Batting July 30, 2010

Chalk and Cheese: a look at the two halves of Test innings

An analysis of batting performances in the two halves of a Test innings

Bradman's 270 was rated the best Test innings by Wisden © Getty Images

It is the responsibility of the first 6 batsmen in a Test innings to score the required runs and the low order batsmen, normally the bowlers, to provide support. There are times when it happens the other way around. The low order batsmen score more runs than the top order. There is an inherent charm and excitement in these innings. Often these also turn out to be match-winning innings. More often than not one of the top order batsman stays on to shepherd the late order. It could also be that these are true cases of innings revival controlled by genuine late order batsmen. In this article I have taken a comprehensive look at such innings.

I may be wrong. However there is only one innings in test cricket in which, for strategic reasons, a captain sent his entire low order first on an a "gluepot" of a wicket, and then he himself came on to play one of the greatest Test innings ever. This match is discussed later. So this is the only innings in which the late order was expected to outscore the top order.

First some summary facts. These are current up to match no 1965, the second Pakistan-Australia match.

Number of innings played: 6187 (Maximum-7860)

Number of innings played in which the late order (wkts 6-10) has out-performed the top order (wkts 1-5).

All tests:  1431 (in 6187 innings - 0.23 times per innings).

Pre-WW2: 206 (in 883 innings - 0.23 times per innings). Pre-WW1: 120 (in 454 innings - 0.26 times per innings). WW1-WW2: 86 (in 429 innings - 0.20 times per innings). 1948-1969: 251 (in 1242 innings - 0.20 times per innings). 1970-1989: 340 (in 1426 innings - 0.24 times per innings). 1990-2010: 634 (in 2636 innings - 0.24 times per innings). 1990-1999: 264 (in 1087 innings - 0.24 times per innings). 2000-2010: 370 (in 1549 innings - 0.24 times per innings).

Overall there has been an average such occurrence of around 0.23 per innings (slightly below once a test). There have been distinct moves away from this mean value of 0.23. During the pre-WW1 period, at times of uncovered pitches, wide disparity in batsmen techniques, "gentlemen" teams playing etc., the low order had to come to the rescue of the top order more often, around 0.26 times per innings. Then came the batting era, with top batsmen playing, the move was in the other direction. Either side of the WW2, the number oscillated around 0.20 per innings. Bradman, Hammond, Headley before and Hutton, the three Ws, Sobers, after. During the period 1970-1989, the figure picked up to 0.24, just above the mean. Then from 1990 to 2010, the figure has oscillated at around the mean of 0.24. Even when I split up the period into two halves, the figure has not changed. It can thus be concluded that barring the Pre-WW1 and either side of WW2, the frequency has remained at around 0.24, during the past 40 years.

The complete table of 1431 innings is available for view, import and analysis for users. The table is in reverse chronological order. To view/down-load the complete table, please click/right-click here.

How do we view the information in a summary form in this analysis. I have created three summary tables for viewing. In all tables I have also shown the highest scorer and the batting position he batted in to give an idea of who coordinated the revival.

The first is a table ordered on the ratio between the second half runs and first half runs. For the selected innings this value is 1.00 or more. In this table I have selected only innings in which this ratio is 4.00 or more. There is no rocket science in this number. It is a high enough number to limit the number of table entries to a reasonable number. Also we are pushing up the bar. Anyhow this is only a cut-off for display. By definition these will be innings such as 25 for 5 recovering to 200 all out or 100 for 5 recovering to 400+ for 8 and so on. It is more likely we have low scoring innings in this table. Let us look at the table.

Year MtId Bow Bat R 5Wkts Final <2nd Half> Highest Score
e Score Score Runs To1Hf Runs(BP) Batsman

