All-time visiting XIs - part one
I must thank Pawan Mathur for providing the spark, by asking for the best all-time XI to tour India. That enabled me to think of going all the way and creating this pair of articles. This was indeed a massive task inasmuch as it combined raw analysis, objective deductions, some subjective considerations and finally, nuanced team-selection processes. Some elements of my own personal preferences come in but I would put that at no more than five percent. I have been true to the numbers as much as I could have been.
What is the objective? Stated in a single sentence, it is to determine the best XI to tour each of the major Test-playing countries. All the selected players should have performed outstandingly in the respective countries but must also form an effective balanced XI which should be capable of doing well in imaginary series on the Elysian fields against the best XI the home team(s) could put together. Of course, the local conditions of the country would be replicated for these Tests: a few renowned pitch curators would already have reached there! All the players who visited the concerned country should be considered and should have equal chances. Most importantly I had to constantly remind myself that this was a true case of peer performance appraisal, despite the fact that I was analysing players from different teams. The common ground between the players was the country they visited.
There is a world of difference between selecting an all-time XI that visited a specific country and performed well there and selecting an all-time XI for the country. A lot of data is necessary but we cannot let the data overwhelm us. We must be able to clearly understand the importance of players being productive, performing and having a clear understanding of the level of opposition faced by the player. At rare times we should be prepared to take left-field decisions. We should not select the bowler combination blindly, based on pre-conceived ideas. We must be able to go against the wind.
Finally I want this analysis and tables to provide a comprehensive insight into the away performances of players: batsmen, bowlers, allrounders and wicketkeepers.
Let me outline the steps below. Let us first take batsmen.
1. A cut-off of 500 runs is used to determine all the qualifying batsmen. England had 129 batsmen and Sri Lanka 14. But some base has to be laid down. This means that the player participated in a very good four-match series or a good five-match series at a minimum. If the player achieved this in a three-Test series, let us accept that he has achieved something spectacular. Enough care has been taken to ensure that flashes in the pan do not find their way in.
2. Scoring runs is an important factor. More runs indicate productivity and experience in handling the local conditions. But it is also true that more tours and more Tests mean more runs. Hence this is assigned 33.33% weight. The points to each batsman are allocated in a firm top-down manner. The scorer of the maximum runs in this country is assigned 33.33 points and the others are assigned proportionate points. This will ensure equality whether it is Australia (highest by Jack Hobbs, with 2493 runs) or New Zealand (Javed Miandad leading with 928 runs), and ensure that the concept of peer comparisons is very strongly implemented.
3. If there is a single reason why I considered abandoning batting average in favour of the highly nuanced RpAI (Runs per Adjusted Innings), it is the case of Jimmy Adams. Adams played three Tests in India, scoring 39, 81, 125*, 23, 174* and 78*, and came out with an average of 173.33. This is a monster number which would kill all logical processes. No way am I going to take the 400*, 365* and in Adams' case, 174* as "zero innings". Most readers would be familiar with the RpAI concept. A brief explanation follows.
Let us say that a batsman has a career RpI (Runs per Innings) of 40 and scores of 15, 24*, 86*, 12*, 18, 30*, 121*, 140, 91 and 23* in ten innings under consideration. His batting average is a way-out 140.00 (560/4). His raw RpI is an unfair 56.00 (560/10), since he has remained not out on four small scores. To determine the RpAI, my own creation used extensively for my analyses, I take the innings as completed (1.0) if it exceeds the career RpI or he is dismissed. This is a very fair base indeed. To take care of the contentious low not-outs I take the proportionate value of the innings. In this example, the RpAI is 68.50 (560/(1.0 + 0.6 + 1.0 + 0.3 + 0.75 + 1.0 + 1.0 + 1.0 + 1.0 + 0.525). It is a very balanced measure and does not penalise the middle-order batsmen for their not outs (as the raw RpI would have done), nor the early batsmen for their getting out more often than others (as the average does).
So I will use the RpAI, which is a total performance measure for this analysis. In view of the importance of this measure, it is assigned 44.44% weight (two-thirds of the remaining 66.667%). The points to each batsman are allocated in a firm top-down manner. The highest RpAI value in the country is assigned 44.44 points and the others are assigned proportionate points. This will work well whether it is New Zealand (Highest by Wally Hammond, with RpAI of 214.00) or Australia (Hobbs, leading with a middling RpAI of 73.20).
