Top, from left: Bobby Simpson, Eddie Barlow, Aubrey Faulkner, Jacques Kallis, Andy Flower Bottom, from left: Garry Sobers, Imran Khan, Keith Miller, Shakib Al Hasan, Ravichandran Ashwin, Richard Hadlee    Getty Images
Top, from left: Bobby Simpson, Eddie Barlow, Aubrey Faulkner, Jacques Kallis, Andy Flower
Bottom, from left: Garry Sobers, Imran Khan, Keith Miller, Shakib Al Hasan, Ravichandran Ashwin, Richard Hadlee Getty Images

Most XIs, real or fantasy, consist mostly of specialist batsmen and bowlers with one or two all-rounders and a wicketkeeper-batsman if they are available. What about a team consisting entirely of all-rounders? Michael Jones picks one.

It is said that having a top-class all-rounder on your team is like having two players in one.What, then would having 11 all-rounders (classifying a wicketkeeper-batsman as such) be like? A team that effectively has 11 batsmen and 10 bowlers. No let-up for the opposition if they dismiss the top order cheaply, since every member of the team can wield the willow. No hope of keeping out the best bowlers and trying to pick up runs off the rest, since there are plenty more bowlers to call on even if one or two have an off day or get injured. With perhaps six or seven seamers and three or four spinners, the attack would be able to exploit any conditions.

Before selecting the team, it is necessary to decide the eligibility criteria: what is an all-rounder? The traditional definition has been a player who would be good enough to make the XI as a specialist batsman if he did not bowl, and vice versa but this sets the bar very high; only a handful of players in Test history can genuinely claim to have met that requirement.

On the other hand, the mantle has at times been bestowed on a player who could bat a bit and bowl a bit, but do neither of them well enough to make the team as a specialist. In an era when the best batsmen average 50 or more and the best bowlers in the low- to mid-20s, a player who averages in the low 30s with bat and high 30s with the ball cannot reasonably be considered a Test-class all-rounder.

Since most players are stronger in one discipline than the other batting or bowling all-rounders we can set an intermediate requirement: an all-rounder is a player who would be good enough to play as a specialist in his stronger suit, while also making regular contributions in his weaker one. The double of 1,000 runs and 100 wickets has long since ceased to be any indication of an all-rounder: in the modern era, any player who holds down his place in the team for 10-12 years can expect to play 100 Tests, and it takes an exceptionally poor batsman to fail to score 1,000 runs in the process. When Waqar Younis and James Anderson have done the double , it is probably a sign that some other yardstick is needed to measure an all-rounder.

Having a batting average greater than one s bowling average is a good start, but not in itself a sufficient requirement Don Bradman s Test bowling average of 36 was comfortably below the famous 99.94 batting average, but one cannot class him as an all-rounder since he only took two wickets. There needs to be some minimum qualification for a player s bowling to be counted as regular enough to class him as an all-rounder. A fixed minimum number of wickets would benefit players with long careers a number of part-time bowlers have taken 50 or even 100 wickets simply by playing long enough; better bowlers have taken fewer wickets due to playing in eras when Tests were less frequent. Let us instead specify that the player should have taken an average of at least one wicket per Test played; that seems a reasonable indicator of a regular bowler.

In a pantheon of all-rounders, four players who, for differing reasons, never played Test cricket, nonetheless deserve a mention: Alfred Mynn, because his career predated it; Bart King, because his country did not play it; Clive Rice, who had the misfortune that his prime years coincided with South Africa s ban from international cricket, and at 42 was well past his peak when he finally had a chance to play three ODIs; and Franklyn Stephenson, who was ostracised after touring South Africa during the same period. There is also Rice s contemporary Mike Procter, who was devastating with the ball and scored some useful runs in Tests but only played seven of them.

All five have a string of superlative achievements to their name at First-Class level, from King ripping through a batting line-up which included CB Fry and KS Ranjitsinhji to Stephenson making a hundred in each innings and taking 11 wickets in the same match, and Procter s joint-record six consecutive centuries. Two of them also excelled in single-wicket contests; Mynn remained unbeaten in what was a popular format of the game in the first half of the 19th century, but after his back-to-back victories over Nicholas Felix in 1846 it went out of fashion for over a hundred years. In the 1980s, the presence of several great all-rounders in the game simultaneously made the idea of watching them go head-to-head in single combat appealing again Rice took on Imran Khan, Kapil Dev, Richard Hadlee, Ian Botham and Malcolm Marshall, and won three of the four competitions held (Imran won the fourth).

