Who was the greater opener? © Getty Images
Who was the greater opener? © Getty Images

Virender Sehwag’s retirement sees the departure of a great, great opening batsman who revolutionised the game. In spite of the freakish strike rate of 82 over eight and a half thousand runs, a near 50 average and boasting the three top scores by an Indian batsman in Test cricket, sceptics still raise their stubborn eyebrows when he is compared to the other great Indian opener Sunil Gavaskar. Arunabha Sengupta and Abhishek Mukherjee try to look at the two batsmen through a statistical lens. The results are revealing.

With the retirement of Virender Sehwag, Indian cricket has bid adieu to ‘the other great opening batsman’. And what an opening batsman he was. And what a tale of contrasting styles!

While Sunil Gavaskar batted following the sound principles of giving the first hour to the bowler, Sehwag would consider it a prodigious waste to do the same for even one delivery. It is of little wonder then that while Gavaskar compiled his runs at a steady 44 runs or so per hundred balls (estimated), Sehwag blitzkrieged his way to a strike rate of 82. What remains a freakish feat on the part of the latter is that his average of 49.34 is within a couple of runs of Gavaskar’s 51.12.

They were batsmen as different as can be, with the approach and principle of each enough to make the other squirm. The same goes for the fan following of the two.

And yet, in spite of seriously encroaching blasphemy and sacrilege, one cannot really expect the two to rest on their feats in the record books without attempts at comparison.

Here is an attempt to do the same.

Disclaimers

There are some disclaimers.

There are some who voice that cricket is an art and cannot be effectively measured. Of course one can appreciate the game as an art, but the details and results remain dependent on runs scored and wickets taken and that is how wins and losses are registered, unfortunately not as artistic appreciation of a group of judges as in rhythmic gymnastics or synchronised swimming. (Even in those sports based on evaluation of the judges, there are strict parameters for scoring and aesthetics is just one of the aspects) There are runs scored, wickets taken and the figures left in the wake of the matches provide plenty of material to evaluate the game.

The claim that eras cannot be compared is also a rather airy and unsupported one. Economists and professional statisticians have been comparing indices of different eras as part of their job for ages, with accurate results. Cricket is not the only sphere of life that has changed drastically. In spite of technological remodelling of the world, comparisons can take place. However, the methods have to be somewhat more sophisticated than comparing crude averages.

We cannot say whether the great of one era could have succeeded in another or not. One cannot say whether Sehwag would have been successful in the 1970s against Lillee without protective gear, but it makes no sense to assume that he would not have been— especially if he had geared to play in those days. Similarly, there can be vociferous arguments about Gavaskar’s inability to turn a match around with a double hundred in two sessions. We have not seen him bat with modern-day monster willows. He could have been a much faster scorer, he might have remained the sedate accumulator. It is useless to speculate. But that does not mean we cannot measure them against one another. We can definitely evaluate whether one great in his era was more successful than another in the latter’s corresponding era.

Finally, we will approach analysis of cricket as a science. There were calypsos dedicated to Gavaskar, and YouTube and Facebook tributes in the honour of Sehwag. Those cannot be parameters for objective evaluation. All that adds up to the lore of cricket, not necessarily the correct picture. One is free to choose between lore, facts or both.

All those who would still maintain eras cannot be compared, numbers don’t give the proper picture, one of these batsmen cannot be measured but only experienced and understood through some esoteric mechanism and so on … you can either keep an open mind and proceed, or gracefully remove the page from your presence.

The Method

In this article we are looking at a definite aspect of how the two performed.

There is the general belief that Gavaskar batted against more challenging bowling attacks and performed splendidly against them, whereas the corresponding perception is that Sehwag made merry against some of the weaker sides. We will try to evaluate the facts and figures and try to see how they correlate with these beliefs.

There are several problems we face when we try to evaluate how a batsman has performed against quality attacks. The most obvious — and fallacious — way is to consider performance against different countries. It has obvious disadvantages as we discuss in the next paragraph.

