Wednesday, December 22, 2010

Tendulkar and Bradman

Now Sachin Tendulkar's got his 50th Test century, the media have predictably revived the debate about whether he's the greatest batsman of all time. Well, of course not. Donald Bradman's clearly the greatest batsman who ever lived by a wide margin.
Batting averages have not shifted all that much in the past century. After a long career, a good Test batsman typically ends with an average of between 40 and 45; an exceptional one with an average of between 45 and 50; and an all-time great with a 50 plus average. That, at least, was the case before helmets became the norm, grounds grew smaller and bats more powerful. When helmets were rare, in the late 1970s and early 80s, there were four batsmen in the entire world with an average of over 50: Greg Chappell, Sunil Gavaskar, Vivian Richards and Javed Miandad. Chappell was replaced by Allan Border as the 50 plus Aussie in the 1980s. In the first sixty years that India played Test cricket, Gavaskar was the only batsman to end a substantial career with a Test average above 50 . Contrast this with the fact that the current Indian team itself has four batsmen averaging 50 plus: Sachin, Dravid, Sehwag and Gambhir. Pretty much every major team has one or two players in that category.
Judging by the stats, batting's become easier than it used to be; certainly no harder. It is reasonable to assume, then, that batsmen of the past would have had more or less the same success if transported to the present; their average would be in the same ballpark.
How, then, is there even a suggestion that Sachin Tendulkar, with an average in the mid-fifties, might be the equal of Bradman who averaged virtually 100? There simply is no comparison. Bradman is one of those outliers that defies comprehension; he is so far above any other batsman to have ever played cricket that he becomes a serious contender for the title of greatest sportsman of the twentieth century. I can't think of any sport where one person has left his contemporaries quite so far in his wake.
In any debate about the greats, Bradman should be left out of the equation, he's way greater than everybody else. What Sachin's done in the past year, I believe, is to put himself at the top of that group of 'everybody else'. About five years ago, Brian Lara and Sachin had similar records, and there was a valid debate about who was the preeminent batsman of their generation. Ricky Ponting had hit such a purple patch it seemed he might in the end overtake both Lara and Tendulkar. Well, Ponting's career graph has described a trajectory quite normal among great batsmen: a peak between the ages of 28 and 32, when the career average rises to the high 50s, and then a gradual falling off till it gets to about 50 at the time of retirement. Dravid's graph shows the same pattern, a rise to 57 or 58 and then a decline back to 52 - 53; The careers of Richards and Miandad also followed a similar arc. Sachin has defied this trend, by getting an improbable second wind late in the day. Having played more tests, scored more runs and hit more Test and ODI centuries than any other batsman in history, he's cemented a place as the second greatest batsman to have ever lived.


Anonymous said...

good that you finally put this article on. however, I feel it is not a right thing to compare Bradman and Sachin as the advantages and disadvantages of both the times they have played are very very different. Bradman might have an exceptional average, but I feel you need to keep the number of matches he played also to account. Sachin has come in at a time when cricket itself has taken a steep growth which no one could have foreseen.

Bradman also never had the baggage as much as Sachin had in his career to carry the expectations of a billion plus people, My take on this is if Bradman were to play today , his average would be in 60's if not in the 50's. Srikanth

Girish Shahane said...

Srikanth, variance decreases as the number of games played and the number of runs accumulated increases. Batsmen like George Headley or Graeme Pollock, who follow Bradman in the averages (though they're far, far behind) played less than 25 tests each, and we can't be sure if they'd have stayed at 60 if they'd played twice as many. But others like Barrington and Hammond played enough for us to be certain their averages were rock solid. Same holds true for Bradman. Besides, Bradman scored nearly 30,000 first-class runs at an average of over 95. There's usually a very strong correlation between first class and Test averages over the long run, and the 30K runs prove Bradman's 7000 in Tests were no fluke (it's tough to have a fluke run over 20 years in any case).
If you look at Graeme Pollock's first class record, he scored about 20K runs at an average of about 54. This suggests his Test average would have fallen a little had he played for longer, but clearly places him among the all-time great batsmen.
I, for one, have never understood why peoples' expectations would have a negative effect on performance. There are sportsmen who wilt under pressure and others who thrive. Sachin clearly falls in the latter category. The plain fact is that there are half a dozen current players who have records that are not that distinct from Tendulkar's; he isn't in any sense, an outlier, but more like first among equals. Bradman's record defines what the word 'outlier' means: an observation that is numerically distant from the rest of the data. I haven't come across that kind of outlier (an average over 50% higher than anybody else that has ever played) in any other sport.

