistency and the other that has shown amazing improvement, for which I chose Australia and Sri Lanka, respectively. All of this is also confined to only one-day cricket.

Even though Cricket WC is the ultimate prize, finally it is one tournament. In the following analysis, I have only concentrated on non-WC matches in the period between world cups. For example, in the charts below 1983 means the matches played between the day after the finals of the 1979 WC and the day before 1983 WC, both days included.

First a general chart, for the teams (here I have included only 8 teams) between 2003 WC and 2007 WC. From this chart, it is not a surprise that the four teams that went to semi-final stage, were well on course over these 4 years.



Chart 1: Winning % for all matches played between March 24, 2003 and March 12, 2007



Let us begin with the 3 teams mentioned above. Let see the change in winning % for these 3 teams from pre-1975 WC. Actually there is no data for SL prior to 1979 as they played only in WC in those days. Post 83 is the period during which India has shown improvement, which is largely stagnant ever since, except 96 when it went to 60%. I still believe that was the best shot India had in all these post 83 years to win the cup again. Over the years, India has always hovered around the 50% winning % whereas Australia and SL have shown continuous improvement. There are dips in Australian performance too, especially during the period of pre-1979-1987, when lot of issue were affecting australian cricket. Packer series, the retirement of Chappell, Lillee, Marsh and the rebuilding of the team under Alan Border were the important ones. SL has really grown a lot since the pre-1996 WC season. Look at the chart below.




Chart 2: Win % during the intervening periods between world cups.


Now let me jump ahead for co-relations. One of the constant theme coming out of Indian cricket team selectors is that we want to play 7 batsmen and 4 bowlers. The reason being that the famed(?) can get any score. Over the years India has completeley ignored the bowling department. I don't even want to get into fielding. I wanted to see whether there is any co-relation between the win% and the R/6b parameter of cricinfo. R/6b is basically runs scored every 6 balls, run rate. The bowling counter part is C/6b, which is runs conceded per over. I plotted win % as a function of both these parameters and i don't see any co-relation. See for yourself in charts 3 and 4.

Chart 3: Win % over all years versus R/6b

Chart 4: Win % over all years versus C/6b.



On the contrary, when I plotted the difference between R/6b and C/6b, there is a very strong co-relation with win%. If this difference is close to 0, no matter which team, the win % is close to 50%. You lose every other match. Notice that for Australia and SL, the increase in this difference has lead to increased win %. It would be illustrative to do this for other teams as well. What this means is, you score as well as you can and you should also bundle out the opposition also. Actually this is more exemplified in the chart between win% and the difference between batting average and bowling average, which also takes into account the wicket lost/taken.




Chart 5. Win % versus the deficit R/6b - C/6b


Chart 6. Win % versus the deficit Bat. Ave- Bow. Ave

What do these charts tell about the composition of the team? Let us look at that at a later post.