Greg Chappell is caught and bowled by Karsan Ghavri for 76

India's greatest Test win is not Kolkata 2001, okay?

© Getty Images

Stats feature

Which are the greatest comebacks in Test history?

Headingley 1981? Galle 2015? Brisbane 1960? We present a list of 13 for you to mull over

Anantha Narayanan |

As an analyst, I can vouch for the fact that the Test format offers the most nuanced and complex challenge of all. Any Test analysis is multilayered and the most obvious solution need not be the correct one. It turned out to be so in this exercise, which looks at Test cricket's greatest comebacks, with their spectrum of heroic and logic-defying feats.

The details of the methodology are in the sidebar. But here, to get us started, is a summary.

  • The analysis starts at the mid-point of each Test, i.e. right before the third innings commences. The computation of the index values is carried out before the start of the third and fourth innings and at the fall of a wicket in these innings. Thus, there are a maximum of 20 measuring points.
  • A notional target score is estimated for the third innings. This is a complex process since it relates to setting up a potential fourth-innings target. The target for the fourth innings is, of course, known.
  • At each measuring point, a projection is done based on the team score, historic data, and the first two/three innings scores.
  • Using the target and the projected scores, a WinIndex value is determined for the winning team.

Target determination

Third-innings targets

Difference over 200 runs

A follow-on is treated differently to a non-follow-on situation. For eg:

Australia 445 and India 171 means that India batted again after the Australian bowlers had bowled 60 overs. The bowlers were tired and India could exploit this to work towards a fourth-innings target of 250. The third-innings target was 524.

Sri Lanka 600 and India 223 means that the deficit was 377. India could only think of going for a lower target of 200. The third-innings target was 577.

England 284 and South Africa 521 means that England batted again after bowling 195 overs, and the South African bowlers were fresh. Therefore England have been given a fourth-innings target of 200. The third-innings score target was 437.

Differing targets such as 200 and 250 are used in order to differentiate between 245-run and 377-run leads and to allow for the freshness of bowlers.

Difference below 200 runs

There is no difference whether the third batting team has the lead or is in deficit. Sequences of 130/110 or 450/500 should not lead to similar pre-determined targets. The average Runs per Wicket (RpW) for the combined first-innings scores (limited between 10 and 40) is multiplied by 10, 11 or 12 to determine the target.

Differing multiplying factors are used so that varying RpW values are treated differently. At an RpW of 13, the team should aim for something more, in this case, 20%. At 38, they can, at best, aim for the full innings score.

Freak pitches, which led to real-life scorelines of 75/123/475 or 523/598/173, should be ignored.

Fourth-innings targets

After the complexities of the third-innings targets, this is a piece of cake since the targets are set in stone and could range from 1 to 836.

Determination of winning chances

The winning chances are always determined for the batting team since the targets are available only for the batting team. In this article, the WinIndex values pertain to the ultimate winning team. Since each winning team has a batting and bowling innings, the required transformation is done for the bowling innings.

The winning chances are determined ten times during each of the third and fourth innings - at the start of the innings and the at the fall of each wicket - leading to a maximum of 20 values.

Let me define the qualifying Tests. I will consider all decisive Tests that went into the fourth innings, including India v West Indies in Bridgetown 1982-83, which had a single no-ball in the fourth innings. As of October 2017, at the completion of the series-clinching win for Sri Lanka over Pakistan, 1132 Tests qualify. Incidentally, this last Test threatened to push itself into serious consideration until Sarfraz Ahmed decided to sweep from the rough. The South Africa v England Test in Centurion in 1999-00 and the England v Pakistan Test at The Oval in 2006 are not considered, for the obvious reasons of contrived declarations and forfeit.

Using the WinIndex pattern across the 20 measuring points, I made a shortlist of around 25 Tests. Then I scrutinised each match and completed the final selection of 13. I have tried to accommodate all eras of Test cricket and represent most of the top teams. There is a pleasant surprise at the end. In general, I looked for Tests that have a WinIndex pattern of at least 3-17. This was to ensure that only matches in which the winning teams came back from desperate situations are considered. There are some exceptions, which are explained later.

In the history of Test cricket there is just one Test with a WinIndex pattern of 0-20: that means that the winning team had no chance from the beginning of the third innings through the 19 fall-of-wicket instances. This was when unfancied South Africa stole a magnificent one-wicket over England in Johannesburg in 1905-06. Needless to say, this was an automatic entry into the final selection.

There are six Tests with patterns of 1-19, and three of them were selected. Also slotting in was the famous Neil Harvey-inspired win in Durban in 1949-50, which had a pattern of 1-15.

Three Tests were selected outside these set criteria. The one-run win by West Indies over Australia in Adelaide in 1992-93 could not be denied its place. It was poetic justice that the WinIndex equation was a beautiful 10-10.

I did some soul-searching while selecting a Test for India. I am strong in my belief that the greatest Indian come-from-behind win was neither Kolkata 2001 nor The Oval 1971, but Melbourne 1981. Kolkata was 13-5 in India's favour (Australia ahead only at the beginning, and till the fall of the first four wickets of India's second innings). The Oval was 14-3 in India's favour. Melbourne was 4-16 and gets in on merit.

Another selection that required some discretion was in selecting a Sri Lankan win from the abyss. The classic in Galle in 2015 featuring star performances from Dinesh Chandimal and Rangana Herath got the edge over the chase of 326 at the SSC in 1998 that was led by Aravinda de Silva. The Galle match clocks in at 4-16 while the SSC win was at 1-15. The final tilting of the scales here was the strength of the Indian team.

These, then, are the final selections, presented chronologically. The chasing wins are in green graphs and the defending ones in blue. The WinIndex percentage is shown on the Y-Axis and the Innings-Wicket on the X-Axis. The values run from 5 to 95.

1. Australia v England in Sydney, 1894-95
England won by ten runs

© Ananth Narayanan

That was the first-ever Test win after a follow-on. England's notional target as they batted in the third innings was 511(250+261). Albert Ward's hundred was supported by a number of useful innings, but at no stage did England look likely to reach the target. The final target for Australia was 176, somewhat below par. They were coasting at 130 for 2, and the WinIndex values remained below 40 for the first half of the innings. Then came a horrific collapse, from 158 for 5 to 162 for 9. Only at the fall of the ninth wicket did England forge ahead: their only WinIndex value exceeding 50. The WinIndex distribution was 1-19.

2. England v Australia at The Oval, 1902
England won by one wicket

© Ananth Narayanan

Australia batted efficiently for a first-innings score of 324. Hugh Trumble then dismissed the strong English line-up for 183. The lead of 141 meant that Australia's notional third-innings target was 163 (12*25.3-141). They could only score 121, yet the final target of 263 was quite stiff, seeing how England had floundered against Trumble and Jack Saunders, who now ripped through the top order to reduce Australia to 48 for 5. That is when the "Croucher", Gilbert Jessop, walked in to play an innings fit to be ranked among the best chasing innings ever. He added over a hundred with Stanley Jackson and 30 with George Hirst and was dismissed at 187 for 7. The match was in the balance, but clearly heading Australia's way. The ninth wicket fell at 248, and who will ever know whether Wilfred Rhodes told Hirst "Let's get' em in singles" or "Let us have a bash" or "Give 'em boogers, hell" or whatever else. They got those 15 runs and England won a memorable Test, against all odds.

3. South Africa v England in Johannesburg, 1905-06
South Africa won by one wicket

© Ananth Narayanan

This is the only Test match in which the winning team did not have a single WinIndex value exceeding 50: in other words, the only 0-20 in the history of Test cricket. A somewhat depleted English team was dismissed for 184 by a raw South African side. South Africa's inexperience was revealed when they did not even reach half that. The two low scores and the big lead meant that the notional third-innings target was just 72 (12*13.75-93), but England scored an impressive 190. Chasing 284, the Windex value for South Africa was 25% when they slumped to 105 for 6 and Dave Nourse played one of history's great match-winning innings, adding over 100 with Gordon White. At 226 for 7, the WinIndex was 49%. Then South Africa lost two quick wickets and were struggling at 239 for 9, but Nourse calmly added the required 48 runs with Percy Sherwell and won the match.

4. South Africa v Australia in Durban, 1949-50
Australia won by five wickets

© Ananth Narayanan

If I ever create a list of the five best Tests of all time, the Durban classic of 1950 would find a place. Eric Rowan anchored a potentially match-winning total of 311 against a potent Australian bowling attack. Australia floundered against the spin wizardry of Hugh Tayfield and were dismissed for 75, conceding a lead of 236. The fourth-innings target for Australia, 232 (12*19.3), was already contained in the lead, and the actual innings target was a notional one (48). This meant that even though South Africa were dismissed for 99, Australia was firmly stuck at WinIndex values either side of 30% for the entire third innings. The target of 336 felt like a mirage, and as they slipped to 59 for 3 and 95 for 4, the WinIndex value dipped below 30. Then Neil Harvey made 151 not out, a masterpiece fit to be ranked with the 153 not out nearly 50 years later by that other elegant left-hander, Brian Lara. Harvey added over a hundred with Sam Loxton, and then Colin McCool provided mature support as Australia romped home to a completely unexpected win. The WinIndex tally was 1-15.