Anonymous said...

Jeez...That was a great analysis. you are right, I have to agree , Bradman is definitely ahead of Tendulkar. Personally though, I would have loved to see how Bradman would have done in today's times. Srikanth

Ajith James said...

ok.. feel compelled to comment here.. first off, I have never watched Bradman play.. having said that, i have to differ with your utter conviction that he is better than Sachin. I am of the opinion that you simply cannot compare players from two different eras..

Bradman played 52 tests to Sachin's 177

Bradman's 19 away tests were all against England

Bradman has never faced quality spin bowling in the sub continent

Am not giving the above stats to prove that Sachin is better. Merely pointing out that its not fair to compare them. Bradman might have played Shane Warne and scored heavily against him.. the point is, you never know !!..

Finally, i think this expectations thingy that many ppl say is bull shit.. yes, there are expectations, but like you said, if people thrive on pressure, then it is even an advantage !!

Girish Shahane said...

Ajith, for everyone who speaks of Tendulkar facing immense pressure of expectations, and Bradman never having played on subcontinental wickets, there is another who will counter that Sachin has had it easier because he has played only on covered wickets, and played wearing protective gear like helmets and thigh pads. And that he has had a number of easy tests against opposition like Zimbabwe and Bangladesh, while Bradman faced constant tough opposition from the best or second best team in the world, namely England.
My central point, which none of those who have criticised my post address, or perhaps understand, is a statistical and factual one. It is this: In EVERY era of cricket the average score of batsmen in tests has remained fairly stable. Very few batsmen in any era have ended substantial careers with averages above 50. In fact, the percentage of test players with 50+ averages is highest in the current era, indicating that, if anything, batting is easier now. But let's disregard that, and agree that a 50 plus average over the course of fifty tests is a good indicator of greatness in batting. Tendulkar's stats, while extraodinarily impressive, are not that different from those of somebody like Jaques Kallis. However, Bradman's are way ahead of even his most talented contemporaries, the Hammonds and Huttons. If we agree, as most people do, that Hammong and Hutton were more or less on par with Tendulkar and Gavaskar, where does that leave Bradman? Streets ahead is the answer. But you know, second best of all time is not a bad position at all. I've placed Tendulkar ahead of Viv Richards, Brian Lara, Jack Hobbes, all those superb players. It's really OK if one guy was much better.

Ajith James said...

Bradman played majority of his matches - 37 out of 52 against England
..avg - 89.78 (cricinfo)

since you mentioned averages, i think the following questions are relevant:

If Sachin played against the same/similar English bowling attack at that time, what would have been his average ? For that matter, what would have been the averages for Ponting/Lara/Kallis against a similar attack ? Would it have been closer to 99 ?

If Bradman played today against the variety of attack that Sachin has faced, spread over about 200 + innings, what would have been his average ?

Sachin played ODI's in addition to Tests.. If Bradman played ODI's as well, would his Test average have been affected adversly ?

In my opinion, no one can answer the above questions..and these are just a few variables..

You have brought up a crucial point regarding the difference between Bradman and his contemporaries.. and also regarding the point that batting has got easier now..

Bradman and his peers grew up playing under conditions that existed during that era.. without protective gear, uncovered pitches etc.. when you grow up under a set of conditions, you are better suited to adapt to those conditions.. Sachin and his peers play during an era of covered pitches and protective gears.. that does not mean that they would have fared worse had they grown up playing under different conditions..

the difference between Sachin and his peers is less.. agreed.. but then, to discuss that point, you would have to start comparing the Laras and Pontings against the Hammonds and Huttons and you are back to square one ! :-)

Just saying that when you start comparing players from two different era's, there are too many variables and assumptions that makes such comparisons meaningless..

Girish Shahane said...