One of the lasting myths about Gavaskar’s career, for example, is his supposed excellence against the celebrated West Indian pace of that era. Indeed, he has 13 hundreds and averages 65 against the Caribbean attacks. The problem is that sort of clustering hides important layers. His 774 runs at 154.80 in his debut series was scored against an West Indian attack with only one regular world-class bowler — a 36-year-old Garry Sobers. The others, Uton Dowe, Grayson Shillingford et al, were less than ordinary, little more than embarrassing footnotes in the Caribbean cricket history.

The fables of Wes Hall, Charlie Griffith, Roy Gilchrist and later Andy Roberts, Michael Holding, Joel Garner and Colin Croft led even the most analytical among us to extrapolate that the 1971 West Indies must have been a fire-breathing pace attack, while in reality they were an insipid side in the midst of a barren period of nine years without a series win.

Similarly, the other hugely successful 1978-79 series against West Indies was in the midst of the Packer era when Roberts, Holding, Garner and Croft were all away in the World Series Cricket. The only time Gavaskar faced the famous four — Roberts, Holding, Garner, Malcolm Marshall — together, his series average was a dismal 30 (1982-83).

Similarly, Sehwag’s exploits against Australia have come against attacks as diverse as the fantastic Glenn McGrath, Jason Gillespie, Michael Kasprowicz and Shane Warne for his 155 in Chepauk; to the less impressive Brett Lee, Brad Williams, Nathan Bracken and Stuart MacGill for his 195 in Melbourne; to the strikingly more ordinary Mitchell Starc, James Pattinson, Peter Siddle and Nathan Lyon in his poor final series. Hence the choice of country as an indicator of the strength of the opposition bowling can be very misleading as one moves through time and seasons.

One other method is to consider the bowlers in question. But even that can be quite deceptive. For example the Marshall of 1978-79 was a greenhorn rookie with one First-Class match behind him, by no means a serious threat. But by the time he returned to India in 1983-84, he was on his way to becoming a terrifying prospect. Similarly, Kapil Dev or Ian Botham in 1979 were far more terrifying than the same bowlers in 1991.  Bowlers evolve with time, and it is far more important to go by the immediate stature and form rather than being confused by the shadow of a full career.

We have circumvented the problems by isolating each match and evaluating how good the bowlers playing in that Test had been at that particular point of time. To get a measure of how good the bowlers had been, we have relied on the Test match rating of the individual bowlers according to the available and retrofitted ICC Cricket Ratings.

The excellence of ICC Ratings is that the bowlers are rated based on current form at the time of the matches, and they don’t fall prey to the heuristics and biases we carry with certain bowler names.

It tells us that in 1991-92 the Australian attack of Craig McDermott, Bruce Reid and Merv Hughes, all bowling at their respective peaks, was one of the most lethal line-ups. However, the same attack in 1985, with all the bowlers finding their feet at the international level, was distinctly ordinary. Normal perusal of the scoreboard will seldom make this immediately apparent. Form of the bowler during that particular period is as important as the retrospective greatness, actually even more so. The ICC Ratings allow us to gauge that.

Finally, ICC Ratings are consistent through the eras. A rating of 850 and above means a very good or great bowler in great form, from Dennis Lillee in 1977 to Allan Donald in 1999 to Steve Harmison in 2004. Lillee obviously stayed in the 850 region way longer than Harmison did; that makes him a far greater bowler. But facing Harmison at his peak in 2004 could have been comparable to facing Lillee in 1977, and perhaps a more threatening option than negotiating the Australian during the his last days of his career in 1983.  ICC ratings allow us to get the temporal picture of the bowler, rather than being overawed or unimpressed by the name.

Ordinary, cruder methods can overlook this. The details provided by ICC ratings do not allow us to be sway by biases.

In the following exercise we have tried to find out how Gavaskar and Sehwag measured up against attacks of different strengths during their day. We have looked at each Test match and evaluated the quality of the attack from ICC Ratings, and then computed their performance against them.

One other word of clarification.

It is commonly believed that batting became a way easier task in the 2000s (Sehwag’s era) than in the 1970s and 1980s (Gavaskar’s time). It may be true, but only to an extent.