5. Australia v India in Melbourne, 1980-81
India won by 59 runs

© Ananth Narayanan

I have already expressed my views on this win. India had a poor start but recovered through a masterly 114 by Gundappa Viswanath. However, 237 was a poor return on the first day of a Test and Allan Border's measured 124, assisted by support from Greg Chappell, Doug Walters and Rod Marsh took Australia to a first-innings lead of 182. India's notional target for the third innings was 542 (11*32.8+182). There were useful contributions from the top five but the last five wickets fell for only 64 runs. India's winning chances at the start of the fourth innings were not even 30%: understandable, with Australia's target a measly 143. But Australia lost quick wickets. The first two went to Karsan Ghavri, who bowled just about slow-medium. Dilip Doshi chipped in with two and Kapil Dev polished off the rest. I would not have been surprised if the odds against an India win before the Australian innings were 500-1 (remember that Headingley 1981 was still six months away). The last four WinIndex values were above 50 and the final tally was 4-16. I am firm in my belief that this is the greatest Indian Test win ever.

6. England v Australia at Headingley, 1981
England won by 18 runs

© Ananth Narayanan

This is a match we normally run out of adjectives to describe. I have written about it many times but never do a copy-and-paste since I always try to get a new perspective. Australia posted a formidable total of 401, made at a pedestrian pace. The Australian pace trio was devastating and dismissed England for 174, a deficit of 227. Today, there would have been no follow-on, but the times were different and England were asked to bat again. The same story repeated itself and England were staring at an innings defeat at 135 for 7. The colossus, Ian Botham, was at the crease and Graham Dilley came in. They added 117 runs and avoided an innings defeat. Chris Old had a wise head on his shoulders and helped Botham add another 67. Bob Willis helped Botham add another 37. During the entire third innings, the target of 477 (250+227) was way out of reach and the WinIndex remained firmly south of 30%. What followed was stirring. Willis produced arguably the best bowling performance ever by an English fast bowler, and dismissed Australia for 111. At 75 for 8, England had the edge, as also at 110 for 9. The Botham and Willis show ended with the WinIndex tally at 2-18.

7. Sri Lanka v Australia at the SSC, 1992
Australia won by 16 runs

© Ananth Narayanan

Batting first, slumping to 124 for 7, recovering to 256, but still conceding a lead of nearly 300 runs, Australia's plight at the SSC in 1992 was miserable. Granted, these were the pre-Muralitharan days, but Sri Lanka boasted a good all-round bowling attack at home. Australia had a notional third-innings target of 491 (200+291). No one scored a hundred, but then no one scored below 10, and they built up to 471. The steady batting is reflected by the WinIndex line hovering around the 40+ mark for most of the Australian innings. However, the final target of 181 was below par, considering the high totals in the match. When Sri Lanka were 132 for 3, a win was theirs for the taking. Aravinda de Silva, with a 32-ball-37 under his belt, went for an ambitious hit and Border completed a magnificent boundary catch. That was it. Greg Matthews and Shane Warne ran through the next six batsmen for the addition of only 32 runs. Australia went on to complete one of their most satisfying wins and the legend of Warne was born. Only at 156 for 9 did Australia go past a WinIndex value of 50 and it was 1-19 at the end.

8. New Zealand v Pakistan in Hamilton, 1992-93
Pakistan won by 33 runs

© Ananth Narayanan

New Zealand has always been a tough cricketing country to visit. On the 1992-93 tour, Pakistan had the better side on paper, especially because Martin Crowe was not available. But they started awfully, losing three wickets for 12, and recovered partially to reach 216, through a masterclass of 92 by Javed Miandad. In response, Mark Greatbatch played a consolidating innings of 133 and helped New Zealand to a lead of 48. Pakistan's notional target for the third innings was 336 (12*24+48). They barely reached the halfway mark, and there was a single spark in the form of Inzamam-ul-Haq's 75. Through most of the innings, the WinIndex values for Pakistan remained around the 30 mark. The fourth-innings target of 127 was way below par. However, Wasim Akram and Waqar Younis, while not generally seeing eye to eye off the field, were arguably the most devastating bowling pair in Test history. Wasim bowled unchanged for 22 overs and Waqar 14, and together they dismissed New Zealand for 93. Only at the fall of the eighth wicket, at 88, did the tide really turn for Pakistan, so it was 2-18 in the WinIndex stakes.

9. Australia v West Indies in Adelaide, 1992-93
West Indies* won by one run

© Ananth Narayanan

A few days after the thriller at the SSC, Australia and West Indies played a magnificent Test in Adelaide: just one run away from being the perfect Test match. West Indies were sitting comfortably at 129 for 2, but could not even double that score, finishing on 252, a poor return for Adelaide. Curtly Ambrose was magnificent and ripped through the strong Australian batting, earning a lead of 39 with his 6 for 74. The notional third-innings target for West Indies was 240 (12*23.2-39). They folded for just 146, almost half of those runs scored by Richie Richardson. The WinIndex values moved either side of 50 through this innings. The target of 186 was a fair one. Australia's chase was a mix of the horrible and sublime: 16 for 2, 64 for 4 and 74 for 7 tell the story. Everything seemed over at 102 for 8. Tim May and a patient Warne added a few runs but the ninth wicket fell at 144. May, batting beautifully, had taken the score to 184 when Craig McDermott fell. The ultimate victory margin of a single run was most appropriate. Also appropriate is the WinIndex tally: a perfect 10-10.

10. Australia v South Africa in Sydney, 1993-94
South Africa won by five runs

© Ananth Narayanan

If you ask me which is the greatest couple of years in Test cricket, I will unhesitatingly say 1992 to 1994. Five of my 13 selections are from this period. The New Year Test at the SCG brought grief to the visiting South Africans on the opening day. Warne was almost unplayable and his haul of seven wickets meant that South Africa scored just 169. In response, Australia secured a potentially match-winning lead of 123. The notional target for South Africa in the third innings was 399 (12*22.7+123). They reached 239, thanks to a plucky 76 by Jonty Rhodes. The graph reveals the WinIndex values fluctuating around the 30 mark. The target of 117 foretold an easy win for Australia's strong batting line-up, but Fanie de Villiers had other ideas. He captured the first four wickets and then Allan Donald chipped in. Then, when McDermott seemed to be taking Australia to a memorable win, de Villiers came and picked up the last wicket to leave South Africa winners by five runs. Only at the fall of the eighth wicket did South Africa look like winning, which is reflected in the WinIndex tally of 1-19.

11. Pakistan v Australia in Karachi, 1994-95
Pakistan won by one wicket

© Ananth Narayanan

This is the fifth Test from our golden 26-month period. Solid middle-order batting helped Australia to a useful 337 in the first innings. Their pace-spin combination then combined to secure a lead of 81. Australia's notional target for the third innings was 274 (12*29.7-81). At 171 for 2, they were well on their way to an imposing total, mainly through David Boon's hundred. Then came a massive collapse, engineered by Wasim and Waqar, and Australia folded for 232, leaving Pakistan 314 to get in the last innings. The chase was well placed at 148 for 2, then floundering at 184 for 7, 236 for 8 and 258 for 9. Everything seemed lost for Pakistan when Mushtaq Ahmed walked in to join Inzamam. Slowly but surely the pair scored the 57 runs to win an extraordinary match by a single wicket, Mushtaq's attacking innings of 20 in 30 balls perhaps as responsible for the win as Inzamam's effort. The final WinIndex equation was 3-17.

12. Australia v Pakistan in Hobart, 1999-2000
Australia won by four wickets

© Ananth Narayanan

Another amazing Test match featuring Australia. This one in from the early stages of their 16-Test winning streak. This is also the Test that almost upset the applecart. The two first innings were similar: both below par, and Australia ended with a slender lead of 24. Pakistan's notional third-innings target was 304 (12*23.4+24). They exceeded it comfortably, with 392, through Inzamam's hundred and fifties from Saeed Anwar and Ijaz Ahmed. The target of 369 and the loss of early wickets worsened the situation for Australia. At 126 for 5, everything seemed lost. Australia's WinIndex value dipped below 30. Justin Langer was joined by Adam Gilchrist, playing his second Test. They produced an extraordinary partnership of 238 and took Australia to within a few runs of the target. The Australian WinIndex value exceeded 50 for the second time at the fall of the sixth wicket. A magnificent chasing win.