One can speculate as to what any of the five might have achieved if they had had the chance of a full length Test career, but one cannot meaningfully compare actual achievements at Test level to speculated ones. While acknowledging the deeds of this quintet, we cannot consider them for an all time XI. Candidates will be assessed on their Test achievements alone.

Some others, while accomplished all-rounders at First-Class level, had less success in one department in Tests. While WG Grace was a titan with both bat and ball in domestic matches and replicated his dominance with the bat at international level, he bowled infrequently in Tests and only took two wickets. He cannot be considered a Test all-rounder.

Wally Hammond, a successor of Grace at Gloucestershire, took 732 First-Class wickets with a best innings haul of 9 for 23; he had his moments as a Test bowler at Old Trafford in 1934, he dismissed Bill Ponsford, Stan McCabe and Don Bradman, the only bowler ever to claim that illustrious trio in the same Test innings but with 83 wickets in 85 Tests, falls narrowly short of the one-per-Test cut-off to be considered a regular bowler.

Which players, then, would earn places in such a team? We may note that comparing all-rounders is, in one respect, easier than comparing specialist batsmen or specialist bowlers. If batsman A averaged 50 and batsman B averaged 40, or bowler A averaged 20 and bowler B 30, that does not necessarily imply that A was better than B in both cases; one must consider the era in which they played.

A batting average of 40 over a period when pitches were usually difficult and scores correspondingly low may be a greater achievement than one of 50 at a time when batsmen s lives were easier; similarly, a bowling average of 30 on flat pitches may indicate a better bowler than 20 on helpful ones. In the case of all-rounders, the two should cancel out: one who plays in a low-scoring era would be expected to have a lower batting average, but also a lower bowling average.

Suppose all-rounder A averaged 30 with bat and 20 with ball, all-rounder B 45 and 30 respectively. Both of them scored, on average, 1.5 times as many runs per dismissal as they conceded per wicket taken, so we may conclude that both were equally good all-rounders, but B played in a higher-scoring era than A. A batting average greater than one s bowling average (given reasonable minimum numbers of runs and wickets) is generally accepted as the hallmark of a Test-quality all-rounder, and the greater the difference (or ratio) between the two, the better. Of course, just as with batsmen and bowlers, one number by itself does not tell the whole story; one must examine whether a player boosted his averages with cheap runs and wickets against weak teams while failing against the best, if he was a home track bully and achieved little outside his own country, and other such factors.

Openers: Bobby Simpson, Eddie Barlow

There is not much choice here, since few genuine all-rounders have regularly opened the batting.

The likes of Mudassar Nazar, Virender Sehwag and Chris Gayle took a few wickets while also going in first, but we cannot seriously term them all-rounders.

Sanath Jayasuriya is a more promising option, but with 98 wickets in 110 Tests he still fails the regular-bowler criterion of a wicket per match (he had more success with the ball in ODIs, but we are not concerned with the shorter formats here).

There are a handful of batsmen who took over 100 Test wickets while opening at least some of the time but the first two on the list, Vinoo Mankad (162) and Ravi Shastri (151) both fail the batting-average-higher-than-bowling criterion, as does Manoj Prabhakar (96).

Wilfred Rhodes (127) and Trevor Goddard (123) are candidates. A few others also meet the regular-bowler measure: Shane Watson (75 wickets in 59 matches), Bobby Simpson (71 in 62), Mohammad Hafeez (52 in 50), and Eddie Barlow (40 in 30). Shahid Afridi, while theoretically eligible, can be safely turned away from a Test team and advised to concentrate on the shorter formats. That gives us a shortlist.

Goddard, with 123 wickets in 41 matches at an average of 26, can stake a claim he can perhaps be rated the best bowler among players who regularly opened the batting. However, his batting average of 34, with only one century, is not what one might hope for from an opener.

Rhodes is similar: although his batting improved over the course of his career, and he averaged significantly more when he opened than when he didn t, it is not really enough to be considered as opener in an all-time XI; either he or Goddard may come into contention for a place further down the order.

Simpson was an opener of the highest calibre, with an average of 55 when he went in first, and runs against all opponents and in all conditions: over 48 runs per dismissal in every country in which he opened the batting. His bowling average was the wrong side of 40, but given that he may only be the eighth or ninth bowler in this team, perhaps we can allow that in view of his exceptional batting. He was also an excellent slip fielder.