The English attacks of Bradman's time were certainly no worse than the English attacks Sachin has faced. Which contemporary English bowler is as fast or accurate as Harold Larwood, for example? So, to answer your question, Sachin would probably have had more or less the same average had he played the English attack of the 1930s that he has against contemporary English teams. But I'm sure you won't believe that, which is fine, since you believe comparisons are meaningless anyway. I think fine comparisons might be meaningless, but when there's an outlier like Bradman, the comparison becomes relevant because we can safely presume human abilities have not changed in the past century.

Ajith James said...

Jardine, Larwood.. great bowlers.. haven't watched them play, but have read about them..

I am sure you would agree that the likes of Akram, Waqar, Donald, Warne, McGrath are great as well..

Bradman was an exceptional batsman from his generation.. Sachin is one of the exceptional batsmen from his generation.. Thats where i would like to leave that comparison.. I must admit though that I have a bias towards a player that i have watched and admired compared to someone who i have just read about..

Girish Shahane said...

You keep shifting goalposts. None of the bowlers you mention is English. But go with your bias, it's an understandable one.

Dsylexic said...

since you are throwing out nice statistical terms like variance etc,let me introduce you to one more : regime change
that alone tells us why comparing avgs are is not a cynical is a statistical truth.bradman and tendulkar played under different conditions .if batting pitches are favorable in the modern era,fielding standards in the past were abysmal and the fielders were not supposed to stop a ball going to the boundary.
since the variables have changed significantly,using just the batting avg is an abuse of statistics

Girish Shahane said...

I disagree; as I said, the averages of the best test batsmen have held steady for almost a century; Bradman's the only major outlier.

Sivakumar Thiagarajan said...

Girish, Agree with your comments fully. There are three additional points I want to bring up in support of your argument

1. In every era the best player is compared to Don Bradman. Even contemporaries were named after him like 'Black Bradman' (George Headley). A 100 years later there will be another hero of year 2111 who will be compared to Bradman not to Tendulkar. I say this because in the long run heroes of each era Jack Hobbs, Wally Hammond, Len Hutton, Everton Weekes, Ken Barrington and even more recently Viv Richards have all faded away from daily talk. Only Bradman is remembered even today. So when people compare someone to Bradman it is an implicit acknowledgement of him being the best.
2. You spoke about variance. In his career, Bradman's LOWEST series average was 56 in the Bodyline series - which is Tendulkar’s average. If you exclude that series the next one was 67 – higher than everyone else’s averages. No player in any game can boast of being in form continuously for 20 years. Never before or after Bodyline has such extreme tactics been used to curb just one individual. It speaks volumes of the extent of respect and hear he instilled into his opponents – like no one ever has.
3. You spoke about all sports. Indeed there has been statistical analysis done on basic measures for all games(like points per game in basketball etc.) and a normal distribution plotted. Don Bradman lies at 4.4 sigma vs the second closest who is Pele at 3.7 sigma. Now that means Pele was the best among 10,000 players and Bradman was the best among 200,000 players.

I know it is difficult to compare two individuals from different eras. The game has evolved and changed drastically today. But if people make such a comparison, it needs evidence and reason. Our logic is that Tendulkar is the best of this era and Bradman was the best of 1930s. They are treated as equals because they were both the best in their era. This brings Viv Richards, Greg Chappel and a whole lot of others to the table as candidates. This is flawed as we need to dig deeper to find out how much better they were than the rest. The best yardstick and the only reasonable way then is to compare how far ahead someone was to his / her contemporaries - not only from the second best but also from the rest of the competition of that era. As mentioned, batting averages have only varied slightly over time and Bradman beat his second best and the rest of the crowd by miles.

Another way to look at it is this way. If someone has demonstrated the ability to rise up and shine consistently every year irrespective of the opponents / conditions (whatever existed then), it is logical to assume that such a person had extraordinary talent and hence will succeed in new situations as well – like facing Shane Warne, playing one dayers etc. It is like saying that computer genius like J J Donovan who programmed on Fortran, Cobol and Basic in his era may not may not be successful today in an era of Java as languages have changed even if he was brilliant during his era. If that is true, then I can probably beat Albert Einstein in today’s physics exams!

Girish Shahane said...

Your last paragraph is a great analogy, Sivakumar, thanks for posting.