13. Sri Lanka v India in Galle, 2015
Sri Lanka won by 63 runs

© Ananth Narayanan

This is the only Test from the current millennium. Sri Lanka did not even reach 200 on the first day since they could not fathom R Ashwin's bewildering mix of deliveries. Centuries from Shikhar Dhawan and Virat Kohli meant that India finished with a huge lead. Sri Lanka's notional target for the third innings was a massive 526 (12*27.9+192). That looked a pipe dream when they slumped to 5 for 3. Kumar Sangakkara and Angelo Mathews retrieved the situation partly but at 95 for 5, the abyss beckoned. Dinesh Chandimal then played an innings fit to be remembered alongside Sanath Jayasuriya's 253, Mahela Jayawardene's 180 and Kusal Mendis' 176. His run-a-ball 162 was a complete bolt from the blue. He added valuable runs with Lahiru Thirimanne and Jehan Mubarak, and the final total of 367 meant that India would have to chase 176. The momentum was with Sri Lanka and Rangana Herath was virtually unplayable. India folded for 112, leaving Sri Lanka winners by 63 runs and Herath with figures of 7 for 48. Only towards the end of the innings did Sri Lanka gain the ascendancy, and this resulted in a WinIndex distribution of 4-16.

Australia v West Indies in Brisbane, 1960-61
Match tied

© Ananth Narayanan

And finally, an extra atop the baker's dozen, a tribute to that most sublime and remarkable of Tests: the first tied Test. The first two innings were in excess of 450 and West Indies started the third innings in arrears by 52. Their notional target was 452. They were behind right through the innings and set Australia a final target of just 233. Only at 49 for 4, 57 for 5 and 92 for 6 in the final innings did West Indies forge ahead. They fell back again because of the huge seventh-wicket partnership between Alan Davidson and Richie Benaud. Australia were then firmly in the saddle until the fall of the tenth wicket. The final WinIndex distribution was 3-17, a terrific comeback result for West Indies.

The other matches in the shortlist are briefly enumerated below, in reverse chronological order.

India v England, Chennai, 2008-09. A successful last-day chase in Chennai after Kevin Pietersen's challenging declaration. The relatively easy win proved a stumbling block in the final selection.

India v Australia, Mumbai, 2004-05. A dead rubber played on an apology of a pitch. A close but thoroughly unsatisfactory Test.

West Indies v Australia, Antigua, 2003. Dead-heat first innings, followed by West Indies completing the biggest chase of all time.

Sri Lanka v South Africa, Kandy, 2000. South Africa's successful defence of 177 through unfancied bowlers like Lance Klusener and Nicky Boje.

West Indies v Australia, Barbados, 1998-99. Arguably the chasing innings to beat all others, Lara's 153 not out taking West Indies to a one-wicket win. However, the WinIndex count was 10-10.

Sri Lanka v Zimbawe, SSC, 1997-98. Sri Lanka chase down a huge target through a magnificent de Silva hundred. A 1-19 Test.

England v Australia, The Oval, 1997. Phil Tufnell and Andy Caddick defend a paltry target of 124 against a strong Australian batting line-up.

New Zealand v Pakistan, Christchurch, 1994-95. A tough chase masterminded by hundreds from Bryan Young and Shane Thomson.

England v Australia, Manchester, 1961. A spin masterclass by Benaud, who used the rough to dismiss England for 201 after they were 150 for 2.

England v Pakistan, The Oval, 1954. A magnificent match performance of 12 for 99 by Fazal Mahmood taking Pakistan to a 24-run win in a low-scoring Test. England had Hutton, May, Simpson, Compton, Graveney, Tyson and Statham in their ranks.

Australia v England, Manchester, 1902. Trumble and Saunders defending 124 with three runs to spare.

In addition to the five Tests between August 1992 and September 1994 in the final selection, there is a sixth from that period in the shortlist. Further, there is one other Test from the 1990s in the final selection and three more in the shortlist. The 1990s were truly the golden decade of Test cricket.

*The result was tweaked to indicate that it was West Indies and not Australia who won the Adelaide Test by one run

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





  • POSTED BY SG on | January 13, 2018, 23:55 GMT

    Sorry for the delayed response hope you publish this ...

    You said:

    "Please don't think I am not aware of the feelings of all of you. It is just that, in my algorithm, Australia was winning at 52/1, 97/2, 115/3 and 232/4. Then it was India all the way."

    See this is the problem ... you still anyhow concluded that Kolkatta'01 was not the best comeback test despite knowing the shortcommings of your algorithm. The damage is done. Same problem as with the 2001 Best inngs list that ignored a lot of great inngs. Do you have any idea how trolls and people with hidden agenda use these things to run down Indian Cricket ? I guess not. This is the same problem (Except not as serious) that Narendra Modi had to face for years from vested interests with a big media mouth piece.. I hope you realize the gravity of the matter. With power comes great responsibility.

  • POSTED BY GV on | December 21, 2017, 13:38 GMT

    Ananth, to Jasprit's point - if there is a hypothetical match identical to Calcutta test upto the point where Australia collapsed on the last day. Except that in this match, Australia score more and reach within 50 runs of the 4th innings target but lose so the loss margin for Australia is significantly lower. Will this be considered a bigger comeback by India since the margin of victory is lower?
    The key question is whether Australia had any chance of winning, in any of the hypothetical situations. If not, there is likely to be no change. The magnitude of Winindex does not get taken in. Only whether it is above 50 or below 50. That concept itself might reauire some rethinking. I may have to treat the 20% different from 45% and the 85% different from 55%.

  • POSTED BY jasprit on | December 21, 2017, 5:27 GMT

    Third....real situation....There was a game in county cricket once where the scoreline was 223/15/521(f/o)/158. Win by 155 runs after being bowled out for 15. Would this qualify for a comeback using your current parameters. I look forward to your responses. Then we'll chat further. Cheers!
    The fact that 20 wickets fell for 238 means that the first innings RpW is 11.9. As such the third innings score is a complete outlier. This match might have to be selected by me as a special selection even though it may not meet the 3-17 type of guideline.
    I could have done this for the Calcutta Test. But I genuinely feel that the MCG was a greater win. I still think that the 281 is the greatest innings played by an Indian batsman and is an all-time top-10 innings, the Laxman-Dravid partnership is one of the greatest ever, Calcutta 2001 might be Harbhajan's greatest Test ever, but the MCG win was the greater one. Give me that.

    (I know the point I am trying to make. And it is not recency bias. I will make it later. But first, I will wait for your responses. )

  • POSTED BY jasprit on | December 21, 2017, 5:23 GMT

    Second hypothetical...let us say a game had a scoreline of 250/0/1000(f/o)/250. How would this game rank among the above games for comebacks?

  • POSTED BY jasprit on | December 21, 2017, 5:22 GMT

    And now to you Ananth....Just out of curiosity....Can you please help me with 2 hypothetical and one real situation? First hypothetical, let us say, in Calcutta, India were all out for 500 with a lead 226 and Aus were then all out for 212 and lost by 14 runs. How would Calcutta 01 compare to the above games then using your parameters?
    These are totally different situations. Defending 226 and coming out the winner by 14 runs would certainly be a much greater comeback win because there is a real danger of losing. The Winindex numbers will reflect that.
    Anyway, defending 150 and winning by 20 runs is a greater win, in certain measures, than defending 500 and winning by 250. Ultimately it is the greater chances the other team has to win matter a lot. Of course, the 20-run win might be a dead rubber while the 250-run win might be a series decider. That is another issue.
    The passionate Indian cricket follower like you places a lot of value on the win after a follow-on (one of three such instances), bringing down the Australian juggernaut, walking the Ganguly-talk, ending the 16-win streak, the euphoria of the last three days' play and so on. As a, then, strong Indian supporter, I am with you. But these are all emotional/off-the-scorecard feelings.
    The cynic in me also tells that Ganguly never really went for a win (he would then have declared overnight), the win happened, and also that if Laxman had not got out at 281, the match might have been a draw.
    The analyst in me says that considering only wicket-fall situations, the last time India looked like losing was at 232/4 (300/4 was not a wicket-fall situation). Also that India looked like losing the MCG Test by a mile. Also that Headingley 1981 was a far greater comeback. England was so far behind that the 500-1 stories must be true. Also that I must be an analyst first and not be swayed by off-the-scorecard factors.

  • POSTED BY jasprit on | December 21, 2017, 5:21 GMT

    Next, I must admit, this omission did upset me at first. There have been others disagreements for me before like Ponting not being in Aus home X1 and Kapil's wicket being more valuable than Gavaskar's and Murali Vijay's being more valuable than Virat Kohli's, and I was extremely vocal about those at those times. But they never upset me. For some reason, (I know what they are. I have mentioned them before in my earlier comments, but I am not going to repeat them) this one did tick me the wrong way. But now I ve gone through the games chosen and I see the reason why. This algorithm/article favors the games that went close in the end and were more of nail biters/thrillers than biggest comebacks. Calcutta's eventual margin of victory was much larger for it to be selected using this article's parameters.