1.1952 0354 Eng Ind 6/5 98 ao 92 15.33 38 ( 4) Hazare V.S 2.1995 1306 Slk Pak 15/5 212 ao 197 13.13 117 ( 7) Moin Khan 3.2004 1683 Zim Bng 14/5 169 ao 155 11.07 61 ( 8) Khaled Mashud 4.1935 0239 Win Eng 23/5 258 ao 235 10.22 85 ( 8) Holmes E.R.T 5.2005 1765 Ind Zim 18/5 185 ao 167 9.28 52 ( 5) Taibu T 6.1898 0056 Eng Aus W 32/5 323 ao 291 9.09 188 ( 3) Hill C 7.1888 0030 Eng Aus 7/5 70 ao 63 9.00 32 ( 6) Lyons J.J 8.1967 0623 Eng Pak 26/5 255 ao 229 8.81 146 ( 9) Asif Iqbal 9.2008 1875 Win Aus W 18/5 167 ao 149 8.28 79 ( 7) Symonds A 10.2000 1520 Aus Win 22/5 196 ao 174 7.91 96 ( 7) Jacobs R.D

Before any reader comes in with his comment, let me confess that this is an odd table. The ratio depends on how quickly the first 5 wickets have fallen rather than on how many runs were scored by the last 5 wickets. However, having set out the base methodology, I did not want to exclude any innings based on an artificial lower limit for the innings size.

The highest ratio reached is 15.33 when India recovered from 6 for 5 to 98 all out, assisted by Hazare. Readers should remember that if the sixth wicket had fallen soon after, India might not have reached 26, the record low total of New Zealand. The bowling attack was a fearsome one, viz., Trueman, Bedser, Laker and Lock. Pakistan's recovery, controlled by Moin Khan, is lot more substantial, with a ratio of 13.13. However the innings which catches one's eye is the Australian recovery from 32 for 5 to 323 all out, orchestrated by Clem Hill's 188, which is in the top-10 of the Wisden-100 innings table. Recent recoveries have been led by the two keepers, Khaled Mashud and Taibu.

It can also be seen that very few of these tests are likely to be won, considering the low-score nature of recovery. Hill's innings was one of the successful ones and Symonds, which was in the second innings. 18 out of 76 have resulted in wins.

To view/down-load the complete table, please click/right-click here.

The second is a table ordered on the number of runs added by the last 5 wickets during the selected innings. In this table I have selected only innings in which the runs added are 300 or more. By definition these will be innings such as 150 for 5 recovering to 475 all out or 300 for 5 moving on to 700+ for 8 and so on. It is more likely we have high scoring innings in this table. Let us look at the table.

Year MtId Bow Bat R 5Wkts Final <2nd Half> Highest Score
e Score Score Runs To1Hf Runs(BP) Batsman

1.1955 0414 Nzl Pak W 87/5 561 ao 474 5.45 209 ( 8) Imtiaz Ahmed 2.1937 0257 Eng Aus W 97/5 564 ao 467 4.81 270 ( 7) Bradman D.G 3.1955 0406 Win Aus 233/5 668 ao 435 1.87 137 ( 5) Miller K.R 4.2009 1911 Eng Win 334/5 749/9 415 1.24 291 ( 3) Sarwan R.R 5.1966 0609 Win Eng W 130/5 527 ao 397 3.05 165 ( 4) Graveney T.W 6.2010 1953 Bng Nzl W 158/5 553/7 395 2.50 189 ( 5) Guptill M.J 7.1972 0695 Nzl Win 171/5 564/8 393 2.30 183 ( 5) Davis C.A 8.2005 1774 Eng Pak W 247/5 636/8 389 1.57 223 ( 4) Mohammad Yousuf 9.2009 1933 Ind Slk 375/5 760/7 385 1.03 275 ( 4) Jayawardene D.P.M.D 10.1996 1336 Zim Pak 176/5 553 ao 377 2.14 257 ( 8) Wasim Akram

This is a more interesting table since it is ordered on the number of runs added.

At the top of the table, Imtiaz Ahmad, batting at no.8, scored 205 and helped Pakistan recover from 87 for 5 to 561 all out.

The next match is an all-time classic. The innings by Bradman was determined to be the best ever Test innings in the Wisden-100 exercise. Australia's 200 for 9 was countered by England with 76 for 9, on a diabolical pitch. Then Bradman countered by sending his low order batsman, to let the pitch dry out. These batsmen promptly lost their wickets, but consumed valuable time. Bradman walked in and scored 270 to take Australia to 564 and a comfortable win. It was a tribute to Bradman the tactician as much as Bradman the batsman.

The most intriguing innings is by Wasim Akram who scored 257 at no.8 and took Pakistan from 176 for 5 to 553 all out in the company of Saqlain Mushtaq.