Watson s record with the bat was not that of a world class opener: he failed to average 40 in any country he played in, including his own. With ball he holds the unusual distinction that his home and away averages are both worse than his overall one: the final figure is improved somewhat by 11 wickets at 10.63 in the two matches he played at neutral venues, against Pakistan in England in 2010 when he picked up two five-wicket hauls; he only managed one in the other 57 matches of his career. While he was a handy player to have in a team, he is not going to make it into an all-time XI.

The bespectacled, schoolboy-like Barlow has one of the best records as opener among all-rounders. He took a particular liking to Australia in South Africa s visit there in 1963-64 he hit one double and two single centuries; and finished with an average of 75; at home in 1969-70, in his team s last series before isolation, there came two more centuries and an average of 51. While apartheid restricted the range of opposition available to him, he averaged over 40 against all three countries he did play. His record with ball formed a curious contrast: 33 wickets at 24.57 against Australia including 15 at 21.60 at home in 1966-67 and 11 at 23.36 three years later but 5 at 64.00 against New Zealand and only 2 at 115.50 against England.

These being the opening positions, I will pick the best batsmen among the available candidates over others who may be better bowlers: Simpson and Barlow. The former may not fit many people s idea of what an all-rounder is, but having set our criteria, we must consider any player who meets them.

The obvious choices: Garry Sobers, Jacques Kallis, Keith Miller, Imran Khan, Aubrey Faulkner

Aside from the openers, let us consider who may comprise the rest of the team, before deciding the batting order: since the objective of this exercise is to pick a team every one of whose members is accomplished in the use of both implements, some will naturally end up lower in the batting order than they ever did in their careers. No member of this XI would ever bat at 11 in a normal team, but one of them is going to have to do so when in such illustrious company.

Some of the selections are obvious. During his career and for some decades after it, Garry Sobers was almost universally acclaimed as the greatest all-rounder the game had seen. Then the achievements of Jacques Kallis led some to argue that he was even better. We need not concern ourselves here with arguments as to which of them was greater, since both clearly walk into an XI of all-rounders. Batting averages in the high 50s would make them strong candidates for many all-time XIs on that criterion alone; add more than 200 wickets and 100 catches for each. Kallis has a unique double triple of 2,000 runs, 200 wickets and 200 catches; only four others, including Sobers, have even a single triple of 1,000-100-100) and there can be no doubt.

Despite coming from different countries and eras, two other all-rounders stand out for the remarkable similarity in their averages: 36.97 to 37.69 batting, 22.97 to 22.81 bowling. They are Keith Miller and Imran Khan: a formidable pair of bowlers (although in this team there is likely to be competition for the opening slots), and highly capable batsmen, with 7 and 6 Test centuries respectively. Their places, too, are automatic.

Sobers, Kallis, Miller and Imran are the only players to achieve the double of 1,000 runs and 100 wickets while maintaining a batting average at least 1.5 times their bowling average, but there is a fifth player who kept the same ratio between averages while falling narrowly short on the tally of wickets. Aubrey Faulkner only played 25 Tests, but that was a consequence only of the infrequency with which South Africa played them not being considered serious opposition, until they forced England to sit up and take notice by beating them 4-1 in 1906-07 along with the interruption caused by the First World War.

At the outbreak of War, Faulkner was the tenth-highest run-scorer in Tests; of the top 10 only Jack Hobbs had a higher average and let us remember that the 10 included Victor Trumper and Clem Hill. Any player who can hold his own in such illustrious company as a batsman, while also making significant contributions with the ball, must be an all-rounder of the highest class. He only played one Test between Wars, but in 1921 showed that he was still more than capable of succeeding against the best in the world, with a second-innings 153 to go with figures of 4 for 50 and 2 for 13, playing the lead role in the victory of Archie MacLaren s amateur team over the previously unbeaten Australian tourists. His life off the field was a troubled one, however, and ended in him committing suicide at the age of 48 by gassing himself in a sealed room. Aside from Sobers mixture of spin and seam, Faulkner has a valid claim to being the best spinning all-rounder in the game s history. His place in the team is guaranteed.

Wicketkeeper: Andy Flower

Given the greater emphasis which has been placed on the keeper s batting skills in the last 20 years or so, the position of wicketkeeper-batsman is almost certainly going to be filled by a player from the modern era. If we discount the claims of Clyde Walcott and Hanif Mohammad on the grounds that they were only part-time keepers and decide that the brevity of Denis Lindsay s career should exclude him from consideration, Les Ames is really the only player from before 1990 who merits a place on the shortlist, and even he is likely to lose out to one of the modern players who have filled the position.

