Which batsmen thrive against the best bowlers?
Kumar Sangakkara made a triple-hundred and a hundred in the same Test in Chittagong recently. In the last 12 months, Sri Lanka have played four Tests against Bangladesh. In these, Sangakkara has scored 142, 105, 139, 55, 75, 319 and 105. In his last 12 Tests against one of the other seven top Test-playing teams, Sangakkara averages 48. This is a broader phenomenon in Test cricket. Bangladesh and Zimbabwe have conceded runs at an impressive rate to the world's top batsmen. Players from the subcontinent play these two teams more often than players from Australia or England.
A few weeks ago I described a method to determine the bowling and batting strength of a Test team. I have since found that a simpler method of measuring bowling strength is similarly correlated with Test results. I describe it briefly below using the example of India's first innings in the Cape Town Test of 2011. This method could be improved substantially with detailed ball-by-ball data, but since this data is not available for all Tests, I use innings-level data.
Sachin Tendulkar made 146 in this Test in Cape Town and had a memorable battle with Dale Steyn. But South Africa's bowling attack as a whole was not particularly strong. The table below shows the calculation of bowling strength for the South African attack. The individual averages are the bowler's averages at the start of the Test innings. Tendulkar made 123 in 231 balls against Morkel, Tsotsobe and Harris, and 23 in 83 against Steyn. Steyn bowled 120 balls while Tendulkar was at the wicket. Tendulkar faced 83 of those. In the other 37, Steyn took three wickets.
Tendulkar made 169 at the same ground in 1997 against a much stronger all-round attack (Donald, Pollock, McMillan, Cronje, Klusener and Adams) that had a bowling strength of 26.1.
The bowling strength for a particular team innings is simply the weighted average of each bowler in a bowling line-up at the start of the said innings. Weights are assigned according to the share of the bowling for each bowler in this innings. For example, Steyn bowled about 26% of overs in that Indian first innings in Cape Town, Tsotsobe bowled 22% while Morkel and Harris bowled 22% each. Kallis, who was part of the South African XI for that Test, did not bowl at all in this innings.
I've used this method to calculate the bowling strength for all team Test innings. The median bowling strength for a Test innings from 1877 to 2014 (including the Sydney Test of the 2013-14 Ashes) is 31.54. Bowlers on Test debut are assigned the median figure.
In this, the first of a series of posts, I look at all batsmen who have made at least 10,000 Test runs and divide their careers into innings where the opposition bowling was better than the median, and innings where the opposition bowling was worse than the median. I also do the same for these players' home and away Test records.
For the purpose of this post, I will designate attacks with bowling strength better than the median as strong attacks and the rest as weak attacks. One final point to be noted about this method of determining the strength of a bowling attack: an attack including the exact same bowlers can have two different strength measures in different Tests. For example, a South African attack playing in Sri Lanka, where Nicky Boje would bowl a lot of overs, would have a weaker strength measure than the same attack bowling in South Africa, where Boje's share of the bowling would be much smaller.
Eleven batsmen have scored at least 10,000 Test runs as of January 2014. Sunil Gavaskar and Allan Border, the first two players to reach this milestone, played 49% and 45% of their innings against strong attacks. In contrast, Ricky Ponting and Jacques Kallis played 32% and 33% of their innings against strong attacks.
It is no surprise that top batsmen perform strongly against weaker attacks. Some more so than others, though. For Brian Lara and Tendulkar the difference in their performance against stronger attacks and weaker attacks is about 7 runs per innings. For Mahela Jayawardene, it is about 22 runs. For Jacques Kallis, it is 20 runs. For Ponting, Border and Rahul Dravid, it is about 13 runs.
Only Tendulkar, Border, Sangakkara and Steve Waugh average over 50 against strong attacks away from home. Jayawardene has struggled against such bowling, averaging only 27 over 49 innings, scoring one century. Ponting made eight centuries in 42 innings against strong attacks away from home, but averaged only 44. These innings constituted 14% of Ponting's innings in Tests. By contrast, Lara played 30% of his Test innings against strong attacks away from home. Sangakkara has built an imposing record against such attacks. Border averaged higher than 50 despite making only five centuries in 52 innings against such attacks.
In home Tests, only Lara and Sangakkara average more than 50 against strong attacks, while Kallis, Border and Waugh average less than 40. Lara has done equally well against strong and weak attacks in Tests at home. Tendulkar's record against strong attacks in India is his worst performance of the six categories shown in this post. He averages a mere 46.
Over the past 15 years or so, we often heard how Dravid was a more dependable player than Tendulkar. A similar observation was often made about Gundappa Viswanath. Observations have also been made about Ponting and Kallis being better players than Tendulkar or Lara.
These figures suggest, fairly clearly in my view, that Tendulkar and Lara were superior all-round Test players compared to all the others mentioned in this list. They also suggest that Sangakkara is a freak of nature whose bullying of the minnows only turns a great Test record into a freakishly brilliant one. The same cannot be said of his great friend Jayawardene, whose bullying of the minnows turns a very good record into a great one. Dravid, Ponting and Kallis were contemporaries. Each played about 250 Test innings. Yet Dravid played about 15 more innings against strong attacks compared to Kallis and Ponting.
These are some of the things that can be studied in these charts. If you would like to see comparisons of specific batsmen, please propose groups of such batsmen in the comments below. I will present analyses in my next post.