The global average of batsmen during Gavaskar’s days (1971-1987) is 30.48. The same figure during Sehwag’s time (2001-2013) was 32.58. There was indeed a two point rise, but not something that cannot be compared. If the reader so wishes, he can always hike up Gavaskar’s average in the following tables by a factor of 1.07 (32.58/30.48), a difference of less than 1 per cent. But surely it is not recommended to add it by a round figure like 5 or 10 as we tend to hear from mathematically uninitiated commentators every now and then.

Gavaskar played attacks without a contemporary in-form bowler in 37 completed innings, compared to Sehwag’s 11.

Definitions to be used throughout the article:

Ordinary bowler: ICC rating <550 just before the match
Decent bowler: ICC rating 550 — 649 just before the match
Fairly good bowler: ICC rating 650 — 749 just before the match
Very good bowler: ICC rating 750 — 849 just before the match
Great bowler: ICC rating 850+ just before the match

Against ordinary bowling sides (all bowlers with rating below 550)

Our first objective is to look at the comparative extent to which the two batsmen made merry against weak bowling attacks (remember, ‘weak’ is defined as on the date of the Test based on immediate ratings).

As cut off for this, we have taken is bowling rating less than 550. We define a weak bowling side if there were no bowlers of the rating 550 or above in the opposition team.

Examples of such attacks:

Ashantha de Mel, Ravi Ratnayeke, Ajit de Silva, Somachandra de Silva, Chennai 1982; Richard Ellison, Norman Cowans, Phil Edmonds, Pat Pocock, Mumbai 1984; Matthew Hoggard, Andrew Flintoff, Craig White, Ashley Giles, Bangalore 2001; Dhammika Prasad, Dilhara Fernando, Angelo Mathews, Suraj Randiv, Ajantha Mendis, in 2010.

As one can see, in the case of the third group, some of the names could have been much more lethal in 2004. But, in 2001, none of them were top-ranked bowlers.

Contrary to popular belief, Gavaskar got more opportunities against such ordinary attacks. He played attacks without a contemporary in-form bowler in 37 completed innings, compared to Sehwag’s 11. And he did exploit these attacks and score heavily to an extent. Sehwag, however, could not quite capitalise and cash in.

Against ordinary attacks (ICC Rating less than 550)

Batsman

Runs

Ave

Gavaskar

1,810

48.92

Sehwag

415

37.72

Against fairly good bowling attacks (at least one bowler between 550 and 649, none above 650)

We next turn to teams who had one or more bowlers of decent ability and form (between ratings 550 and 649) but no bowler who had a rating of 650 or more. (Again, the onus should be not on the names but the combination of name and season)

Examples of such attacks:

Vanburn Holder, Grayson Shillingford, Garry Sobers and Jack Noreiga, Port-of-Spain 1971; John Price, Peter Lever, Richard Hutton, Ray Illingworth, Old Trafford 1971; Richard Collinge, Dayle Hadlee, Hedley Howarth, Bevan Congdon, David O’Sullivan, Eden Park 1976; Tahir Naqqash, Azeem Hafeez, Mudassar Nazar, Mohammad Nazir, Jullundur 1983; Matthew Hoggard, Andrew Flintoff, Simon Jones, Ashley Giles Lord’s 2002; Jerome Taylor, Pedro Collins, Corey Collymore, Dwayne Bravo, Basseterre 2006.

Gavaskar maximises big time on these attacks scoring 2454 runs at 72.18. Sehwag also does well, but not to that extent. His 2044 runs come at 51.10

Against attacks with fairly good bowlers (ICC Rating between 550 and 649)

Batsman

Runs

Ave

Gavaskar

2,454

72.18

Sehwag

2,044

51.10

Against bowling attacks with good bowlers (at least one bowler between 650 and 749, none above 750):

We next turn to teams who had one or more bowlers of good ability and form (between ratings 650 and 749) but no bowler who had a rating of 750 or more.

Examples of such attacks:

Keith Boyce, Shillingford, Sobers, Lance Gibbs, Georgetown 1971; Rodney Hogg, Geoff Hurst, Jim Higgs, George Dymock, Madras 1979; Hoggard, Harmison, Dominic Cork, Flintoff, Trent Bridge 2002; Chris Martin, Kyle Mills, Daniel Vettori, Iain O’Brien, James Franklin, Hamilton 2009.