  • POSTED BY jasprit on | December 21, 2017, 5:20 GMT

    First....Saumik, my friend, Many Thanks for your kindness and support. I Appreciate it. Second, Everyone....I am not in a fight with Ananth. Third, I am extremely disappointed that this discussion took the shape of Indian victory and appreciating Indian performances only vs non-Indian performances.. Fourth, I am extremely disappointed that this discussion was given the shape of Calcutta 01 vs MCG 81. That is just a shame. Atleast from me, that was never the point. The point was to appreciate a good game of Cricket and see it be recognized fairly. Thats all.

  • POSTED BY Saumik on | December 20, 2017, 11:14 GMT

    Will egarly await your next. No love lost! Cheers

  • POSTED BY Saumik on | December 20, 2017, 8:35 GMT

    Dear Ian and Ananth... am sorry if I came out strong. I loved the way the article was narrated and I agree to it. What I did nit appreciate was the way Jasprits comment was handled. Apologies if I came out strong idea was not to bullie anyone. We all are humans and we have scope to improve it applies to all including me and Aanath as well
    Two things, Soumik.
    I got quite upset when Jasprit repeatedly said that he did not and would not read the article. That is not correct and is very unfair to the writer.
    The second is that I promptly apologized and if you want to know, I have taken a couple of Jasprit's ideas to form the next follow-up article. I am as quick to apologize as take offence. I guess one does not come without the other.

  • POSTED BY Cric on | December 20, 2017, 2:01 GMT

    Ananth- et al is a list-ender for names and not things.

  • POSTED BY GV on | December 19, 2017, 17:06 GMT

    Hi Ananth, I am surprised at the reactions on one plane about Calcutta win and not surprised on another plane. Am surprised since readers of this blog must know that your statistical method is thorough and while improvements can be debated, it will not fail to excavate even the most forgotten of matches, so if Calcutta is not featuring, then surely your method is offering clues on why not, and equally why the other matches are greater thrillers - they have to be, since you are applying the same yardsticks. I am not surprised since this seems to be recency bias, which largely explains Sachin phenomenon as well - 10 years later Kohli will be defended equally raucously. Sachin, Australia as the greatest team ever, Calcutta win - they are all of the same period, and go together to the emotional exclusion of equal players and wins from previous eras.
    What bothers me is that what is pushed down and ignored is also Indian - The MCG win now or Laxman's innings or Kumble's 10-wkt performance, then. The yardsticks are so different.

  • POSTED BY Saumik on | December 19, 2017, 3:28 GMT

    Hello Ananth Been a great fan of your articles and always cherished them. Looking at the way you defend your thoughts is commendable but comes out very strong. We are readers and have like you have our own views and enlightened to our opinion. I have never ever commented or expressed y thoughts though I use Cricinfo regularly. You are surely enlightened to defend your logic or algorithm or math to determine the way you choose the test matches but remember emotions are stronger than maths and that's why Kolkata 2001 not being there is disturbing a lot of folks. Respect that thought and move on. Agree emotions cant be measured but do respect the emotions of your readers rather than pointing out your logic repeatedly - it comes out a bit adamant. Many of your readers like self not only come to read the article but also read the comments. Cheers!!! Jasprit - bro Kolkata 2001 - I was at the ground when we won respect your is the greatest test ever and we all know it :)

  • POSTED BY Cricinfouser on | December 18, 2017, 2:54 GMT

    Where does the Indian win against Australia in Kolkota 2000-01 series stand? The epic partnership between VVS and Dravid eventually won the series for India. India were 1-0 down and were following on.

  • POSTED BY Bijendra on | December 17, 2017, 5:27 GMT

    one entry i think should be here #Kolkata 1999 test where Pakistan had slumped to 26 for 6 after Srinath & Prasad's thunderous spell..#Moin Khan brought them into the match with a fighting inn. and even in 2nd inn while chasing 270 Ind wre pretty much in at 100 odd for 1..yet Pakistan won cmprehensively in the end..
    A very good suggestion. If you take the match as a whole, the 26 for 6 and the recovery become very significant. However, at the end of the Indian first innings, Pakistan trailed by 31 runs only and the all-time top-20 masterpiece of Saeed Anwar (188*) ensured that they were almost always ahead of India. The WinIndex count at the end was 13-7 in Pakistan's favour. Even these 7 instances were the result of India's good start in the fourth innings: They were 108/0, 145/2 and 183/4, chasing 279.
    Thank you for pointing out to a nice match.

  • POSTED BY Murray on | December 17, 2017, 1:08 GMT

    Hope one day you write an article on snatching wins from almost certain draws ? Adelaide 2006 Springs to mind.
    For that matter, SCG 2008, Adelaide 2003 et al.
    My problem really is that the TCM articles are 'Premium' articles and are scheduled only once in four months or so. I put in a lot more effort into these. The editing process takes the better part of a month. The next year's ideas are already agreed upon. So some of these follow-up articles chould be done in EspnCricinfo, where I will re-start in January. There I have a monthly slot.

  • POSTED BY Murray on | December 17, 2017, 0:56 GMT

    Nothing to do with this article, but I wonder what happens when Mitchell Marsh realises he's a better batsman than Smith. I've waited 8 years for yesterday.
    Many comparisons are valid. All-rounder entry, overall better batsman than bowler, attacking instincts and so on. Marsh is still young at 26. So he has enough time to make a real mark as a batsman at around no.5/6.
    We could see a 300 and 200 today.
    Unfortunately, neither 300 nor 200. But the cracks could see an England scoreline of 200 all out.

  • POSTED BY sukalyan on | December 16, 2017, 23:26 GMT

    This is a beautiful analysis quantifying the vague notion of coming from behind. The methodology, however, only considers how late the tide changed to rate the test match as most competitive or not. This is why the matches, which are ranked, have a late value of the win-index crossing 50%. Another question could be how long the ultimately defeated side dominated substantially the proceedings. It would be a rather geeky measure, where we have to take area under the curve defined by a function (50-winindex) until the match is over. Then, we may see a different set of test matches in the list. Is it worth investigating, or is it too geeky and complicated?
    This is one article for which a few excellent suggestions have been received. Yours is one. Similar ones have already been received. To handle differently the WinIndex between the sub-30 values and the 40-60 values. Many of the matches featured had quite a few 40+ WinIndex values which were all pushed into the 'losing' situation. As such, a 3-17 match didn't really have that wide a variation. The idea is to differentiate between the really low WinIndex values from the middle level values. There is something intriguing in this.
    As I have mentioned in my response to Murray (Archer), the problem is to find a slot. Maybe I will use the EspnCricinfo slots.

  • POSTED BY GV on | December 16, 2017, 12:44 GMT

    Ananth, does Smith have the highest average for any batsman with 59 tests and 5800 runs?
    Yes, of course. Once the number of Tests crossed 52, Smith stood and stands supreme. If he is out first ball tomorrow, his average will be a mind-boggling 62.21. Now, he is unlikely to fall below the three others in the 60+ range, as happened after the Adelaide Test. And who knows, he might add another 80 runs tomorrow.

  • POSTED BY Srivatsan on | December 15, 2017, 13:33 GMT

    Mere statistics cannot define the greatest comebacks!, any model leaving out the 2001 Kolkatta test is questionable. Even a naive cricket fan would put that test first in the list at least in Greatest comebacks in Tests by India!.

  • POSTED BY shinde4240297 on | December 15, 2017, 8:24 GMT

    Well, I am glad that cricket is not a statistical game. The name of this article should have been "Which are the greatest statistical comeback in Test history". This is actually great analysis but why people love this game cannot be put in statistics. In similar fashion, Sachin's record will be broken but it cannot be calculated by love, passion and happiness it brought to millions of people. So many people have broken Lara's record of 34 centuries but it does not mean that those players are great than Lara. Similarly, all the 13 wins are great but cannot be compared to a win which was against a team which had 16 Match winning streaks, possibly the 2nd best team ever in the entire history of cricket. The statistic is just one side of the coin as there is a different aspect of a game which statistic cannot bring it to light. Make no mistake this is an awesome article but reflects only one side of the story.

  • POSTED BY Rangarajan on | December 12, 2017, 8:47 GMT

    One more on the actual article - Most had a pattern. High first innings deficit, low second innings score and small target messed up by chasing teams. Where does India's loss in 1996-97 Barbados rate in this?
    Match 1363. The WinIndex comparison was only 3-17 for West Indiws. This was in the first shortlist. I do not remember. But probably I could not accommodate this and the 1-run win.

