December 18, 2012

Test XV - the final readers' selection

Selection of an all-time Test XV based on readers' votes

Shane Warne received the second-highest percentage of votes after Don Bradman © Getty Images

Finally we come to the moment billions (not that often misused word), a few hundred, have waited for. All the entries for the all-time best XV have been received, processed, cleaned up, shenanigans identified & put in place and the final Readers' XV determined.

The geographic distribution of responses has been given at the end.


1. All the valid entries were placed in a text file. I wrote a custom program to tally the reader selections to create a Player-Reader matrix. Each set of fifteen players was assigned a weighted index using a complex algorithm to pin down identical selections. Then I created multiple Excel files. One for the number of reader votes received. Another for the number of selections by each reader that made into the final XV. Which brings us here.

2. I imposed certain restrictions on the selection of each XV. I did not want the reader to specifically exclude any period whatsoever. This was essential to have a fair field of selection across the years. Another restriction was that a reader had to specifically confirm if Bradman was excluded. Only one reader, out of such three, reverted with a series of convincing explanations. The other two simply did not bother, hence their entries were not included.

3. A few readers had raised the possibility of other readers misusing the system to put in multiple entries. I had always been aware of such attempts and set up diverse steps to locate and exclude such attempts. These steps have been documented in the form of a MS Word document. But why start with the unsavoury??? A link to that document provided at the end of this article.

4. The final selection is strictly bases on 'one reader, one vote' system. Be it Martin Crowe, Rajesh, Madhu or me, we all get one vote each. However I would have given slightly more weight to these selections to resolve a tie but it was not required.

Summary of numbers

No of valid entries received: 205
No of players selected:       119
No player was selected by all readers
Highest selection :           204 (Bradman-99.5%)
No of players with single selection each: 31
The votes gathered by the top 15 players: 2210-72.5%

Top 30 selections

NoPlayerVotes% of Total ...No PlayerVotes% of Total
1Bradman 20499.5...16Hadlee 8642.0
2Warne 18791.2...17Lillee 6431.2
3Sobers 17786.3...18Kallis 6230.2
4Marshall 17082.9...19Hutton 4120.0
5Gilchrist 16982.4...20SF Barnes 4019.5
6Murali 15676.1...21Waqar 2512.2
7Lara 15475.1...22Dravid 2411.7
8Hobbs 14470.2...23RG Pollock 2411.7
9Tendulkar 13967.8...24Holding 2311.2
10Akram 13766.8...25BA Richards 18 8.8
11Viv Richards13465.4...26Flower 18 8.8
12Gavaskar 13364.9...27Sangakkara 18 8.8
13McGrath 11556.1...28Sutcliffe 18 8.8
14Ambrose 9847.8...29Sehwag 18 8.8
15Imran 9345.4...30Hayden 17 8.3

7 players were selected by over 75% readers, 13 by over 50%, 18 by 25% or more and 24 by just over 10%. The 30th best selection, Hayden, received 17 votes (less than 9%). 40 players have 10 or more votes. There are big drops after Gavaskar (#12 with 133 votes), McGrath (#13 with 115 votes), Hadlee (#16 with 86 votes), Kallis (#18 with 62 votes) and SF Barnes (#20 with 40 votes).

The Readers' XV

The following final XV has been selected strictly based on votes received. There is no doubt that this is an excellent group of players. While there could be individual differences of opinions from readers, there is no doubt that this is an all-time best XV representing countries and eras in a very fair manner.

Viv RichardsWin854050.24

There are 5 West Indians, 4 Australians, 2 Indians, 2 Pakistanis, 1 Englishman and 1 Sri Lankan in this elite list of fifteen players.

Bradman is in with a record 99.5% votes. It may resemble the ballot to elect the President of a banana republic. Well, Paul Sime need not fear any reprisals from the President's secret service. He had clearly explained his selections. My estimate is that if we had received 1000 votes, Bradman would have got 99.94% selection (Thanks, david !!!). The lily needs no gilding. And before someone comments that it is absurd to destroy the lily because melting is an essential step for gilding, let me include the original quote by Shakespeare - ''To gild refined gold, to paint the lily, to throw a perfume on the violet,... Is wasteful and ridiculous excess." (Milind, thanks!!!). So let us move on.

I must admit that I was startled by the second highest selection. While I expected Sobers at this position, to my pleasant surprise, it is Shane Warne who garnered 92.7% of the votes. This indicates the very very high opinion of informed readers in considering Warne as the best spinner of all time and a genuine match-winner.

He is, as expected, followed by Sobers who secured 86.2% votes. I am bewildered by the rather high 14% exclusion. I feel Kallis edged out Sobers in some selections while a few top batsmen may have elbowed out Sobers for the batting spot.

Gilchrist is next with 82.8% votes. Once again I am amazed that over 17% did not choose Gilchrist. I think Andy Flower, no less a keeper-batsman, took away quite a few votes. Possibly the solidity of Flower was preferred by some (nearly 9%). It is also possible that keepers with better keeping ability (on what basis, I cannot understand) like Knott may also have got the nod.

In fifth position is the undoubted master of aggressive pace bowling, Marshall with 82.3%. Quite on the expected lines. A true match winner and the best fast bowler amongst all the greats.