Let us stipulate that in order to qualify, a player must have kept wicket in at least half the Tests he appeared in; that excludes AB de Villiers, who has an exceptional average of 57 when keeping, but has done so in fewer than a quarter of his Tests. Kumar Sangakkara can also be ruled out; his performance with the bat dropped significantly when he was required to keep wickets. Mark Boucher was an exceptional keeper, but his achievements with the bat cannot match up to those of some of his contemporaries. MS Dhoni s Test record is more than respectable, but his true place is in an ODI XI. While Quinton de Kock, Johnny Bairstow and BJ Watling all have excellent records when swapping one set of gloves for the other, it is too early in their careers to be able to assess them against others who have completed theirs.

Adam Gilchrist is certainly a candidate, and many observers consider him the automatic choice as keeper in an all-time XI. However, those who claim that he revolutionised the role of the wicketkeeper, permanently raising the bar for the expectation of his batting ability, overlook the fact that another keeper-batsman who made his debut a few years earlier averaged significantly more when entrusted with the larger gloves (53 to Gilchrist s 47): Andy Flower.

There could hardly be a greater contrast in styles: Flower ground out his runs, Gilchrist blasted them out a difference reflected in the chasm between their strike rates, 45 runs per 100 balls against 82. However, in comparing those two we must consider their vastly different circumstances. Gilchrist played for the best team in the world, coming in at 7 with a stellar batting line-up above him. As often as not, he came to the wicket after the top six had laid the foundation for the innings: if he made a big score it would hammer home the advantage, but if he was out cheaply the team would be on top anyway. When he made his highest Test score of 204* at Johannesburg in 2002, it was after coming in at 293 for 5 (one run more, as it transpired, than South Africa would make in the match): all the top six had got starts, with Matthew Hayden making a century and Mark Waugh fifty. Gilchrist s innings merely inflicted further agony on an already demoralised attack, after Allan Donald had limped off the field with the injury which would end his Test career. Of course, he was also capable of making runs when others had failed: at Hobart in 1999 Australia, chasing 369 to beat Pakistan, had been reduced to 126 for 5. Gilchrist, in only his second Test, counter-attacked in style; his 149* off 163 balls dominated a partnership of 238 with Justin Langer which made victory a formality.

Flower was the best batsman in his team by some distance, and often went to the crease knowing that if he failed, the rest of the team would fold. For him to play like Gilchrist did would have been irresponsible: he could not afford to take such risks. Like Gilchrist, one of his best performances came against South Africa, and it too did not affect the result of the match but for the opposite reason. At Harare in 2001 the visitors, batting first, piled up 600 for 3, rarely looking troubled, before showing some mercy by declaring. Gary Kirsten made 220, Jacques Kallis 157*, Herschelle Gibbs 147; Neil McKenzie, dismissed for a mere 52, must have felt that he had missed out.

When Zimbabwe batted, Flower played a lone hand: 142 out of 286 in the first innings and 199* out of 391 in the second, batting over 14 hours in total and scoring more than half his team s runs in the match deprived of a second-innings double-century when Andre Nel finally ended Douglas Hondo s resistance after almost an hour, in which he had scored 6 of the last wicket partnership of 47. Dion Ebrahim gave him some support with 71 in the first innings, Hamilton Masakadza with 85 in the second, but no other batsman from the home team made more than 30; all they could do was cling on while Flower scored the runs. Allan Border remains the only batsman to score 150 or more in both innings of a Test, but Flower came closer to emulating him than anyone except Tillakaratne Dilshan (162 and 143) and Dilshan s innings came against Bangladesh. Flower s two knocks against South Africa stand as one of the greatest batting performances of this century, yet the most they achieved was avoiding an innings defeat; the visitors had no problems knocking off the target of 79 to win by 9 wickets. It was a similar story throughout Flower s career: the degree to which his team depended on him prevented him from taking risks. For him to attempt to bat like Gilchrist did would have been irresponsible.

If we temporarily disregard the very different how of the two and concentrate on the how many , the most obvious contrast is in their performances in India: Flower excelled there (average 117), Gilchrist flopped (29). Both flourished in New Zealand and West Indies, although it is worth noting that when Flower made 113* at Port-of-Spain in 2000 (again, in a match which his team still lost: chasing 99 to win, they were all out for 63 the first time since the birth of the Ashes in 1882 that a team had successfully defended a sub-100 target in the fourth innings), it was against Curtly Ambrose and Courtney Walsh. Gilchrist s only series in the Caribbean came after that deadly duo had retired, and he faced the less threatening Merv Dillon, Pedro Collins and Jermaine Lawson.