  • POSTED BY Rangarajan on | December 12, 2017, 8:45 GMT

    Hi Ananth - Good analysis... The basic premise of so many arguments here is that most readers look at a "match" whereas this analysis talks about a "game"... For example, between Mel 81 and Kol 01, probably the readers were looking at Australia, 16 match winning streak, Waugh, McGrath, Warne, etc. Whereas, inside that match, which progressed, we had multiple stages and then after some point, say the point of inflection, this match was turned firmly in one direction. If I were to bet on India "bowling" out Australia on day 5, I would have won a fortune (which also caused Ganguly to delay his declaration so that we won't lose). So it was a surprise turn-around and definitely one of the best wins of millennium. Probably the new generation readers don't know about India before 1983 :)

  • POSTED BY jasprit on | December 11, 2017, 21:09 GMT

    And one more thought....' In my algorithm, Australia was winning at 52/1, 97/2, 115/3 and 232/4. Then it was India all the way. '......It would be interesting to see an article based on turnarounds where the magnitude of turnaround was determined.. In other words, turnarounds which involved the biggest change. In other words, where a team got far behind in the game, turned around, made a comeback and pulled far ahead. Some set of games where the magnitude of this turnaround was biggest would be interesting to see. Cheers!
    In some matches, the turnarounds have inched forward across 10/12 wicket-falls from, say, 40 to 60. But this sort of 30% turnaround was there also in the Gilchrist-Langer stand - from 29% to 55%. This gives another possibility. Truly match-changing partnerships. THAT could very well be a TCM piece.
    Ah! Ideas galore. You guys rock. We might tick ourselves off quickly but also come around to get great ideas going.

  • POSTED BY jasprit on | December 11, 2017, 21:03 GMT

    Thanks Ananth. I think ICC should acknowledge your database and use it to develop new insights. Get some IIT engineers to work on it with you and develop some new viewpoints and perspectives on the game. For eg, it would be interesting to incorporate some batting strength index into this article to see what kind of batting strength India had left at 232-4. And then combine it with the bowling strength index of the opposition as well. And see where we get. And I so wish your articles on cricinfo like before became a norm again. Good luck.
    Since this was an initial attempt at determining the match status, I probably went half-way. Your idea of the batting strength available at 232 for 4 and matching that against the (notional) target occurred to me but I left it out as too complicated. But that is a very good idea worth exploring.
    It is eminently possible that the four wickets might all be top order batsmen, as in this case, or that there was a night-watchman hidden in the four. That would make a huge difference. Also, today, four down could bring in Ashwin, Saha and Jadeja - all having averages around 30. At Kolkatta we had Mongia and the tail.
    I already have 2/3 ideas to make this work better. I might do a follow-up piece in EspnCricinfo rather than TCM, since I have only limited aricles with TCM.

  • POSTED BY jasprit on | December 11, 2017, 7:53 GMT

    . But, the truth is, nothing is perfect and we always look to improve ourselves, everyday. Isnt that what we live for? Cheers!

  • POSTED BY Sitaraman on | December 11, 2017, 7:35 GMT

    A rigorous piece but only goes to prove that a purely statistical approach to deciding what the great comebacks were is a woefully inadequate method. Kolkata 2001 was a win after a follow-on (a near miracle in itself) against one of the world's best-of-all-time teams that was on a red-hot streak. It featured an unforgettable innings, possibly the greatest of all time. It was unexpected and magnificent, and turned the cricketing world on its head. The Melbourne 1981 win was a delightful surprise but there was an element of fluke about it -- an Australian collapse on a poor 5th-day wicket. It did not lead to any great days for Indian cricket. Using stats, however cleverly, to arrive at this result is like using solely the yo-yo test to determine who should be in the team: You would then have the Dhawans and Pandeys of this world displacing the Gavaskars and the Laxmans. And that, is we know, would be a travesty.
    May I request you to look at the other Tests featured carefully. Wasn't the 336 on the last day against Tayfield and company nearly as difficult task as what India faced. The point is that so many things happened off the ground regarding the Kolkatta Test that all these are used to put the magnificent win on an even higher pedestal. The needle between Steve W and Ganguly, the unbeaten run, the swagger of the Australians, the statements made by the visitors, the Mumbai result, the near-invincibility of the Australians et al. But these are not on the scorecard. Almost always, my analysis is based only on scorecards. Let the others write on the off-field happenings. I stick to the scorecards. And am dispassionate about it.

  • POSTED BY jasprit on | December 11, 2017, 7:21 GMT

    And you say Laxman's innings was made greater by Tendulkar and Harbhajan. Exactly! Afterall, we are talking about team comebacks, Right? Nobody is talking about Laxman's innings. Just fyi, I am not even a Laxman fan. I dont even like him. But the team to comeback from where it was especially against the all conquering side on world record setting 16 straight wins was just monumental. A true landmark. Cheers!

  • POSTED BY jasprit on | December 11, 2017, 7:16 GMT

    I told you I understand your point why Kolkatta is not on the list. Its not because you are algorithming the games where the odds of win were low. When India started their second inning, I agree India had 10 wickets in hand. So, the odds were not low and India could always score big. But the odds of that happening were 1 in 2250 ( Actually 0 in 1535 as that is what that Test no. was). And today, its 1 in 2287 as it has still not happened again. And you say India were not in desperate situation. That is what you should call your article. Wins from unlikely situations. A comeback is a comeback. It has nothing to do with winning. And when you talk comebacks, there is none bigger than Kolkatta. Well, I would not say none, because calling greatest this or that is a fool's game. But it would definitely belong in the Best!

  • POSTED BY Paaji on | December 11, 2017, 6:55 GMT

    Contd. 3) This is linked to my first comment in some sense...the uneven split in the test match win index (1-19 or similar) tends to favour matches where the decisive action happened later in the match. Going back to the Kolkata test, without that partnership the match would have probably been over. It was key to India avoiding defeat. But don't you think getting wickets out when teams could have chosen safety is better than teams "going for the target"? Perhaps you should also include a draw index which shows the chance for the team to have lost from such a position. PvP
    A split of 1-19 is spread over both the third and fourth innings.

  • POSTED BY Paaji on | December 11, 2017, 6:43 GMT

    Hello Ananth, nice to see you back on these pages. Good selection of tests but I feel compelled to add a few thoughts: 1) I feel Win index hides some information--whether the opposition had a chance for a draw or not. Like others, I'll use the Kolkata test as an example to highlight it. Once India batted itself to a great position, it couldn't lose the match. Only a collapse by Australia/great bowling by India got the win into the equation. This is different from other examples where draw wasn't an option due to time left in the match. Any chance where the time or possibility of a draw for the losing team could be factored in? Reason being, the team that has lost had 2 options--playing out for safety, or going for the target and this might have dictated their approach in the match.
    I understand that the numbers add to 100. But if I say that right at the top, I select only the Tests which had the appropriate result, doesn't this get cleared up. Since the Test ended in a result, there is no draw to be taken care of. The WinIndex for the batting and bowling teams would add to 100.
    2) You've not considered team strengths here but I feel a note should be added here (CTD values or some other way). An average team is less likely to mount a result against the grain, isn't it?
    Right from the beginning I was only looking at the turnarounds without bothering about the team strengths. Once I consider the team strengths, then it would become "The most unlikely wins by weaker teams". You remember that I had the TPP (Team Performance Points - which takes into account the result, team strengthes (location based) and the margin). When I see that value, the highest TPP is for match 1140 during 1990 (2.940-0.283) when a weak English team defeated a very strong West Indian team by the huge margin of 303 runs.

  • POSTED BY jasprit on | December 11, 2017, 6:37 GMT

    Copy Paste... 'the last time India looked like losing was at 232 for 4. Afterwards, and my algorithm is not flawed, let me assure you, India looked like winning always. '
    Yes, Jasprit. The last time India looked like LOSING was at 232 for 4. Afterwards, from 608 for 5 onwards, India looked like winning. The measuring is done only at wicket-falls. You would have known it if you had cared to read the article.
    It is unfortunate that even after making a huge mistake, you are refusing to admit it. But it does not matter. I would appreciate that you comment henceforth only if you read the article.
    Good bye.

  • POSTED BY jasprit on | December 11, 2017, 6:06 GMT

    To say that at 232-4, India looked like winning is just laughable. At 232-4, India were still 40 in deficit with last pair of recognized batsmen at the crease, with Nayan Mongia to follow. In the tail were Zaheer Khan, Harbhajan Singh(nowhere near the batsman he later bacame in his career. Infact an absolut mug at that stage), Venkatpaty Raju and Venkatesh Prasad. And what happened after that is History. One of the most beautiful and astonishing moments in Cricket. Enough said. Cheers!
    Where did I say that at 232 for 4, India was in a winning position. I said that "the last time India looked like losing was at 232 for 4" That meant that at 232 for 4, Australia was still winning. Do not make up the words to make a case for your argument.
    I expected better things from you rather than drumming up your own fiction to bolster your case.
    It takes a long time to gain respect. But one such deliberate misrepresentation is enough to lose all the respect gained.
    We live and learn. I would appreciate if you do not waste your time posting such deliberately misrepresenting mails. Even the others who are supporting the Kolkatta Test have not resorted to this.
    And finally, if you have not read the article, you clearly lose the right to comment on it. That is fundamental courtesy and what an author deserves.

