A new measure for cricket greatness?

G=12N2μ i=1Nj=1 NRi- Rj

That, you have to admit, is a pretty snappy start to an article. Seeing as Spin readers are such a learned bunch, most of you will already have volume 6, issue 3 of the Journal of Quantitative Analysis in Sports folded up in a corner of your bathrooms, its pages all dog-eared from excessive use. But just in case there are a few unfortunates who missed out on the August edition I'll quickly explain that in this case N is the number of innings a batsman has played, the number of runs he has scored in completed innings and the bit in the middle is his consistency adjusted average.

And the G? That stands for something called the Gini coefficient, a means of measuring inequalities in distribution of statistical results. Tempting as it may be to take a diversion into the life and works of the fascist statistician who devised the coefficient, Corrado Gini, that can surely wait till next week. For now, the G can just stand for greatness, because that is what the formula attempts to capture. Supposedly those symbols sum up greatness with an economy and precision which would surely elude an army of sportswriters hitting random keys on a typewrite for an infinite amount of time.

It is part of a paper called The Bradman Class: An Exploration of Some Issues in the Evaluation of Batsmen for Test Matches, 1877-2006. Let's hand over to the authors, Vani Borooah and John Mangan, for a moment. "Suppose that batsmen are viewed as run producing machines of varying degrees of efficiency; that every time a 'batsman machine' is started he/it would score, with certainty, the same number of runs before stopping. The CAA represents the 'certainty score' of a batsman." So a Trescothick machine would produce X runs before stopping, whereas a less efficient Trego machine would produce Y runs. That is what the sum at the top of page will tell you.

Frankly the Spin can make as much sense of the intricate workings of this paper as a grizzly bear can of a teapot, but basically the two authors have set out to correct two problems that will be familiar enough for most cricket fans. The first is that batting averages do not take account of consistency across innings – a man who manages to make four fat centuries among 21 largely lean Test innings can finish his career with a deceptively good career average. Just ask Vinod Kambli. Secondly, averages take no account of the value of an innings to the team. The man who can graft his way to 75 out of 150 is surely worth more than the one who makes a habit chipping in with a century every time his team rattle off 500. (No, I didn't mention Ian Bell. Don't blame me for the leaps you've made in your own mind.)

The two men behind that snappily-titled statistics paper have devised calculations that produce averages unaffected by these two imperfections. Then they have used them to recalculate the rankings of the top 50 batsmen with the highest Test averages of all time.

The upshot is that the paper soon turns into an exercise in elucidating what everyone already knew – between Don Bradman and every other batsman, and yes, that includes Sachin Tendulkar – there is a gap as wide as the Andaman Sea. If you started up your "Bradman machine" up in the first innings of a Test, it would score 46 runs. No one else manages 40. In the second innings, when the pressure tends to be greater still, Bradman will make 68. No one else manages more than 50.

Cricket has always been backwards when it comes to using statistics, though I suppose that would seem a preposterous statement to any of the unconverted who find themselves leafing through the Records section of Wisden. Recently Andy Flower has been praised for adopting the kind of analytic techniques that were first used in baseball by Bill James in the late 1970s. James's work evolved into sabermetrics, the statistical study of in-game evidence in an attempt to reach objective conclusions. There are clearly areas in cricket where such study would be useful. I've written before about John Buchanan's early adoption of these kinds of techniques when he was coaching Australia.

Buchanan explained his approach to the numbers game like this: "1) Ignore existing cricket statistics - these are just the 'outcome numbers' of a process of getting there. 2) Search for valid and reliable process numbers that give a truer indication of performance. 3) Seek the numbers that 'guarantee' my team a win. 4) Use these numbers over time to look at trends in the game, and trends in individual and team performance." If you exclude No3, then that is an accurate summary of what Borooah and Mangan have tried to do with players' Test averages. But Buchanan wanted to know whether or not nightwatchmen were worthwhile, not compare players from across generations. An outsider can find a lot of interesting things in Borooah and Mangan's work, just as Gulliver did in Laputa. But the statisticians, just like Swift's scientists, are spending their time trying to extract sunshine from cucumbers.

The players who gain the most from their consistency calculations are Douglas Jardine, Charlie Davis, and Colin Bland. None of the three could be called a great batsman, though two of them had a claim on the title in other ways, one as a captain and the other as a fielder. On the other hand Virender Sehwag suffers, and drops down the rankings because of his inconsistency. Overall though, he still finishes up above Sachin Tendulkar, because Sehwag has a habit of contributing big scores to small team totals.

The trouble with their work is that beyond Bradman, greatness in cricket is never objective, always subjective. It is not a game of absolutes, and cannot be measured in terms of faster, higher or stronger. Style, character and charisma are far more important in shaping players' reputations. If you are interested in numbers do go and read the Bradman paper. But if you are interested in understanding greatness, go and get a copy of the Wisden Cricketer's anthology of their My Favourite Cricketer columns. It is a gem of a book, one I have read and re-read since it landed in my pigeon-hole at work. Good writers on great cricketers. They may not be as succinct as the statisticians, but they are a hell of a lot more revealing.

If you really do not have your August copy of the Journal of Quantitative Analysis in Sports to hand, you can read Borooah and Mangan's paper by following this link, signing in as a guest and then downloading the article.

THIS WEEK'S GEM FROM THE PATHÉ VAULT

Seeing as we have had WG Grace and Jack Hobbs, it must be time to have a look at Len Hutton. This is the report from the first day's play of the fifth Ashes Test at the Oval, 1938. "Even if Edrich and Hutton were to knock up a million," the reporter tells us, "the match score would still be level and Australia would keep the Ashes."

Hutton settled for 364. I shan't spoil Pathé's report on the rest of the match by telling you what happens, but will instead implore you to go and watch it yourselves the next time you've five minutes to spare.

This is an extract taken from The Spin, guardian.co.uk's weekly cricket email. You can sign up here.