The infrequency with which some countries agreed to host Zimbabwe limited Flower s opportunities: he only played two Tests in England, one in South Africa and none in Australia. Gilchrist excelled in England in 2001 (average 68) before being brought down to earth four years later (22); successfully containing him was a key factor in England s series victory. In South Africa, too, he had one brilliant series (473 runs at 158, including that whirlwind double hundred), followed by a dismal one four years later (50 at 10), and he did poorly against the same opposition at home. Flower averaged 71 against South Africa overall, almost entirely due to those 341 runs in two innings at Harare; a couple of scores in the 60s were the best he managed in his other four matches against them.

What of their keeping? Gilchrist s figure of 2.18 dismissals per innings far outstrips Flower s 1.57. This is only natural: Gilchrist kept wicket to Glenn McGrath, Jason Gillespie, Brett Lee, Shane Warne and Stuart MacGill. Contrast that with Zimbabwe s attack of the era: Heath Streak was the only genuinely Test-class bowler Flower kept wickets to, the rest mediocre or worse. They frequently failed to take 20 wickets in a Test, which Australia rarely did. It was natural that Gilchrist would make more dismissals, since his bowlers created far more chances; one can only speculate as to how he might have fared if he only had Streak, Henry Olonga and Paul Strang to create chances, or how many more dismissals Flower could have made with the benefit of better bowlers. Dismissals per innings is thus useless as a way of comparing the two.

Charles Davis has conducted an analysis of dropped catches, primarily covering the period since the start of ball-by-ball online commentary, when they have been more reliably recorded; his data show that over the course of his career Gilchrist dropped an average of 12% of the chances he received among the best keepers over the period covered, with a lower proportion early in his career increasing later.

Flower played most of his international career before the ball-by-ball era, when usually the only source of information on dropped catches is match reports and those are often incomplete. Wisden has habitually given significantly more detailed reports of matches involving England than of those not; in the four Tests Flower played against England, it records that he dropped one catch, at Bulawayo in 1997 (it didn t matter too much, since the batsman reprieved was Phil Tufnell). He took nine catches in the same four matches, giving him a drop percentage of 10% based on that extremely small sample size; for most of Zimbabwe s other matches in the period Wisden does not mention dropped catches at all (the report of one Test against Pakistan in 1993 mentioned uncharacteristic fielding lapses by Zimbabwe, without naming particular fielders as being guilty of them). The 10% figure appears, then, to be the only one obtainable, but it is subject to such a large margin of error as to make comparisons impossible.

Overall, then, there appears to be little to choose between them, but I will go for Flower on the grounds both of his better overall average and the balance he brings to the team the ability of Sobers, Miller and Imran able to score quick runs contrasts well with the steadier pace of Simpson, Kallis and Flower.

Spinners: Ravichandran Ashwin, Shakib Al Hasan

Before picking the last three positions, let us check the balance of the team so far. There are four seam-only bowlers, two spin-only, plus Sobers who could bowl both and one of the spinners is Simpson, the weakest bowler among the seven. Thus in order to achieve a balance overall, two of the last three should probably be spinners.

There have been far fewer spin-bowling all-rounders than fast-bowling ones, so our choices for the two may be limited and a large number of fast bowlers left to compete for a single final spot in the team, but such is the difficulty of selecting fantasy XIs.

Considering the spin candidates chronologically, we start with Johnny Briggs, one of the best bowlers of his day. His phenomenal Test average is boosted by some mind-boggling returns 21 wickets at an average of 4.80, and no, that is not a typo against the weak pre-1900 South Africa teams which were not considered to be of Test standard at the time, and whose retrospective award of that status is difficult to justify; take those away, however, and he still had 97 at 20.55 against Australia, an impressive return against the best opposition of the time. Briggs did make a Test century against Australia, at a time when that was no mean feat, but even by the generally lower-scoring standards of the late 19th century, a batting average of 18.11 was a fairly poor one.

Moving through the decades, we come to Wilfred Rhodes, whom we rejected as an opening batsman with the caveat that he might still fit in further down the order. His list of records at First-Class level is impressive: most matches, most wickets, 17th on the list of most runs (and the most of anyone who failed to score 100 career centuries). In Tests, his records are for longevity: when he was called up for the second-string team which toured West Indies in 1930 after having played only one Test in the preceding nine years he was both the oldest player ever to appear at international level (52) and recorded the longest career (just short of 31 years).