  • POSTED BY jasprit on | December 11, 2017, 6:03 GMT

    And I am So Sorry to say, instead of explaining with reasons, you ve gone on to belittle one of the Greatest achievemnts in Cricket by Any team, let alone India. That comeback is Monumental. Infact, the Vastness of it actually makes your head spin and leaves you dizzy. To start from 274 behind and reach 350 ahead, if that is not the greatest comeback, then I dont know what is. It is the crowning jewel for the game of Cricket and takes its uncertainty and Greatness to a different level. If Cricket was in a competition with another sport(s) and wanted to show its vastness, I am sure That Comeback would be one of its proudest moments.

  • POSTED BY jasprit on | December 11, 2017, 5:53 GMT

    And now coming to this article, I ve not read this article. And I am not going to. Obviously, because of Kolkatta 01. But now that I ve understood your point a little bit through the comments, I understand your point that this article lists the matches where the team had ALMOST lost the match, like 80 or 90% according to your algorithms.. And with Kolkatta, when India started their second inning, they still had 10 wickets in hand and there was always a chance that India could go huge and score big. But the history tells us the likeliness of that actually happening are 1 out of 2250. It has Never happened. I think you should (have) named this article as the unlikeliest victories or winning against the odds or something, and then written a short paragraph about how the odds were not as low as some others in Kolkatta. Because when talking comebacks, Kolkatta is the Mother of all Comebacks. There is Nothing like it. No list of comebacks can start without Kolkatta or Headingley at the Top.
    I only decide on the title. I do not decide on the photographs or captions.
    How is it that Kolkatta is a 1 in 2250 situation while a scoreline of "311/75/99/336-5" is not given its due. Wasn't the 336 as unlikely as India scoring over 500. Or for that matter "256/547/471/164" or for that matter "169/292/239/111" and so on.
    My point is that it was probably much more difficult for South Africa to defend 116, for Australia to score 336 on the last day, for Australia to defend 180, for Australia to reach 369 after being 126 for 5 or for Sri Lanka to defend 175 and so on.
    Why should the greatest comebacks only talk of Kolkatta 2001. And why should an analysis which does not include this Test, for clearly outlined reasons, be not worth further consideration.
    And let me say in no uncertain terms, Kolkatta and Headingley are not even remotely comparable. India was NEVER in a situation as desperate as England was. Not in the third innings, not in the fourth innings. This, I say, with a firm belief that Laxman's innings was not for the mortals, but for the Gods. But his innings attained greater sheen because of Harbhajan and Tendulkar.
    I expected better from you, Jasprit.

  • POSTED BY jasprit on | December 11, 2017, 5:48 GMT

    From earlier aticle 2....

    By reverse article, I mean exactly what you said. Games in which one team had lost the game 80% or more. And then came back to win or draw it. It could be thru one partnership, or combined effort. Main point is the team's loss %age was 80% or more. And then they came back to win or draw. That might be an interesting set of games. Thanks! [[ Let me see what I can do.

  • POSTED BY jasprit on | December 11, 2017, 5:48 GMT

    From the earlier article....

    I think a reverse article might be good too. This article shows games where teams lost from a strong position. How about games where teams won from positions where they almost lost the game. It could be due to individual wicket/partnerships or combined team efforts. Kolkatta 2001 and Headingley 1981 would probably take the cake, but it might show some interesting games. Thanks! [[ In reality this article highlights wickets which changed the win % from above 80% for the batting team to either losses or draws/ties. In other words the wickets which changed the course of a game completely. It is difficult to define what you want. A wicket can only change a certain win to a loss. Not vice versa. A partnership can. Ananth

  • POSTED BY jasprit on | December 11, 2017, 5:47 GMT

    What a Shocker Ananth! A few months ago, you'd written an article called Golden wickets where a fall of a certain wicket changed the course of the match and the team in lead eventually lost. In essence, it was an article in which a team lost after being in a very strong position. In that article, you'd chosen a team lost after their win %age was 80% or higher. At that time, I'd asked/suggested you to do a reverse article where a team wins from being far behind in the game. The only difference in this article is you ve changed the odds from %age to fraction and called it WinIndex (I think, I ve not read the article. And I am not going to.)
    I agree that there are similarities. However, the two indices are totally different if you take the trouble of going into the details. Of course, it is your prerogative. Only I know the difference and the amount of work I have put in. Superficial reading, and certainly 'non-reading', will not help.

  • POSTED BY SG on | December 11, 2017, 3:42 GMT

    The explanation for not considering Kolkatta 2001 are just shocking !! Bowlers were tired and being 4 Down minus 40 at the fall of 4th Wkt somehow means Aus were no longer ahead? Seriously? I will be interested in knowing who else here will agree with that. Anybody who understands Test Cricket will tell you that Aus(With *THAT* team would win 99 times out of 100 with both SRT and Ganguly out) But since you obviously believe more in stats I suggest you lookup how many times a team having enforced followon has lost. Thats just three matches out of the many such matches. So much for tiredness! BTW VVS Batted for about 2 days worth of batting from one end (or 90 overs on his own) while listing).I guess he is immune to being tired? When VVS got out all commentators got up and started applauding. How often does that happen.

  • POSTED BY Palak on | December 10, 2017, 17:53 GMT

    In this article, you have only for how many wickets team were below the 50% mark but that can also happen in a close match where the team is floating just below the required target like 9, 11 & 12 in the article. In these instances WinIndex was greater than 40 most of the time. But for a real comeback, it is also important that the team reached a situation from which nobody in their sane mind would agree to bet on their winning. i.e. the WinIndex was 25 or less and then the team did a turnaround. Hope you would consider that come up with another article.
    A similar point has already been made. I will look at doing something on the lines of really looking at the sub-30 WI values and dumping the 40-60 together as a 'difficult to call' situation.

  • POSTED BY espn28518479 on | December 10, 2017, 17:33 GMT

    Thought we had closed this :) - " then what about Australia moving from 3 down to all out in 90 minutes 20 years later. That is as bad a collapse as the earlier one" - Precisely the point, how many times has that Aussie team collapsed before or after that match in a chase? Can't think of even one. Edbagston 2005 or perth 2007, one of the few matches that the team lost, they still fought incredibly in the 4th innings. With due respect, when Ghavri bowled Greg Chappel around the legs in Mel 81, u knew the pitch had deteriorated as test pitches do on last day. Kol 01 was not so, that pitch was still decent, not 1 of the dismissals was due to the pitch misbehaving. It was a team playing quality cricket to beat an opposition of the greatest quality. Ian chappel once said he had asked Warne about his bowling during that second innings n Warne saying he hadn't bowled too bad and Ian said ' you didn't mate, it was just unbelievable batting". Mel 81 was gr8, Kol 01 was a miracle

  • POSTED BY Harsh on | December 10, 2017, 16:17 GMT

    Sorry Ananth also 3rd test at Trinidad between England and West Indies in 1994 where England were skittled for a mere 46 in the fourth innings after Curtly Ambrose ripped through the heart of the English batting unleashing a series of thunderbolts .Also feel the 2nd test at Adelaide of the 2003/04 series between India and Australia where India ressurected themselves from the grave to pull of a stunning win.Also South Africa's astonishing comeback in the 1st test of the 2011 series bowling out Australia for 47 in the 2nd innings to win by 7 wickets eventually.Finally also the 2nd test of the 2010 series between Australia and Pakistan where Mike Hussey single handedly revived a sinking ship to win a famous battle.
    1. From 5 for 3, it was West Indies' game.
    2. Once 556 was met by 523, the match went to an even keel. So there is no great comeback. Please note that I do not consider the match status at the end of the first innings.
    3. The final innings in the 2011 win was ridiculously easy for South Africa.
    4. It was indeed in the final reckoning until the last few days. The WI score was 2-18 in favour of Australia.

  • POSTED BY Harsh on | December 10, 2017, 16:04 GMT

    Amongst the drawn games I would select the 4th India-England test at the Oval in 1979 when India fell short by 9 runs of achieving a record 4th innings run chase of 438runs as well as the 2nd test in 1989-90 between Australia and Pakistan in Sydney where Pakistan almost turned the tables rising like a pheoenix from the Ashes.Arguably even the 1st test at Brisbane of the 1976-77 Australia-Pakistan series.after Pakistan came close to winning after conceding a huge 1st innings deficit.