4. David Gower scored 746 runs in the West Indies against a collection of the most devastating pace bowlers who set foot on the Test arena: Michael Holding, Andy Roberts, Malcolm Marshall, Colin Croft, Joel Garner, Patrick Patterson and Courtney Walsh. Ricky Ponting scored 1169 runs in the West Indies against a collection of average bowlers. These two are not comparable. To take care of this, I have determined the weighted average of the bowling quality faced by each batsman. Every innings he played is weighted by the bowling quality specific to that innings, and an average arrived at. This weighted average of bowling quality is measured against the median value and an index arrived at.
This index is determined at either side of 1.0. Gower's index is a very high 1.33 and Ponting's a somewhat low 0.84. Thus it can be seen that this is an exact measure and does not make any assumptions or estimates. This measure is assigned 22.22% weight. Points to each batsman are allocated in a different manner. An index value of 1.0 is allotted 22.22 points and the others proportionately. This is to take care of the wide variations in bowling quality across the ages. The index is an exact number and the batsman gets above-par or below-par values depending on the type of bowling faced.
The three values are summed and a total Batting Index arrived at. The top ten batsmen for each country are taken for selection. To the extent possible, I would do the selection from this lot of ten batsmen per country. The only reason I might have to go out is if there is a dearth of batsmen for a specialist batting position, like opening. I am happy to state that I had to do this just twice.
1. A cut-off of 25 wickets (20 for Pakistan, because of paucity of Tests) is used to determine all the qualifying bowlers. This indicates a very good performance in a four-Test series. England had 86 bowlers who qualified and Sri Lanka 11.
2. Capturing wickets is an important factor. More wickets indicate productivity and experience in handling the local conditions. But it is also true that more tours and more Tests mean more wickets. Hence this is assigned 33.33% weight. The points to each bowler are allocated in a firm top-down manner. The bowler with the maximum wickets in this country is assigned 33.33 points and the others are assigned proportionate points. This will ensure whether it is England (highest by Shane Warne, with 129 wickets) or Sri Lanka (highest, again Warne, with 48 wickets), the concept of peer comparisons is very strongly implemented.
3. There is no problem with the wonderful measure of bowling average, which is a lovely composite of the strike rate and bowling accuracy. The bowling average, in my opinion, is the best cricketing measure available, especially across many Tests. So I will use the bowling average, which is a total performance measure for this analysis. In view of the importance of this measure, it is assigned a 44.44% weight (two-thirds of the remaining 66.667%). The points to each bowler are allocated in a firm top-down manner. The lowest bowling average value in the country is assigned 44.44 points and the others are assigned proportionate points. This will ensure that whether it is South Africa (lowest by George Lohmann, with 5.80, which I have normalised to 10.00) or Pakistan (Muttiah Muralitharan with 21.46), the concept of peer comparisons is very strongly implemented.
4. Iqbal Qasim captured 29 wickets in India against a collection of some of the toughest batsmen who ever assembled: Sunil Gavaskar, Dilip Vengsarkar, Mohinder Amarnath and Mohammad Azharuddin. Fazal Mahmood also captured 29 wickets against India against a very much pedestrian collection of batsmen. To take care of this, I have determined a weighted average of the batting quality faced by each bowler. The wickets a bowler captured in every spell are weighed by the batting quality of that innings and an average arrived at. This weighted average of batting quality is measured against the median value and an index arrived at, as explained under batting.
This index is determined at either side of 1.0. Qasim's index is a high 1.12 and Mahmood's, a somewhat low 0.71. Thus it can be seen that this is an exact measure and does not make any assumptions or estimates. This measure is assigned 22.22% weight. The points to each bowler are allocated in a different manner. An index value of 1.0 is allotted 22.22 points and the others proportionately. This is to take care of the wide variations in batting quality across the ages. The index is an exact number and the bowler gets above-par or below-par values depending on the type of batting he bowled to.
The three values are summed and a total Bowling Index arrived at. The top ten bowlers for each country are taken for selection. To the extent possible, I would do the selection from this lot of ten bowlers per country. The only reason I might have to go out is if there is a total dearth of a spinner. I am happy to state that I did not have to go out of this collection of ten top bowlers for any country.
The wicketkeeper has his own index which is 60% for the keeping ability, the sole measure available being DpT (Dismissals per Test) and 40% for the batting index. The allrounders are evaluated on an equal 50-50 allocation between batsmen and bowlers.