Gavaskar is up there again, with 2068 runs at 49.24. Sehwag trails at 1404 runs at 43.87.

More than the difference in averages, it makes sense to note that Gavaskar got far more opportunities to bat against less than potent attacks than Sehwag: quite a massive blow to perceptions.

Against attacks with good bowlers (ICC Rating between 650 and 749)

Batsman

Runs

Ave

Gavaskar

2,068

49.24

Sehwag

1,404

43.87

Against very good bowling attacks (at least one bowler between 750 and 849, none above 850):

We now look at teams who had one or more bowlers of very good ability and form (rating between 750 and 849) although no bowler who had a rating of 850 or more. These were decidedly more difficult to play.

Examples of such attacks:

John Snow, John Price, Richard Hutton, Ray Illingworth, Norman Gifford, Lord’s 1971; Dennis Lillee, Len Pascoe, Bruce Yardley, Jim Higgs, SCG 1981; Andy Roberts, Michael Holding, Malcolm Marshall, Winston Davis, St John’s 1983; McGrath, Gillespie, Kasprowicz, Warne, Chennai 2004; Dale Steyn, Shaun Pollock, Makhaya Ntini, Andre Nel, Jacques Kallis, New Wanderers 2006; Jimmy Anderson, Stuart Broad, Tim Bresnan, Graeme Swann, Edgbaston 2011.

This is where Sehwag scores surprisingly better than Gavaskar. The latter’s average takes a beating as the quality of attack becomes better.

This should put to rest all the myth about Sehwag piling runs in easy matches and Gavaskar taming extraordinarily difficult bowling attacks. The numbers, looked at in detail, prove quite the opposite.

Against attacks with very good bowlers (ICC Rating between 750 and 849)

Batsman Runs Ave
Gavaskar 2,275 42.92
Sehwag 2,700 48.21

Against great bowlers at the peak of their form (at least one bowler above 850):

Very few bowlers achieve the mark of 850 and above. If some do, they are truly great or are very good ones in the midst of a dream spell when they are virtually unplayable.

We finally look at teams with such bowlers in their midst.

Examples of such attacks:

Snow, Price, Illingworth, Derek Underwood, The Oval 1971; Holding, Marshall, Roberts, Garner, Port-of-Spain 1983; Imran Khan, Wasim Akram, Abdul Qadir, Tauseef Ahmed, Madras 1986-87; Pollock, Ntini, Lance Klusener, Nantie Hayward, Kallis, Bloemfontein 2001;  McGrath, Lee, Shane Watson, Warne, Stuart MacGill, SCG 2005; Steyn, Morne Morkel, Ntini, Paul Harris, Kallis, Chennai 2008.

Once again, two things are evident from the results. Counting scientifically without relying on legends, Sehwag actually played more knocks against the most difficult attacks, albeit marginally.

Sehwag played 21 Tests against teams with at least one bowler of a rating over 850, Gavaskar played 20. This again is drastically different from the age-old perceptions.

Moreover, while Gavaskar performed adequately enough in these Tests, Sehwag was almost unparalleled against these great attacks.

Against attacks with very bowlers of supreme ability and form (ICC Rating 850 or higher)

Batsman

Runs

Ave

Gavaskar

1,515

47.34

Sehwag

2,023

57.80

At least according to this study, Sehwag performed better when pitted against greater bowlers.

Let us, for now delve into the percentage runs scored by each batsman plotted against the quality of bowling attacks:

20151020GavaskarSehwag_Pi
The first myth is busted: Sehwag had definitely scored more runs than Gavaskar against quality bowling attacks. He has also faced them more frequently.

The final two categories deserve a closer look. Obviously it is not the same thing to score runs against a bowling side with one great bowler and the rest ordinary ones, as opposed to another side with one great bowler supported by multiple very good ones.

Hence, for greater insight, we have:

Against attacks with very good bowlers (of rating between 750 and 849)

Opposition Gavaskar Sehwag

Runs

Ave

Runs

Ave

One very good bowler
Without support

1,090

40.37

1189

66.06

With one supporting good bowler

613

34.06

648

40.50

With two supporting good bowlers

20

10.00

185

23.12

Two very good bowlers

552

92.00

559

55.90

Three very good bowlers*

119

29.75

Three very good bowlers (ICC rating 750-849) in the same side is a rarity. Sehwag came against two such attacks — against Australia in 2004 at Bangalore and Nagpur, with McGrath, Gillespie and Warne all at the peak of their forms.