What of his achievements in that span? He certainly cannot be accused of boosting his figures against weak teams, since 41 of his 58 Tests were against Australia, and his 109 wickets against them came at the more-than-useful average of 24. A record of 10 wickets at 45 against West Indies can be excused on the grounds of age (only one other bowler, Bert Ironmonger, took any Test wickets after his 50th birthday at all), although he also performed poorly with the ball in the 13 matches he played against South Africa, the best part of two decades earlier.

The 1913-14 series was the one in which Syd Barnes took a record 49 wickets, which didn t leave many for anyone else although Rhodes did better with bat in that series, making 152 at Johannesburg to ensure that Barnes had enough runs to bowl at: he took 17 wickets as England won by an innings. Rhodes started his career as a tail-ender, but gradually moved up the order to open, and for a period of five years or so before the war, his record was comparable to those of most specialist batsmen.

Richie Benaud, for all his brilliance with both implements in the 1957-58 South Africa series 329 runs at 54.83, 30 wickets at 21.93 fails to make the cut, with a career bowling average higher than his batting average. There are surprisingly few who do: from Rhodes last Test in 1930 until the turn of the 21st century, only two spinners other than Sobers meet the criteria of a positive difference in averages and at least one wicket per match.

The first is Mushtaq Mohammad. At Dunedin in 1973 he dominated with bat, ball and hands: he scored 201, added 350 with Asif Iqbal, then took two wickets and two catches in the first innings, five wickets and one catch in the second, as New Zealand failed to match his individual score in either; he remains one of only two players to score a double-century and take a five-for in the same Test, Denis Atkinson being the other. New Zealand were rather a weak team at the time, but when Mushtaq did the century-5-for double again four years later this time as captain there was no doubting the calibre of the opposition: at Port-of-Spain he scored 121 against Andy Roberts, Colin Croft and Joel Garner, then dismissed Alvin Kallicharran twice and Viv Richards once on the way to figure of 5 for 28 and 3 for 69. Pakistan levelled the series with a crushing 266-run victory, although they lost the decider at Kingston two weeks later.

That one century aside, Mushtaq s performances with the bat against West Indies were sub-par, although perhaps he can be forgiven that given the quality of the sides he faced; modest records with both bat and ball against Australia were not helped by their reluctance to tour Pakistan, which meant that he only ever faced them away from home. His overall statistics are flattered by the easy pickings offered by New Zealand: take them away and he has batting and bowling averages of 37.19 and 32.65 respectively still useful, although they don t look quite as good as the 39.17 and 29.22 against all teams.

The next candidate in chronological order is Tony Greig, who, like Sobers, bowled both seam and spin. In 1974 West Indies bowling attack was yet to reach the heights it did in the following years Keith Boyce and Bernard Julien were a useful opening pair, backed up by Sobers and Lance Gibbs, but were not in the same league as their successors but their batting line-up was as strong as it would be at any time during their period of domination: a top six of Roy Fredericks, Lawrence Rowe, Alvin Kallicharran, Clive Lloyd, Rohan Kanhai and Sobers. It was against that line-up that Greig took 8 for 86 and 5 for 70 in Port-of-Spain, bowling England to victory by 26 runs and with it a share of the series; it would be 30 years before they next avoided defeat in a series in the Caribbean. Greig finished the series with 24 wickets at an average of 23, and 430 runs at 48. Curiously, he did much better away than at home, with both bat and ball: if his away averages of 46.92 and 28.09 had been career figures they would have bracketed him alongside Sobers, but figures of 34.63 and 37.91 at home looked ordinary in comparison with the bat he averaged more in every away country he visited than he did at home, although he was less consistent with the ball, with little success in Australia or Pakistan. India was a happy hunting ground for him with both implements: two centuries and an average of 52, 21 wickets at 28.

Spin-bowling all-rounders, it appears, are like buses after only four in the first 130 years of Tests (aside from Faulkner and Sobers, who are already in the team), three more have made their debuts in the last decade: Ravichandran Ashwin, Ravindra Jadeja and Shakib Al Hasan. Ashwin and Jadeja have both been deadly at home: England have been the only team ever to get on top of Ashwin in India, pushing his average the wrong side of 50 when they won there in 2012-13. He improved in his second home series against them in 2016-17, although his 28 wickets still came at an average of slightly over 30, short of his usual standards.