  • POSTED BY Harsh on | December 10, 2017, 15:55 GMT

    Great list Ananth.Wished to mention the 1999 iindia-Pakistan test at Kolkata where Pakistan ressurected themselves from the grave at 26-6 to win the test.Even in the 2nd innings they were in a precarious position.Also the 3rd test at Melbourne in 1972-73 between Australia and Pakistan where Australia defended a paltry 159 in the 4th innings that too because of a stubborn stand of 84 for the ninth wicket.England's 29 run win at Edbaston in the 1981 Ashes also runs close as well as India's then record run chase at Trinidad in 1975-76. chasing a target of 406 in the fourth innings.
    In the Calcutta Test, since the deficit was only 38, Pakistan was always there or thereabouts. Through the latter two innings, the WinIndex oscillated either side of 50. Mainly because of Anwar's 188 (almost certainly all-time top-20 innings), the WinIndex ratio was 13-7 in Pakistan's favour. The other Tests have been covered earlier.

  • POSTED BY GV on | December 10, 2017, 15:27 GMT

    Ananth, nice work again. On MCG 1980-81, do remember Gavaskar's innings - it came on a terrible wicket - just see the youtube of this innings - the first ball from Pascoe nearly takes Gavaskar's head off. Come back from behind in the match and in the series. Under pressure due to poor form throughout the series. Very good bowling attack - Gavaskar aced it. When I read Lillee saying that Gavaskar could not get a century against him, I think neither could the whole Oz team, making only 83.
    Unfortunately people do not realize how much was achieved by players with very little resources and incentives. Today, when I watched the collapse of India against a couple of very good, but not great Lankan quicks, I realize that for all our bravado, we are going to find it really tough when we travel abroad.

  • POSTED BY Pawan Mathur on | December 10, 2017, 13:15 GMT

    Contd - Take for example, West Indies vs India 199 7 Barbados . There were atleast 3 points in the test where India were in a strong position - Reducing West Indies to 193/7 in the first innings, then being 212/2 in their first innings in reply to West Indies total of 298 and managing to secure only a lead of less than 24, then again bowling out West Indies for 140 (and at one point the score was 95/8) and finally failing to chase 120 and managinng only 81. This is more of a test India lost than West Indies won through a comback
    Very fine request with its own multiple nuances. A challenge I would look at seriously. Thanks, Pawan.

  • POSTED BY Pawan Mathur on | December 10, 2017, 13:08 GMT

    1. "If you ask me which is the greatest couple of years in Test cricket, I will unhesitatingly say 1992 to 1994" -. Apart from the number of tests in this period covered in this article, one thing that further validates this comment is the fact that 3 teams - Pakistan, South Africa and Australia were evenly matched during this period and looked ready to take over the number 1 spot from West Indies,who were still a decent side in this period. It was absolutely remarkable how Mark Taylor's Australia surged ahead in the mid 90s with the series win in West Indies and then more significantly with a 2-1 home win over a dangerous Pakistan . 2) I would also suggest a follow up article to this- examining tests where teams committed hara kiri, could not capitalize on multiple positions of advantage and lost tests they should have won. (Contd..)

  • POSTED BY Murray on | December 10, 2017, 12:00 GMT

    Brilliant work as always Ananth. A super list ! I'm personally very happy to see Harvey's classic. The 12th man swears that pitch in Durban was tougher to bat on than the one in Brisbane next year. For my own interest, what measurement was Australia v Pakistan at Sydney in 72 ? That all seemed impossible.....

  • POSTED BY DAVID on | December 10, 2017, 11:39 GMT

    You ask Who knows what England's last pair said as they went after the 15 runs needed to beat Australia at The Oval in 1902? I can tell you. In an interview many years ago I asked Rhodes (then blind and in his nineties) about the "We'll get 'em in singles" fable. He firmly dismissed it. "It wouldn't have made sense, would it? We might have got a two!" In fact there was a four and a two as they nudged England to victory. How people love to cling to myths.
    Many thanks for your informed comment.
    You will see that I have not referred to the 'myth' at all and have rather gone on to speculate on what might have actually been said. Might very well be my third phrase.
    In my library of cricket books, amongst the most treasured ones are three of your books, including "The Fast men" and "The Archie Jackson story". Once again my sincere thanks for finding time to respond.
    Take care and all the best.

  • POSTED BY Murray on | December 10, 2017, 11:39 GMT

    Brilliant work as always Ananth. What measurement was Australia v Pakistan at Sydney in 72 ? That all seemed impossible.
    Great Test. Australia ended up at 3-17 and won. Through their second innings they were around 20%. This Test was in my first shortlist. Australia got ahead only at 95 for 7. Could easily have got in.

  • POSTED BY paulma0962217 on | December 10, 2017, 10:59 GMT

    one thing is for sure. i have never seen a team lose so many close games as Australia have during the last 30 to 40 years. in fact i don't think i have ever seen them win "a close one".
    Yes, you are correct. They seem to be missing on many close matches. The Cummins debut match and Mark Waugh century match are the exceptions.

  • POSTED BY uday on | December 10, 2017, 10:53 GMT

    Very interesting analysis as usual Ananth. I would argue though that, more than measuring "comebacks", the Winindex parameter actually measures matches where the ultimate winning team was more likely to lose for the longest time over the last 2 innings. This is why, I think, that Kolkata 2001 does not qualify - after the VVS-Dravid partnership, India were on top and Winindex must have been in India's favour thereafter (not that I am out to insist on it being in the list - like everyone else, its just the first one that comes to mind when I think of a "comeback" win).
    Very nice point. You have got the key point of the exercise. Let me recollect my own thinking on the Calcutta Test. I was there for the first day's play and returned to Bangalore during the second day. After the follow-on and until 254 for 4 I had given up the match as most people did. But let me say this. From around lunch time on day 4, the idea of an India loss never entered my mind. Maybe that is what is reflected in the WinIndex.
    It is another thing that no one, including Ganguly, though that India would win. Otherwise, Ganguly would have declared overnight.
    Conceptually, therefore is it possible that the deficit after the 1st 2 innings may be very small, but if the team batting third keeps losing wickets at just the right time, you may end up with a one sided Winindex ratio despite the team not really being very far behind? The contrast between the India-Sri Lanka at Galle (no. 13) and the two Australia Pakistan matches (Nos 11 and 12) seem to illustrate this best.
    Yes, you are correct. A team could hover around the 45-49 mark for a number of wicket-falls and come out x-X in the WinIndex comparisons.
    However, look at match 905. How far behind England was. almost always.

  • POSTED BY sathis9817239 on | December 10, 2017, 7:09 GMT

    After looking at the scoreboard of India vs Australia Melbourne, it was more of a freakish collapse from Australia than a strong comeback from India. It seems India was hapless for the first 4 day and suddenly there was a collapse.. It happens all the time in subcontinent.. I agree with AMBIAN6198981 in his assessment ..
    If MCG was a win caused by the Australian collapse, then what about Australia moving from 3 down to all out in 90 minutes 20 years later. That is as bad a collapse as the earlier one.

  • POSTED BY ambian6198981 on | December 10, 2017, 6:17 GMT

    The two batsmen are still batting and looking at adding a million more runs - of course , they are looking to add , but what s the guarantee one of them s not going to b dismissed next ball ? Of course, all the matches in this list are mostly classics, but Kolkata 2001 was one from the heavens. Let us agree to disagree n close this

  • POSTED BY anthon3591782 on | December 10, 2017, 5:28 GMT

    Just wondering where Edgbaston in 05 Ashes comes into calculation obviously Aus didn't win but had they won would it have been an 0-20??
    Let us look at this nice match and thanks for taking the discussion into a new realm.
    England gets a 99-run lead. But at 29 for 3, Australia moves ahead. Stays there to the end. Because of the 407/308 and the RpW of 35.7, England's target for the third innings was 258 (357-99). So they fell short. Overall it was Australia which led 13-7 in the WinIndex stakes. Australia fell away after 82 for 3.

  • POSTED BY ambian6198981 on | December 10, 2017, 5:03 GMT

    " The last time India looked like losing was at 232 for 4" - Exactly the problem , I feel, with such algorithms. As a cricket fan , who watched this game, India were not safe at 332 for 4, 432 for 4 or even 500 for 4 . It was a matter of just 1 wicket, just take a look at the previous 3 innings of that series, how the tail crumbled. That was a period when tail could not score 12 runs after Sachin got out ( Remember Chennai 99? ) . It was only at the end of that 4th day wen VVS - Rahul ended at 589 for 4, felt safe India were not going to lose. I don't think anyone was thinking about the win yet , draw seemed most likely , more so when Aus reached 150 for 3 at tea. What Bajji n Sachin did in that final session was just as unbelievable. Leeds 81, Mel 81, Ade 92, Galle 2015 , all same pattern - team falls behind , 150 odd target ,team collapses on final day. How many matches in history have pattern of Kolkata 01? That is why it s unique.
    Please understand that my algorithm looks at the situations at wicket-fall situations. Let us not forget that at 332/4 and 432/4, the two batsmen were still batting and looking at adding a million runs more.
    As someone who has watched/followed all three of India's referred matches, I can say that all were terrific wins. It is just that the MCG win fitted into my current definition a lot more. Anyhow let us close this argument. Please look at the other matches featured.