Against attacks with bowlers in supreme form (of rating over 850)

Opposition Gavaskar Sehwag

Runs

Ave

Runs

Ave

One very good bowler
Without support

884

49.11

342

85.50

With one / two supporting good bowlers

159

31.80

1041

69.40

With one supreme and two very good bowlers bowling together

472

52.44

640

40.00

There are some categories in these tough scenarios where Gavaskar scores better, but overall Sehwag is consistently more successful against better attacks.

In Conclusion

Two openers side by side

Opposition Gavaskar Sehwag

Runs

Ave

Runs

Ave

Ordinary attacks

1,810

48.92

415

37.73

Attacks with decent bowlers

2,454

72.18

2,044

51.10

Attacks with fairly good bowlers

2,068

49.24

1,404

43.87

Attacks with bowlers of very good ability and form

2,275

42.92

2,700

48.21

Attacks with bowlers in great form

1,515

47.34

2,023

57.80

Total

10,122

51.12

8,586

49.34

*The definition of ordinary, decent, fairly good, very good, and great bowlers are given in the respective sections above, but here is a recap:
Ordinary bowler: ICC rating <550 just before the match
Decent bowler: ICC rating 550 — 649 just before the match
Fairly good bowler: ICC rating 650 — 749 just before the match
Very good bowler: ICC rating 750 — 849 just before the match
Great bowler: ICC rating 850+ just before the match

Here comes the shocker: the better the bowling attack became, the more prolific Sehwag became.
Here comes the shocker: the better the bowling attack became, the more prolific Sehwag became.

Gavaskar had been outstanding against weaker attacks, and with a 40+ average, he was certainly good against the toughest men, though not outstanding. As for Sehwag, well, an average of 57.80 against the best bowling attack tells the story. That, possibly, is one of the biggest myth-busters of all.

From the tables above it can be seen that Gavaskar was a superb batsman against decent and non-penetrative attacks, while he was a good batsman against very good to great bowling. Sehwag, on the other hand, was not really a great batsman against the weakest attacks (but then, he was Sehwag), but was supreme against the strongest attacks.

But what about overseas performances? It is generally accepted that batting overseas, at the bowlers’ den, is a tougher proposition than facing them at home. It is also unfair to compare a performance by anyone against, say, Daryl Tuffey in India and in New Zealand.

Let us keep the above conditions intact, and concentrate only on the overseas numbers.

Two openers side by side, overseas

Opposition

Gavaskar

Sehwag

Runs Ave Runs Ave
Ordinary attacks

751

62.58

248

35.42

Attacks with decent bowlers

1,396

66.47

1,086

47.21

Attacks with fairly good bowlers

705

58.75

330

27.50

Attacks with bowlers of very good ability and form

1,297

40.53

1,193

44.19

Attacks with bowlers in great form

906

45.30

1,073

56.47

Let us further explain this with a graph.

20151020GavaskarSehwag_OverseasColumn

Yes, it is counter-intuitive, but figures cannot really be disputed. Free from pink tints, the world can look a lot different.

Again, we cannot say that Sehwag would have been a champion batsman if he had faced Marshall and Holding with just a skull cap. Neither can we say he would not have been one. We cannot say that Gavaskar, with better bats, could have scored two triple-hundreds at a run-a-ball. We perhaps also cannot say for sure that he would not have.

But the figures do tell us that it is a myth that Sehwag scored his runs against weaker attacks.

The figures also do tell us that it is not really blasphemous to at least ask the question: “who was a better batsman?” Neither is the question rhetorical.

(Arunabha Sengupta is a cricket historian and Chief Cricket Writer at CricketCountry. He writes about the history of cricket, with occasional statistical pieces and reflections on the modern game. He is also the author of four novels, the most recent being Sherlock Holmes and the Birth of The Ashes. He tweets here.)

(Abhishek Mukherjee is the Chief Editor at CricketCountry and CricLife. He blogs here and can be followed on Twitter here.)