Scientific analysis ranks batsmen

[vcs]

Bradman best, Sehwag greater than Sachin: Study - Top Stories - Cricket - Sports - The Times of India

A 'scientific' analysis of batting achievements in Test cricket has found what most people already know - that Don Bradman was the greatest batsman in the game. A paper, written by two economists, analysed factors such as consistency of scoring, value of a batsman's runs to the team and home-away record, and found the Australian great leagues ahead of the rest on all counts.

That's hardly a surprise. But the paper has some far more interesting conclusions.

While just five Indians qualify in the top 50 on the basis of career averages, in none of the parameters does Sachin Tendulkar, the highest run scorer in Tests, emerge on top even among Indian batsmen.

For instance, in terms of value of runs scored to the team, Virender Sehwag is India's top batsman (overall rank 6), followed by Rahul Dravid and Tendulkar.

On consistency of scoring, Dravid (rank 4) pips Sachin (5), followed by Sunil Gavaskar (11), Sehwag (12) and Vinod Kambli (13).

In another ranking based on career contribution to the team score, Dravid and Gavaskar (both ranked 5) emerge as India's top batsmen. Then come Tendulkar and Sehwag (both 6) and Kambli (8).

The paper - The 'Bradman Class': An Exploration of Some Issues in the Evaluation of Batsmen for Test Matches, 1877-2006 - ranks 50 batsmen from all eras. Written by economists Vani K Borooah, University of Ulster and John E Mangan of University of Queensland, it was published in the Journal of Quantitative Analysis in Sports.

Some rankings, however, may be misleading as the paper uses batting averages only up to 2006. For instance, in 2006, Dravid was No. 4 in the all-time list with an average of 58. At present, his average is down to around 53.

A 'scientific' study in Australia of Test batsmen down the years has put Don Bradman ahead of all others in every parameter it uses to define batting class.

The paper works on the premise that the method of calculating batting averages glosses over many aspects of good batsmanship. "It does not take into account consistency of scores across innings: a batsman might have a high career average but with low scores interspersed with high ones; another might have a lower average but with much less variation in his scores," it notes. Secondly, the paper argues, batting averages do not reflect the value of the player's runs to the team.

The authors use several equations from economics to suggest new ways of computing batting that could complement the existing method and "present a more complete picture of batsmens' performance." Based on these "new" averages, the paper offers several revised rankings of the world's top 50 batsmen.

For measuring consistency, the authors use the Gini coefficient, a popular method for computing inequality in the distribution of outcomes. After applying the coefficient to the top 50, while Bradman remains No.1, South Africa's RG Pollock (ranked 2 on averages) falls three places to rank 5 and West Indian George Headley falls to 7 from No.2. Among Indians, Dravid and Gavaskar hold their ranks but Sehwag falls four places and Tendulkar two places.

Downloaded the paper and have attached it. Haven't gone through it though.

 

[silentstriker]

"Scientific" indeed.

As Sheldon says...hokum.

 

[Teja.]

Lol, tbh.

 

[vcs]

Turns out the attachment didn't work for whatever reason . Dunno what the problem is.

You can go and read it here.

 

[G.I.Joe]

Was in the papers today. They ranked Kambli ahead of Gavaskar. Credibility lost right there.

 

[Top_Cat]

haha, as if people aren't already suspicious of economists and their funky models. Way to advance the field, guys.

 

[Halfpast_Yellow]

Why is 'scientific' in quotes? Either it's an experiment/analysis conducted using principals consistent with the scientific method, or it's not.

Somewhat controversial results don't change whether a scientific process was used to arrive at the conclusion.

 

[pasag]

Can't access it through the uni database.

 

[Top_Cat]

Why is 'scientific' in quotes? Either it's an experiment/analysis conducted using principals consistent with the scientific method, or it's not.

Somewhat controversial results don't change whether a scientific process was used to arrive at the conclusion.

The above is a broad definition, though. I mean, you can call everything from a case study analysis to RCT 'scientific' but there's massively varying levels of generalisability/quality of evidence between them.

Eco analysis falls in that continuum too. Mathematical models, perfectly scientific, don't have to be based on any real data.

 

[vcs]

Can't access it through the uni database.

You don't need any license or login id, just go here, click on the "Download" link, enter a few details and you're done. I'd attach the paper but attachments aren't working for me for some reason (any ideas, Mods?).

 

[silentstriker]

Somewhat controversial results don't change whether a scientific process was used to arrive at the conclusion.

What does that even mean in this context? I don't see how you could do the equivalent of a double blind study of batsmen and bowlers to even begin to get enough relevant data so you can come to a 'scientific' conclusion.

 

[honestbharani]

Can some just post the top 50 here? I cant get anything with that atachment.. It is asking me to create a log in

 

[Athlai]

Jesse Ryder, WAG

 

[Himannv]

Can some just post the top 50 here? I cant get anything with that atachment.. It is asking me to create a log in

Just click that link that vcs posted. Click download and enter a few details. Cant paste it here cause its a PDF and the table doesn't paste properly.