Aside from the sole Test he played in the 2012-13 series his debut, which he finished with match figures of 3 for 117 Jadeja has never averaged over 26 in a home series; his worst performance, 26 wickets at 25.84 in the 2016-17 clash, was still pretty good by most people s standards, and overall he averages a shade under 20 at home. In 2015-16 South Africa were caught like rabbits in the twin spinners headlights: Ashwin took 31 wickets at 11.12, Jadeja 23 at 10.82.

Neither has fared so well away from home, although this is hardly unusual for Indian spinners. Ashwin s overall away average of 31 may look respectable, but that is largely a result of success in Sri Lanka and against a weak West Indies team; he has averaged over 50 in two tours of Australia, and failed to take a wicket in his only Test in South Africa to date. Jadeja did rather better in that South Africa match, with six wickets, but has yet to tour Australia, endured a miserable time in his two matches in New Zealand and was little better in England, and in all away Tests he averages just under 37.

Despite Jadeja s three triple-centuries at domestic level, Ashwin has the superior batting record in Tests, and unlike his figures with the ball, his batting average is almost the same home and away albeit with the latter number, like the bowling one, boosted by success in West Indies (the opposition against whom he has scored all four of his Test centuries). Jadeja has yet to make a Test century anywhere, and only has two fifties outside India albeit one of them crucial: his 68 off 57 balls and partnership of 99 with Bhuvneshwar Kumar at Lord s in 2014 enabled his team to set their hosts a target of 319, from which Ishant Sharma bowled India to a famous victory, to which Jadeja applied the finishing touch by running out his old foe Anderson. All things considered, though, Ashwin s record comes out as superior, and his name is added to the list of candidates.

Then there is Shakib the only Bangladeshi who would be in contention for an all-time XI of any sort. Only four players in Test history have scored 100 or more runs and taken 10 or more wickets in the same match, and of those only Shakib has come close to doing it a second time. When he did accomplish the feat it was against Zimbabwe, but there was no disputing the quality of the opposition against which he achieved the near miss: by returning 84 and 5 with the bat, 5 for 68 and 5 for 85 with the ball, he propelled Bangladesh to their first ever Test victory over Australia a prospect which would have been dismissed as unthinkable not-too-long before and when David Warner and Steven Smith put together a partnership of 130 in the fourth innings, threatening to spoil the hosts party, it was Shakib who dismissed them both to swing the game in Bangladesh s favour.

His away record is difficult to assess due to the small sample sizes: most countries only schedule two Tests per series against Bangladesh, leaving him with only 17 away matches in his career to date, and no more than four in any one country. Like Ashwin, his decent overall away average masks significant variation between countries. A strong performance in South Africa in 2008 11 wickets in 2 Tests, including the scalps of Kallis and AB de Villiers twice each and Hashim Amla once; de Villiers had broken the record for most Test innings before making a duck until Shakib finally handed him his first was all the more impressive for the lack of support he received: the rest of the attack took only 8 wickets between them, and Bangladesh lost both matches by an innings. Then there were 8 wickets in the match and an unbeaten 96 in the fourth innings to guide the team to victory at St George s in 2009, their first ever away series win; 6 wickets and 116 at Colombo earlier this year as they beat Sri Lanka for the first time. It s difficult to name a Test victory by Bangladesh in which Shakib hasn t played a leading role it s no wonder he felt that he needed a break, and his team are certainly missing him in South Africa. Curiously and despite their vastly different styles (and heights) his batting and bowling averages are almost identical to Greig s (40.38 and 32.37 to 40.43 and 32.20), from a similar number of Tests (51 to 58) and almost exactly the same number of runs (3,594 to 3,599). Shakib has a comfortable lead in the wickets, though, with 188 to Greig s 141.

We have our shortlist of spinners: Rhodes, Mushtaq, Greig, Ashwin, and Shakib only two of whom can find places in the team. Given that those two will, along with Faulkner, provide the main spin attack, we really want full-time bowlers. That counts out Mushtaq, who only averaged 1.4 wickets per Test where all the others are over 2. With Sobers already in the team, two more left-arm spinners may be overkill: one other left-armer plus one off-spinner would do nicely. Ashwin and Shakib it is, although Greig can consider himself unlucky to miss out.

The last fast bowler: Richard Hadlee

Only one place in the team remains to be filled, which is somewhat unfortunate given the number of candidates there are for it. We have plenty of batting already, so the place goes to the best bowler among those who can be classified as all-rounders. The possibilities are almost endless: Monty Noble, Jack Gregory, Alan Davidson, Ian Botham, Kapil Dev, Wasim Akram, Chris Cairns, Shaun Pollock but one name stands head and shoulders above them all, and that is Richard Hadlee. He completes the XI.