The six Tests Pakistan played at Dacca will be considered home Tests for Pakistan since Dacca was part of Pakistan then. The UAE Tests are strictly neutral since UAE does not play Test cricket. However, there are six other Tests which need some explanation. The three Tests in 1912 in England between Australia and South Africa, the two Tests in 2010 in England between Australia and Pakistan and the Test in 2002 in Sri Lanka between Australia and Pakistan. All these Tests have been considered as visiting Tests for both teams. Out of these six, the last one is the most relevant since Ponting and Warne did very well for Australia. In view of the smaller number of players who have played in Sri Lanka, I cannot really afford to leave out this Test. The matches in England do not add much to the key players' numbers.
We can debate till the cows come home, and until they leave the next morning, whether the base longevity measure like runs/wickets should get 25% weight or 40% weight. I have gone on common sense and my own insights. There are three important measures. The longevity-related measure has been given a third of the weight and the two other performance and quality measures share two-thirds. Any argument on this can only be theoretical.
To the extent possible, I would select six batsmen, a keeper and four bowlers. I would make sure that one of the batsmen is a real allrounder and could easily share the extra bowling load. Among the batsmen, I would strive to select two opening batsmen and if not possible, one opening batsmen and another top-order batsman who has successfully opened in Tests. The bowler combination is likely to be three pacers and one spinner. In rare cases, and if there is an excellent spinner available, this might change to two plus two.
I will provide the top-ten tables for batsmen and bowlers for each country and the final XI. Then the selection will be covered in reasonable detail. In view of the length of the introduction and process descriptions, the first part will include the first three Test-playing countries: Australia, England and South Africa. The second part will cover the other five countries/regions: India, New Zealand, West Indies, Pakistan and Sri Lanka. At the end I will provide a brief summary of the selected players.
I have to apologise to the Bangladesh and Zimbabwe followers. It is impossible for me to get sufficient data for these two late entrants. Tables with three batsmen or three bowlers do not help in team selection.
A captain will be selected based on my own familiarity with the players' achievements. Maybe a selection or two would be a surprise. Readers must remember that I have followed Test cricket for well over half a century.
First, the tables listing the top ten batsmen and bowlers to visit Australia.
The selection of the eleven players to tour Australia is quite easy because we have a history of over 13 decades of Test cricket. The multitude of Tests ensures that the cream always rises to the top. Barring a single position, the team virtually selects itself.
The best ever opening combination that walked side-by-side on to the cricket field is there at the top of the table. Jack Hobbs and Herbert Sutcliffe scored over 4300 runs across a span of 26 years. They opened the innings 15 times and scored 1202 runs at an average of just over 80. This glittering opening saga was studded with six partnerships over 100. They are followed by another equally great England batsman, Wally Hammond. Over 1500 runs at a RpAI value just over 56 makes this trio at the top a formidable one.
Then come two modern greats: Sachin Tendulkar and Viv Richards. Significant experience, but with RpAI values, just below 50. The wealth of experience at the top is unmatched. At No.6 is Aubrey Faulkner, whose performances in Australia are way above what the other allrounders have achieved. The bowling of the two allrounders, Hammond and Faulkner, is a bonus. Note Virat Kohli's presence in the top ten. Another good tour, he would be in the top five and it would be difficult to leave him out.
The keeping position presents a dilemma. Overall, Junior Murray of West Indies has better figures, however achieved across five Tests only. So I have gone for the overall excellence of Alan Knott, exhibited over many years and Tests. In 13 Tests, he dismissed 51 batsmen and scored 643 runs at 29.62.
The three pace bowlers select themselves. The performances of Richard Hadlee, Curtly Ambrose and Sydney Barnes, with over 70 wickets at excellent averages mean that they would just have to walk into the team. The fourth position presents a dilemma. Michael Holding is there in the fourth position. However I need to strike a balance and all the pitches in Australia help spinners, with their bounce. So I have selected Bobby Peel, who played between 1884 and 1896, to fill the spinner's position. He is placed in the tenth position. The next spinner is Lance Gibbs, in the 20th position with a high average of 33. Hence I am going to stick with Peel, who had three good tours of Australia.
The final team for Australia:
I have selected Richards as the captain to lead this team to Australia. Lest we forget, the only time Richards captained in Australia, it was an overwhelming win for West Indies.
The selection for the all-time best XI to tour England is less cut and dry. Surprisingly, the modern players have done quite well in England. Graeme Smith was magnificent in England and scored well over 100 runs per Test. His RpAI was a superb 64 and he walks in as the first choice opener. Gordon Greenidge suffers only in comparison with Smith. Otherwise he has done very well in England and is the obvious choice to partner Smith. The availability of two world class openers precludes the necessity of selecting someone like Rahul Dravid to open the batting.
