Who is the greatest test batsman of all time? In a follow up to a recent paper I created a media furore by suggesting that India’s Sachin Tendulkar had eclipsed Australian great Sir Donald Bradman in terms of career performance.
The spirited reaction to my research is perhaps not surprising. As the Little Master prepares to step into the crease for his final Boxing Day Test against a much-diminished Australian opposition, it’s inevitable his remarkable legacy will begin to be dissected by fanatics of the game. Never mind the fact he’s chasing an elusive hundredth hundred.
That said, it seems the mainstream media has not fully understood the significance of my result, so I’m writing here to explain what it does and does not mean.
Batting averages: just not cricket
The most common way to assess a batsman’s career is to look at their average, which is simply the ratio of runs scored to the number of times the player has been dismissed.
Bradman, who is widely regarded as the greatest batsman ever, has an average of 99.94 runs per dismissal. This is phenomenally high, as an average of 50 is generally considered to be a world class performance. Sachin Tendulkar, on the other hand, has a relatively low (albeit still exceptional) average of 56.02.
But the batting average has problems. Cricket fans will often point to factors such as the quality of opposition, as well variations in playing conditions and equipment over time as reasons not to completely trust a player’s average as a measure of performance.
These concerns are reasonable, but they can be incorporated into a statistical analysis quite easily by comparing batting performance relative to historical means. One factor that is not so easily modelled is the issue of player longevity – some players simply play more cricket than others. This issue is actually implicitly accounted for when people cite Bradman as the greatest ever.
While Bradman’s average is exceptional, it is actually not the highest in test cricket. That honour belongs to little-known West Indian cricketer Andy Ganteaume, who played only one test innings and was dismissed for 112, leaving this as his career average.
Cricket fans are correct in noting that Ganteaume is not better than Bradman, and to ensure cases such as his do not appear in rankings of performance, a “qualification” is usually used to filter out extreme results. In a list of great batting averages it is required that a batter plays 20 tests before their name can be included. Such a system ensures that all players have had ample time to demonstrate their quality at test level.
Better than average
To an economist such as me, the above system looks rather odd. A batsman who plays 19 tests does not get included in the rankings, but a batter who is selected for 150 tests gets treated no differently than one that only just makes the cut. As sustained success is surely better than transitory success, there is a need to alter such a system to account for longevity in performance more effectively.
If we look at it like this, the question becomes, “how much weight should be given to a batter’s longevity versus his average in assessing his career performance?”
This question can be answered with a little economic theory. A batsman is only of use to his team if he averages more than his replacement player could have; a value known as the player’s “opportunity cost”. For test cricketers this value is approximately 40, although it changes from season to season. This means a batsman is contributing to his team when he scores above this, and detracting from his team’s performance – and risks being dropped – when he scores below this.
The opportunity cost may be plugged into an index to provide the correct weighting between these two facets of performance. Such players may be ranked by the extent to which they exceeded a “benchmark” player of their generation over the course of their career. This gives a sense of the total value of their career performance.
It is on this basis that Tendulkar (who has sustained his average for 184 tests so far) has slightly eclipsed Bradman who played only 52. Both players are miles ahead of the rest of the pack.
Mastering the stats
Bradman fans need not be too dismayed by my result, however. While it is fair to say that Tendulkar has been slightly more valuable than Bradman in aggregate, it is equally correct to say that what took Tendulkar 184 tests, Bradman achieved in only 52.
Perhaps the best way to think about the result is this: innings for innings, Bradman is vastly superior to Tendulkar, just as Ganteaume is slightly ahead of Bradman.
But if the Australian selectors could unearth another Bradman (restricted to playing just 52 tests) or another Tendulkar, they would be wise to opt for a local version of India’s Little Master. Just.
Join the conversation
Comments (5)
Neil Pemberton
(logged in via Facebook)
Bradman was good for a century better than every third time he batted in Tests. 29 centuries from 80 innings, or 36% if you like. Twelve of these scores were double centuries (including a 299), and two of those were over 300 (334 and 304).
Sachin scores a ton roughly every sixth time he takes the crease in a Test match. 51 centures from 304 innings, or 16.8%. Sachin's top score in Tests is 248. Six of his centuries are 200 or more.
Most of Bradman's runs were scored against the crack England teams of the 30s. Most of Sachin's runs have been scored against opponents of lesser standard, due to the highly variable standards of international cricket during the 90s and 00s.
Moreover, Bradman was still capable of scoring a Test century at the age of 40. Sachin is younger but is enduring a century drought.
Surely the raw figures tell the story. Bradman's appetite for runs is still unmatched, even today.
Jock norton
Psychologist (logged in via email @vtown.com.au)
Is there not also the fact that much less test cricket played in Bradman's era, so he had far less opportunity for "sustained success" and "longevity"? A Bradman unearthed now would not in fact be restricted to 52 tests.
And here is another analysis: compare batting averages in a team's first vs second innings. Batting in the second innings (often the last innings and on a wearing pitch) can be highly challenging. Few players average more with their second innings contributions than their first…
show full comment
Stephen Tanner
traveller (logged in via email @ozemail.com.au)
Nice 1 Nicholas, its a good bit of fun.
You've taken a really good pub debate to another level.
To Jock and Neil
'But if the Australian selectors could unearth another Bradman (restricted to playing just 52 tests) or another Tendulkar, they would be wise to opt for a local version of India’s Little Master. Just.'
Did you guys read or understand this caveat? Maybe you should read Nicks paper too.
Now Nick, next paper please. If you were in the playground tomorrow, had a time machine and could pick Sobers, Kallis, Bradman etc. first, who would it be.
Jock norton
Psychologist (logged in via email @vtown.com.au)
So where are your data, and your mathematical justification for this claim, so that we can interrogate the method? Otherwise we will simply exercise the Hitchens axiom: WHAT CAN BE ASSERTED WITHOUT PROOF CAN BE DISMISSED WITHOUT PROOF.
Mat Hardy
(Lecturer in Middle East Studies at Deakin University)
I saw a paper a few years ago that modelled and weighted batting contribution according to total team score. That is to say, scoring 50 when the whole team was dismissed for 140 is amore important contribution than 100 when you declared at 6/649. This negates factors like opposition quality.
IIRC Bradman scored something like 1/3 of all Australia's runs during his career, an achievement vastly beyond anything anybody else has ever acheived.