More tests in this table are won since the recovered innings score is almost always in excess of 400. 28 out of 59 have resulted in wins.

To view/down-load the complete table, please click/right-click here.

The third is a table ordered by the final score reached, but with a different criteria for selection. I have selected only innings in which the ratio is 2.50 or more and 200 or more runs are added by the last 5 wickets. This is done to ensure that we get a representative population of truly great late order batting performances. By definition these will be innings such as 150 for 5 recovering to 450 all out but not 7 for 5 to 70 all out nor 375 for 5 to 760 for 7. This table is likely to contain the really relevant innings. Let us look at the table.

Year MtId Bow Bat R 5Wkts Final <2nd Half> Highest Score
e Score Score Runs To1Hf Runs(BP) Batsman

1.1937 0257 Eng Aus W 97/5 564 ao 467 4.81 270 ( 7) Bradman D.G 2.1955 0414 Nzl Pak W 87/5 561 ao 474 5.45 209 ( 8) Imtiaz Ahmed 3.2010 1953 Bng Nzl W 158/5 553/7 395 2.50 189 ( 5) Guptill M.J 4.1966 0609 Win Eng W 130/5 527 ao 397 3.05 165 ( 4) Graveney T.W 5.1955 0406 Aus Win 143/5 510 ao 367 2.57 219 ( 7) Atkinson D.S.t.E 6.1908 0098 Eng Aus W 135/5 506 ao 371 2.75 160*( 9) Hill C 7.1925 0160 Eng Aus W 118/5 489 ao 371 3.14 201 ( 7) Ryder J 8.2002 1594 Nzl Eng W 106/5 468/6 362 3.42 200 ( 5) Thorpe G.P 9.1976 0784 Pak Nzl 104/5 468 ao 364 3.50 152 ( 7) Lees W.K 10.1984 0975 Nzl Eng 115/5 463 ao 348 3.03 164 ( 7) Randall D.W 11.2008 1857 Ind Aus W 121/5 463 ao 342 2.83 162 ( 6) Symonds A 12.1931 0209 Nzl Eng 129/5 454 ao 325 2.52 137 ( 7) Ames L.E.G 13.2005 1759 Zim Nzl W 113/5 452/9 339 3.00 127 ( 8) Vettori D.L 14.1983 0972 Win Ind 92/5 451/8 359 3.90 236*( 4) Gavaskar S.M 15.1994 1264 Eng Saf 105/5 447 ao 342 3.26 104 ( 6) Kirsten P.N 16.1970 0675 Eng Aus 107/5 440 ao 333 3.11 171 ( 5) Redpath I.R 17.2001 1566 Bng Zim 89/5 431 ao 342 3.84 94 ( 6) Wishart C.B 18.2006 1824 Eng Aus W 84/5 419 ao 335 3.99 156 ( 7) Symonds A 19.1984 0997 Aus Win W 104/5 416 ao 312 3.00 139 ( 7) Dujon P.J.L 20.1981 0907 Aus Eng W 104/5 404 ao 300 2.88 118 ( 7) Botham I.T

Truly a list of the greatest recoveries ever. We make our acquaintance with Bradman, Imtiaz and Hill again. We should admire the recent double hundred of Thorpe. Gavaskar's 236*, against Marshall, Roberts, Holding and Davis, at the unusual batting position of 4, must compete with his swan song classic of 97 for being considered his best innings. Symonds has played two such innings.

Since this test balances the ratio and runs added measures, the number of tests won in these matches is also likely to be on the higher side. 34 teams have won out of 87.

To view/down-load the complete table, please click/right-click here.

I would appreciate if readers download the master table, import into an Excel sheet and come out with nice nuggets of information including country-wise numbers. These would be published with due acknowledgement.

The one thing that strikes me at the outset is that there are very few such recoveries by India, barring the one led by Gavaskar. I am not sure whether this indicates a lack of quality of the Indian late order batsmen or the strength of top order batting or a combination of both. Surprisingly, West Indies and Pakistan have many such recoveries.