My two favourites, Muralitharan and Lara, appear next. Muralitharan just edging out Lara by two votes. Two wonderful champions, crowd-pleasers, fighters, elegant performers: all rolled into single packages. No real surprise that over 3 out of 4 readers selected these two players' players.

The next three positions are held by Hobbs, Tendulkar and Wasim Akram, with just under 70% of votes. It is a measure of the readers' understanding of the game that similar numbers have selected two maestros who played nearly 100 years apart. Hobbs and Tendulkar were masters of flawless technique and extremely consistent. That Hobbs received more votes than Tendulkar is a testament to the very fair manner in which the readers had approached this task.

Viv Richards, Gavaskar, McGrath and Ambrose occupy the next four positions. I expected a slightly higher position for Richards and Ambrose. But they are comfortably in. These 14 players were in my selection of 15 players

The last position was a closely fought one. Imran Khan finished comfortably ahead of Hadlee by 7 votes and secured the 15th spot. Maybe a different demographic distribution of readers might have got Hadlee in, as also couple of other batsmen like Kallis or Chappell.

I am extremely happy with the 15 selected and take pride in the fact that Imran was in my XV until the very late stages and Hadlee just about edged him out. At the same time I am happy at Imran's selection since most of the concerns were raised at his exclusion.

In summary, I am surprised that Warne and Kallis got higher votes than expected, and Sutcliffe received fewer. I am saddened that Greg Chappell and Ponting did not get more votes.

The 13 readers who selected 14 of the final XV

No selection matched the final 15. Apart from me, 12 others included 14 out of these 15. It can be argued that my selection closely resembles the final list solely because most readers relied on my list and altered a few spots but a glance down the most commented article reveals the thought behind these selections, The table below lists all these and the selection changes from the final list. KC (from US) matched my selection. Jay and Kaushik had identical selections as also Arnab and Rohith. That is all. These 6 are the only identical choices amongst the 200+ selections. Madhu, Cricinfo Stats wizard, got 14 correct while Rajesh, Cricinfo Editor, got 13 correct. Martin Crowe, who selected the wonderful quartet of Lillee, Hutton, O'Reilly and SF Barnes, matched the other eleven. These four find a place in the second XV.

Readers/WriterSelectedNot selected
Madhu RamakrishnanLilleeAkram
Lucky StrikeHuttonAkram
PradeepBarry RichardsViv Richards

The 4 readers who selected 1 and 2 players of the final XV

Now the other end. One reader, Mike (from Australia), succeeded in selecting only Bradman out of the final XV. However since his selection encompassed all the eras and countries it is fine. Pete, Dennis and Sam39083 selected two players each. All three selected Bradman and Sobers and 13 other players.

The Second XV

The second XV has been selected mostly based on numbers but also with a bit of tweaking to get the team balance correct. It would be silly to just go by the numbers. It may be a good idea if the readers do not come out with comments like "A has got 1 vote more than B, why was B selected?". It would be totally counter-productive.

Barry Richards
Greg Chappell
Graeme Pollock
Andy Flower (wk)
SF Barnes
Waqar Younis

The overwhelming number of votes given for Warne and Muralitharan meant that the other spinners got very few votes. Still the numbers at least were reasonable. It is possible that one could have gone for Kumble over Bedi or Ponting over Dravid/Chappell. But those selected were thoroughly deserving of their inclusions. I am certain that this team would give the Top XV a run for its money. They would certainly win a Test or two in a 5-Test series.

Readers should remember the magnitude of the task I have completed. Cutting and pasting teams sent by readers in different formats, correcting spellings (how many 'Marshal's), affixing Viv or BA as warranted, affixing RG or SM as warranted, following up and effecting changes, tracking all below-15 and above-15 selections and validating these, checking for duplicate entries, taking care of multiple changes to teams and finally the special work to weed out suspicious entries and so on. All these with a 83.5% physical condition.

So please accept that there could be a few errors. However I am certain that these errors would not cause any change to the final selection. Frankly the only votes that matter are those to Imran and Hadlee. Imran is 7 ahead and is very unlikely to be caught up, on account of my errors. And many of the entries rejected had Imran. Even if you point out some errors please do not expect me to rush in and correct those.

To view/download the complete Player-Reader-Matrix Excel file, please CLICK HERE.

To view/download Word Document about my steps to locate and remove multiple selections, please CLICK HERE.

The geographic distribution of responses

Afghanistan     1
Antigua         1
Australia      32
Bahrain         3
Bangladesh      1
Barbados        2
Canada          1
Germany         1
Iceland         1
India          75
Ireland         1
Israel          2
Italy           1
Japan           3
Netherlands     1
New Zealand     4
Pakistan       19
Saudi Arabia    1
South Africa    3
Sri Lanka       3
UnitedArabEmir  2
United Kingdom 23
USA            24

What next?

Let us all settle back and relax, me the most. My next article will be an alternate review of 2012 by first week of January 2013. I have to do a lot of work to get the simulation programs up. Whether I want to carry out a simulation or not is also unclear. Let us give this a break. All you guys have a great Christmas, New Year, Pongal, Sankranti and Id-e-Milad. Your comments are, of course, welcome and will be published with the normal response criteria.

Have a great holiday season

Anantha Narayanan has written for ESPNcricinfo and CastrolCricket and worked with a number of companies on their cricket performance ratings-related systems