A Rating for Batsmen in Test Cricket

In a recent post, i critiqued the ICC's Ratings method for batsmen. I found that while the ratings methodology is fairly sophisticated, it uses some criteria which seem problematic, namely - the result of the game and the quality of the bowling. On the face of it, these appear to be important criteria, but as i have discussed earlier, neither criterion is within the batsman's control. As such it is unfair to penalize any batsman in a rating based on some statistical construction which suggests that he may have faced inferior bowling to some other batsman. The other problem with the rating is in how it is calculated - the scale of 0  to 1000 is arbitrary, and there is no indication whether the difference between a rating of 875 and 900 is the same as the difference between a rating of 850 and 875. Finally, the criteria considered seem to measure the same thing over and over again. For example, the difficulty of scoring runs is measured both through the difficulty of the bowling and through the match total. Obviously, the match total can come about because rest of the batting may be out of form. The consideration of the match result favors players playing in stronger teams, because these teams are more likely to win.

Having criticized the ICC's ratings method, it is only fair that i should stick my neck out and present my own rating. In this post, i will describe a batting rating methodology for Test Cricket which measures how successful and how valuable a batsman has been in Test Cricket. This to me seems to be a reasonable to measure the "best". I consider three factors:

1. The career aggregate for the individual batsman.

2. The career batting average of the individual batsman.

3. The percentage of his teams runs that the batsman contributes.

These are the three basic parameters which between them completely measure the success and the value of a any given batsman to his team - How many runs the batsman makes, how prolific the batsman is and how significant are these runs to the team.  This third factor is important, because the goal of any batting side in a Test Match is to build partnerships - this is easier to achieve if there are 7 top class batsman, than if there is only 1 top class batsman and a number of other lesser ones. Therefore, it is more difficult for the batsman in the weak batting side to make runs than it is for the batsman in the strong batting side. The runs are more significant and more valuable for the weak batting side as well.

In any given population of batsmen (say all Indian batsmen), if a batsman has the highest batting average, the highest run aggregate, and contributes the highest percentage of runs to his teams total score, then this batsman is the "perfect" batsman within this population. All the batsmen in that population are ranked on a scale for each of the three criteria, ranging from the lowest to the highest. Any individual batsman's rating for each criterion, is the ratio of his record for that criterion to the highest number. So any Indian Test batsman score for batting average is the ratio of his batting average, to the highest batting average for any Indian Test batsman. Similarly, the same is calculated for the other two criteria.

This table is made up of all Indian and Australian Test batsman who have scored at least 1500 Test Runs (this is a basic number of runs which a batsman must score before he can enter the rating list). In addition, i have also included Brian Lara, Jacques Kallis and Gary Sobers. The final "Rating" in the table for each player is the simple sum of the players score for each criterion - aggregate, average and percentage contribution to the team total. My rationale is best explained by comparing the first three names on that list - Bradman, Lara and Tendulkar.

When Bradman retired, he had the highest run aggregate, the highest batting average and probably the highest percentage contribution to his teams score (a whopping 25% - note that this is measured by considering the batting average against the average cost of 10 wickets in terms of runs). This would mean, that when he retired, he had a perfect rating of 3.0. In the years since he retired, many batsmen have played more games than he did, they have scored more runs than he did, value of Bradman's runs to Australia. Australia average 393 per innings when Bradman played, and Bradman contributed 100 of those runs. Thats why Bradman is still head an shoulder ahead of the next batsman in the ratings. Sachin Tendulkar has been more prolific than Brian Lara, but Lara's runs have come in a significantly weaker batting side. Thus Lara has been more valuable to the West Indies, than Tendulkar has been to India. Ricky Ponting has been more prolific than either Tendulkar or Lara, but he plays in a strong Australian side and as such is less central to Australia's success than Tendulkar or Lara have been to India's or the West Indies success.

A batsman's rating can be calculated at any stage of his career provided he has made at least 1500 runs in his career. This rating methodology is also responsive to world records - a world record aggregate would change every batsman's rating because it would change the maximum. If you have two retired batsmen, their rank will never with respect to each other, but their rating will as newer players come along and play more than they do. While a batsman is playing, it is theoretically possible for him to reach the number 1 spot in these all time ratings.

Thus, this methodology allows us to do everything which the ICC's player rating does, and more - it helps us judge a player at any point during his career, as well as at the end of it and also place his position in history.

Do tell me what you think