  • POSTED BY Giri Iyer on | December 10, 2017, 2:48 GMT

    Very interesting Ananth... i am quite the fan of your analytics...this one obviously took a lot of thinking to get a pattern to emerge that would show momentum as a synthetic index. Wonder which one was the most momentum turns? rather than a 1-19 going from one side to the other would constitute a momentum shift. Would love a response.
    You can study the graphs. One reason why the WinIndex values are plotted. My personal favourites are the 0-20 and the 10-10 West Indies win by 1 run. See the number of times the momentum shifted.

  • POSTED BY humik on | December 9, 2017, 22:37 GMT

    The Australia vs Pakistan in Hobart was not a comeback test as the 'comeback' was cheating by the Aussie Umpire when he gave Langer not out - apparently the only person in the ground who did not hear that nick. Even Langer later admitted that he had nicked the ball which was given not out by the umpire.

    There maybe other tests in the list where umpiring of gambling may have been an issue (England vs Australia 1981, where two Aussie players put money on their team losing)

    I think the writer should take these into account before listing 'Greatest' as someone once said 'great is a great word'

  • POSTED BY cricfan92012399 on | December 9, 2017, 18:02 GMT

    What about the India's win in India-Austrailia Test at Bengaluru , 2017. I think that too was a great come-back victory.
    Yes, you have a good point.
    However, the small lead of 87 meant that at various stages in the third and fourth innings, both teams had their winning chances. The final WinIndex count was 6-14. So certainly it was a comeback win but when there are many other magnificent matches around, this does not even get into the shortlist.

  • POSTED BY jcproj1839713 on | December 9, 2017, 17:55 GMT

    India versus Australia - Kolkata 2001. If the statistical methodology does not rate this as one of the most amazing comebacks then whatever is the methodology - it is fundamentally and totally flawed.

    This is more the expression of a personal point of view being validated through some numerical methods.

    The title should be "Which are the greatest comebacks in Test history - as defined by xxxx method".

    Very very disappointed.


  • POSTED BY ambian6198981 on | December 9, 2017, 17:38 GMT

    Well, no test cricket fan wud not appreciate any of the other tests mentioned here. But using a 'customized' algorithm n ignoring arguably d greatest comeback of all times, which several greats have approved of seems odd. Have already quoted the context of the game, with due respect ,the 1981 Aus team had 3 or 4 greats, the 2001 team, barring Kasprowicz, all of them had a case for an entry in an all time gr8 Aus 11 n most of them in an all time gr8 world 11. Such things can never be simulated by stats, bit weird like the scene in the movie ' Sully' where real time action is questioned based on simulation. Test cricket is all about winning sessions, 2001 Kolkata was a case where after Sachin's dismissal , the only way India could win was wining all 7 sessions in a row, even a minor blip in any of those , it was game , set n match Aus. Incredibly VVS, Rahul, Harbhajan n Sachin ensured that . As a test cricket fan, it was pure treat, long, hard, difficult route to success. An epic !
    I agree with you 100%. The only reason I have selected the MCG win was because, all said and done, the last time India looked like losing was at 232 for 4. Afterwards, and my algorithm is not flawed, let me assure you, India looked like winning always. That there was a 10-hour epic stand in between is history.
    Give me credit for taking into account all on-field aspect.

  • POSTED BY Gitans2600138 on | December 9, 2017, 15:37 GMT

    The only issue here with the galle win is the poor umpiring decision, Chandimal survived due to a couple of bad decisions and went on to make a century. Hope you account such factors in the model. Otherwise an interesting piece.

  • POSTED BY Satish on | December 9, 2017, 13:16 GMT

    Hello Sir, how come India's greatest win against WI in 1976-77 when they chased down 402 by becoming the 1st team to chase 400 runs in the 4th innings not on this list? Also what about the Test match against Eng when Gavaskar scored his magnificient 220, even though the Test ended in a draw or even the Tied Test match between India vs Australia?
    Good points.
    Ultimately the Port of Spain win was an unexpected and surprising win facilitated by a surprising declaration by Lloyd. India was behind in the chase, only at 68 for 1. They they were ahead always. As such this match does not qualify as a "Comeback win".
    The Gavaskar Test was a draw. As such it does not qualify.

  • POSTED BY Pramod on | December 9, 2017, 11:41 GMT

    Dear Anantha Sir, If its alright, is there a repository where I could get all these data to work on?
    These are created using my own proprietary database.

  • POSTED BY Pramod on | December 9, 2017, 11:38 GMT

    Dear Anantha Sir, Wonderful analysis once again, good to see you back . I guess there are some intangibles that cannot be captured, like the strength of the opposition and the form of the team, that would make Kolkata 2001 sweeter in the fans mind. But I do agree such analysis previously ignored factors and judgements aren't based on emotion. Recently, Cricinfo compiled a list of greatest innings and I felt the parameters weren't uniform across innings and the weightage was higher for innings that won you matches. Say for example Dravids 233 was elevated to a great innings mostly due to Agarkars flash in the pan bowling performance. If the match was a tame draw, most people wouldn't rate this innings as high. But it would be certainly interesting to see which batsmen had higher averages when faced with a significant deficit. I feel that a batsmen forcing a draw is equivalent to a bowler winning you the match and such innings are match winning in nature even if the result wasn't a win
    This is 100% a team analysis with only the match situation as the guide. Harvey's or Laxman's innings will come in only my brief report of the match.

  • POSTED BY David on | December 9, 2017, 11:08 GMT

    Thank you for an analytical excuse to indulge in some quality reminiscing about some of the great battles of the past. It's interesting, however, to think about the two big chronological holes in your shortlist - 1906 to 1949 and 1961 to 1981. What were the best come-from-behind victories in these time brackets? Did any just miss the cut, or were these simply periods of test cricket characterised by either clear-cut victories or dull draws?
    You will see some of these Tests in the also-considered list.

  • POSTED BY kailas5835897 on | December 9, 2017, 10:39 GMT

    This logic of calculations doesn't suffice a comeback win completely. kolkota test should be at top when the target runs and win margins were more. This was the best comeback when you talk about comeback. Other wins in the lists are like nail biting finishes.

  • POSTED BY shekar46789 on | December 9, 2017, 9:43 GMT

    How come Australia vs India 2001 didn't figure in your calculations? Laxman and Dravid scoring huge tons after following on.

  • POSTED BY Sriram on | December 9, 2017, 7:46 GMT

    I have a question about the basic premise of this exercise - the way WinIndex is being calculated. By considering the part of the test only from the beginning of the 3rd innings seems to miss the case where one team is already down! A comeback in the third innings from a very poor performance in the first two innings will never get a place in this kind of analysis, but it would still be a comeback. It feels like the situation at the end of the first two innings is not being taken into account at all!

    Of course, I maybe wrong - maybe the actual calculation of the WinIndex probably uses the data from the first two innings to come up with the prediction in the first place. Still, I'm not fully convinced!

    Interesting list of tests however. :-)

  • POSTED BY Amey Kanade on | December 9, 2017, 7:40 GMT

    This list seems incomplete without mention of probably the greatest test victory in Indian Cricket - Kolkata 2001.

  • POSTED BY drvish3223668 on | December 9, 2017, 7:08 GMT

    Why is the India v Australia classic of 2001, VVS Laxman's test not included here?

  • POSTED BY ambian6198981 on | December 9, 2017, 6:47 GMT

    A precise example of over doing stats for everything. Probably the greatest side in test history coming on the back of 16 consecutive test wins against a team that was heavily over dependent on one batsman and who has already been dismissed, of the 2 batsmen at the crease , one averages 28 at that point n the other was being questioned for going off color when playing against tough opposition, potentially one wicket away from losing the series n they both bat almost 4 sessions unbeaten. Even after all that, the great team had to just bat out 75 overs for a draw against a bowling attack whose main bowler had earlier planned to move to USA for a career in driving. No stat can get the context of this game , undoubtedly the turning point of Indian cricket . The greatest come back for me of all time , hands down .

  • POSTED BY Pankaj on | December 9, 2017, 6:17 GMT

    9. Australia v West Indies in Adelaide, 1992-93 Australia won by one run In the above match, West Indies won by one run.
    Thank you. Since corrected.

  • POSTED BY Cricinfouser on | December 9, 2017, 6:12 GMT

    Very interesting article. West Indies won the match by one run which is on serial no. 9

  • POSTED BY MuhammadAtifLone on | December 9, 2017, 6:09 GMT

    One test match that I remembered is India vs Pakistan where Kamran Akmal and Abdur Razzaq managed to hold tight and make it a draw. Is that match is among some where in the ratings above?