Now we have the names, what about the order? Kallis sometimes batted at 3, but more often at 4. Rather than shift him from his preferred position, there is an alternative: although Faulkner played only 9 Test innings at number 3, they brought him one double- and one single-century, four fifties and an average of 69.

Putting him at first-drop allows Kallis to take up his regular place; Flower was a fixture at 5 and Sobers most often batted at 6. The others will be forced to take up places lower than they might be accustomed to, simply because the team has no tail-enders in the usual meaning of the term, but still has to fill positions 7-11; I will allocate them to Miller, Imran, Shakib, Ashwin and Hadlee in that order.

As for the captaincy, there is plenty of choice: Simpson, Flower, Sobers, Imran and Shakib all captained their countries. Imran gets the nod, partly for the startling improvement in his figures when he was leading: while in the ranks he averaged 25 with both bat and ball, still a decent performance but when in charge those improved to an extraordinary 52 with the bat and 20 with the ball. Leading an unfancied team to a World Cup title is also a feather in his captain s cap.

Thus the final team, in batting order, is:









Bobby Simpson









Eddie Barlow









Aubrey Faulkner









Jacques Kallis









Andy Flower (wk)






Garry Sobers









Imran Khan (c)









Keith Miller









Shakib Al Hasan









Ravichandran Ashwin









Richard Hadlee









It is a global team, with eight nationalities represented: there are three South Africans a fair reflection of that country s long tradition of world-class all-rounders two Australians, and one each from West Indies, Zimbabwe, Pakistan, India, Bangladesh and New Zealand. Only England and Sri Lanka are not represented; from the former, Rhodes, Greig and Botham can consider themselves unlucky to miss out, but Sri Lanka have never really had a truly world-class Test all-rounder. Jayasuriya came close, Aravinda de Silva and Angelo Mathews both had some success with the ball in ODIs but made little impact in Tests, while Chaminda Vaas batting never quite earned the status of all-rounder despite 3,000 Test runs and a sole century close to the end of his career.

Every player in the side has at least two Test centuries, and everyone except the keeper has at least one five-wicket haul (Flower had one haul of five catches). Only one averages under 30 with bat, and only one again, excluding the keeper averages over 35 with ball.

Even if the team loses early wickets, the opposition will never be able to assume that the innings is almost over, as with Hadlee at 11 there is no tail.

When they take to the field, Imran will have to decide which of himself, Miller and Hadlee does not get to take the new ball; one of them may be unhappy about it. If the opposing batsmen succeed in seeing those three off, then Sobers, Kallis and Barlow are the back-up seamers. Faulkner, Ashwin and Shakib offer the full range of spin options: one leg-spinner, one off-spinner and one left-armer and of course Sobers could switch to spin as well. Simpson s less regular spin is unlikely to be required too often, but it is there if needed. With Simpson, Kallis and Sobers in the slip cordon, it is unlikely many chances will go down.

With so many great names failing to get into the first team, there is clearly scope for a 2nd XI. For this one I will relax the criteria, allowing in those who played few or no Tests through no fault of their own, as well as others who clearly qualify as all-rounders based on their accomplishments at First-Class level, even if they miss the one-wicket-per-match cut-off in Tests.

Allowing them in gives: WG Grace, Trevor Goddard, Wally Hammond, Clive Rice, Mike Procter, Tony Greig, Frank Woolley, Adam Gilchrist, Ian Botham, Wilfred Rhodes, Alfred Mynn. WG, Hammond, Gilchrist and Botham and this is a 2nd XI?

What about a 3rd? Although there are still plenty of bowling all-rounders left, we are running rather short of batting all-rounders by now, so most of the top order for this team will be batsmen who bowled a bit, rather than genuine all-rounders. Nevertheless, Virender Sehwag, Sanath Jayasuriya, Greg Chappell, Steve Waugh, Jack Gregory, Mushtaq Mohammad, Les Ames, Kapil Dev, Chris Cairns, Shaun Pollock, Wasim Akram make a more-than-useful 3rd XI.

Like the first two, the batting extends all the way to 11 where we find a player with a highest score of 257* to his name but the bowling does not have quite the same depth. That said, not quite the same depth in this case is relative: there are still five world class fast bowlers, but the three spinners and remaining two seamers are part-timers.

If the celestial fixtures secretary could manage to arrange a tournament between these three teams, it would be one worth watching.