For about fifteen seconds I thought of leaving out the No.3 batsman and then decided to play it safe. I did not want to get lynched. Don Bradman it is at No.3. His RpAI in England at 91.14 is better than his career RpAI of 89.50. He loved touring England.
Richards and Dravid get in at Nos. 4 and 5 by comfortably out-performing their rivals for these positions. Their RpAI values are above 60, truly remarkable performances. This measure gets Dravid ahead of Allan Border, who is placed above him in the table. At No.6 is arguably, the greatest batting allrounder ever, Garry Sobers. The numbers speak for what Sobers achieved in England: 1800 runs at a near-50 RpAI and 62 wickets at just over 30.
The keeper's place is taken by Gilchrist, although I must say that Brad Haddin ran him close. They played ten Tests each, scored just over 500 runs and dismissed 45 and 49 batsmen respectively. Gilchrist had the better RpAI (37.21 vs 28.50). In fact, Haddin had higher rating points. However it came down to whom I wanted batting at No.7 and I decided that would be Gilchrist.
The bowling selection was rather easy. Warne is the only bowler to have captured over 100 wickets in a foreign country. He dismissed Mike Gatting with his first ball, and Steve Harmison with the last ball he bowled in England. In between he delivered 6749 balls and captured 127 wickets. All at a pace-bowler like average of around 22. The next two bowlers, Marshall and Glenn McGrath have almost similar bowling stints in England and finished with identical rating points. They select themselves also. Then come a host of fast bowlers, Dennis Lillee, Terry Alderman, Ambrose and Garner with similar rating points. I decided that I have no reason to go past Lillee, the first among these lot.
The final team to tour England:
It is not rocket science to anoint Bradman as the captain.
3. South Africa
The selection of the team to tour South Africa was not easy. Two positions presented dilemmas.
The first opener selected himself as he did against Australia. Jack Hobbs, with nearly 1000 runs at an excellent RpAI of 60.62, brooks no argument. However his partner was a real headache for me. The candidates were David Warner (very high rating points but only 543 runs) and Andrew Strauss (926 runs at 48.59). For quite some time I seriously considered having Brian Lara open. Finally I plumped for Warner. I did not want to incur the wrath of the South African followers by foisting Strauss on them when Warner was available. His numbers are also outstanding. And to see Hobbs and Warner (not "Plum" but "David") walk out at Kingsmead would have been something. Each could have learnt something from the other.
Hammond, with nearly 1500 runs, is an automatic choice for the No. 3 position. South Africa is not the place where high averages were obtained and most of the top batsmen were at RpAI figures below 50. I thought long and hard and finally settled on Lara, Sachin Tendulkar and Neil Harvey, all within the top 10. Let me say that these are my personal selections. Any one of those who missed out could be selected. Although I must say that Barrington, Ponting and Fleming sounds much less exciting that Harvey, Tendulkar and Lara.
Gilchrist gets the wicketkeeper position quite comfortably. He and Russell are comparable on the wicketkeeping front, but Gilchrist wins hand down in the batting stakes. In six Tests, he dismissed 26 batsmen and scored 523 runs at 52.30.
The single series of four Tests in which Sydney Barnes captured 49 wickets is enough to pencil his name first within seconds of starting the bowler selection. Warne is second; his good haul of 61 wickets at a sub-25 average lets him get in quite comfortably. Clarrie Grimmett's numbers are very good: 44 wickets at 14.59 cannot be ignored despite the fact that he was also a legspinner. It really does not matter. Two great legspinners in action would be a sight to behold. The last position left me in a dilemma. Javagal Srinath and Mitchell Johnson had almost identical bowling numbers, including the average, varying only in the second decimal point. However, Johnson's much better batting skills got him the fourth bowling spot. The allrounder, of course, is Hammond.
The final team to tour South Africa:
Hammond is the designated captain. One of England's best captains ever, he would do an excellent job of leading this collection of greats.
Part two will follow in two weeks' time. I will cover the all-time best XIs to India, New Zealand, West Indies, Pakistan and Sri Lanka in that article. I will also provide a summary of the selections.
Please do send in your comments on the methodology and these three teams. However, you might be tempted to send in your selections for the remaining five teams. Per se, I have no problem with that. But I will not provide any response and will just publish your comment, as it was sent. It may be a better idea to post such comments after part two is published.
Anantha Narayanan has written for ESPNcricinfo and CastrolCricket and worked with a number of companies on their cricket performance ratings-related systems