 

[Avada Kedavra]

Can some just post the top 50 here? I cant get anything with that atachment.. It is asking me to create a log in

Bradman, D.G. 100 65 165 1 1 1
2 Pollock, R.G. 61 24 85 4 5 5
2 Headley, G.A. 61 44 105 2 2 2
2 Sutcliffe, H 61 15 76 7 12 5
3 Paynter, E 59 13 72 9 14 8
3 Barrington, K.F. 59 12 70 11 15 5
3 Weekes, E.C. 59 29 88 3 3 4
3 Hammond, W.R. 58 23 81 5 6 5
4 Dravid, R.S. 58 13 71 10 14 5
4 Sobers, G.S. 58 12 69 12 15 6
4 Ponting, R.T. 58 11 68 13 16 8
5 Kallis, J.H. 57 9 66 15 18 7
5 Hobbs, J.C. 57 18 75 8 9 4
5 Walcott, C.L. 57 13 70 11 14 6
5 Hutton, L 57 24 81 5 5 4
6 Tendulkar, S.R. 56 14 70 11 13 6
7 Tyldesley, G.E. 55 11 66 15 16 5
8 Davis, CA 54 3 57 22 23 6
8 Kambli, V.G. 54 15 70 11 12 8
8 Hayden, M.L. 54 16 70 11 11 7
8 Chappell, G.S. 54 14 68 13 13 6
8 Nourse, A.D. 54 17 71 10 10 4
8 Sehwag, V 54 26 79 6 4 6
9 Lara, B.C. 53 29 83 4 3 3
9 Miandad, J 53 11 63 16 16 6
10 Inzamam U.H. 52 10 62 17 17 7
10 Ryder, J 52 5 57 22 21 10
10 Flower, A 52 15 67 14 12 5
10 Smith G.C. 52 19 70 7 8 7
11 Gavaskar, S.M. 51 18 69 12 9 5
11 Waugh, S.R. 51 1 52 24 24 9
11 Mohammad, Y 51 10 61 18 17 7
11 Border, A.R. 51 5 55 22 21 8
12 Richards, I.V.A. 50 11 61 18 16 7
12 Gilchrist, A.C. 50 -2 48 25 25 11
12 Compton, D.C.S. 50 11 61 18 16 6
12 Worrell, F.M.M. 49 10 59 20 17 7
13 Mead, C.P. 49 14 63 16 13 7
13 Bland, K.C. 49 20 69 12 7 7
13 Mitchell, B 49 14 63 16 13 5
13 Jackson, F.S. 49 11 60 19 16 6
14 Khan, Y 48 10 58 21 17 7
14 Harvey, R.N. 48 14 62 17 13 5
14 Walters, D.S. 48 7 55 23 20 7
14 Ponsford, W.H. 48 9 57 22 18 9
14 McCabe, S.J. 48 14 62 17 13 8
14 Jardine, D.R. 48 -8 40 27 27 11
14 Martyn, D.R. 48 -3 45 26 26 10
14 Dexter, E.R. 48 4 52 24 22 8
14 Jayawardene,D.P.48

 

[honestbharani]

yeah.. seems like they ranked by each category and the rankings are just as we might expect. As G I Joe said, when you place Kambli above Gavaskar, there is no real credibility left for the method, however scientific it may have been.

 

[Sir Alex]

Must be true.

 

[vic_orthdox]

Analysis purely by numbers is rarely going to make perfect sense, and won't reflect what is the opinion of serious cricket followers. What is interesting is looking at those anomalies it throws up because generally the lists will make a bit of sense, but there will be odd things about it. Also how it separates some contemporaries, like Viv Richards and Greg Chappell.

Other examples include Mohammad Yousuf over Viv Richards, Jack Ryder over his contemporaries Ponsford and McCabe, and Graeme Smith over Gavaskar.

 

[Ikki]

I didn't read the pdf but did they take into account giving more points for a bigger sample? Kambli and a few others are placed a bit too high up. A cut-off of about 40 innings makes the list look more credible.

 

[Matt79]

Analysis purely by numbers is rarely going to make perfect sense, and won't reflect what is the opinion of serious cricket followers. What is interesting is looking at those anomalies it throws up because generally the lists will make a bit of sense, but there will be odd things about it. Also how it separates some contemporaries, like Viv Richards and Greg Chappell.

Other examples include Mohammad Yousuf over Viv Richards, Jack Ryder over his contemporaries Ponsford and McCabe, and Graeme Smith over Gavaskar.

Exactly - it's simply a different method of considering the same names. If we use the same method all the time, there's precious little to think about or discuss. These are always interesting discussions both for what such studies confirm from our existing views, and for what they throw up that surprises. We can certainly decide that in the broader scheme of things, whatever model they've used doesn't change our personal views, but I'm always loathe to simply say: "This produced a couple of odd results, hence the whole exercise is pointless/biased/stupid".