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

Comments have now been closed for this article

  • testli5504537 on August 29, 2010, 11:13 GMT

    What a shame that a lovely piece of research like this needs to be updated before it is a month old! In the Lords test against pakistan, England recovered from 47/5 to 446 all out. The last 5 wickets added 399 runs, which was 8.49 times the total added by the first 5 wickets. This means the innings would feature in all three of the tables provided here: it would be in 9th position in Table 1, 5th position in table 2, and 3rd position in Table 3. It would be the ONLY innings to feature in all 3 tables. It is arguably the greatest recovery of all time. [[ Stewart I doff my imaginary hat at you. Wonderful of you to remember the article and provide the update. I was looking for the single which would have given the 400 run mark for the last 5 wickets. It is amazing that a high innings like this would have broken into the first table. When England were 47 for 5, let me confess I thought of my chalk and cheese article and assumed that 200 would have been a great recovery. But 446. The stuff which makes all of us switch to Test cricket. As far as recoveries are concerned, I would place this as the best, ahead of Pakistan's 1955 recovery simply because what happened afterwards. Pakistan only went to 111 for 6. England went to 103 for 7. And, the unsavoury controversy notwithstanding, let us not forget that the Pakistani bowling attack at Lord's was a most potent one. Once again, many thanks. Ananth: ]]

  • testli5504537 on August 11, 2010, 10:50 GMT

    Perhaps more interesting is the number of times a team has won a test with these "negative split" innings. I seem to recall that Australia with Gilchrist at number 7 or even Symonds at 7 and Gilchrist 8 due to nightwatchmen, seemed to recover when teams thought they were on the ropes.One example being the Pak test at Bellrive that saw Gilchrist join Langer to take Oz from less than 150/5 to score 349/7 to win!!

    In fact, wasn't there a series of Oz in SL where in every test Oz was behind on 1st innings but won due to 2nd long innings from Lehman (2x) and Martin with the tail?

  • testli5504537 on August 3, 2010, 9:28 GMT

    Ananth - I also don't recall a single century without a false shot (perhaps only 15-20 ball cameos are fully error-free) although Bradman reckoned his 232 to be one such innings. Anyway, the paradox is that an analysis aims to give a complete picture of reality whereas reality, philosophically it being the loss of ego, mandates no analysis. So, the best analysis is done simply by witnessing the game moment by moment as it unfolds.

  • testli5504537 on August 2, 2010, 22:04 GMT

    Ananth - Taking away the runs scored after dropped catches or poor umpiring decisions (unless the decisions are real shockers) etc. is best left for whiners. E.g., this view penalizes a batsman who was dropped at the 2nd slip but does not penalize another who unintentionally edged a sitter through a vacant 2nd slip region.

    Still, it is good to bear in mind whether such blemishes happened because the purpose of analysis is to impart a better insight. E.g., it is good to remember that Gibbs dropped Waugh and accept the 120 as an all-time great ODI innings without deducting any points from it. Likewise for SMG's 236*; later on, even Marshall praised it. (However, I don't think it merits a place in all-time Top 100. His 90 in the same series feels better. Interestingly, he got out on 90 to a Holding beauty in parts due to a distraction caused by someone at the sight-screen). [[ Agreed, Alex As I mentioned earlier let the asterisk be in the viewer's mind rather than being a part of the analysis. It is also my firm feeling that when a batsman bats for a day so (and scores between 100 and 200, depending on which decade he batted in !!!), he would get at least one life. It may be a poor umpiring decision, a simple drop, a bad fielding error or a defensive captain (Sangakkara in this series-remember Sehwag's escape at short leg). When trhe innings becomes larger and longer, this will multiply. Ananth: ]]

  • testli5504537 on August 2, 2010, 13:31 GMT

    Ananth, Thank you for your great reply. Just put a point across that some folks inc. me feel. I realise that it is impossible to clearly delineate "poor decisions", "dropped catches" etc. Just that when "talking cricket" with the guys and when great innings come up you often hear things like "well yeah, but what about those 3 sitters dropped?" This problem will not arise with those older inn. which not many ppl have seen.But is rather common with the modern day classics.

    However, like you say there is no way to clearly categorise these things. They are nice to discuss! Thats all.

  • testli5504537 on August 2, 2010, 6:38 GMT

    Looking at some recent matches Pak should try this "reverse batting order" trick every time and all the time, irrespective of the pitch.

    I sympathise with what Alex has to say rgd. dropped catches etc. Ofcourse we cannot take away further runs scored after a dropped catch/poor decision etc from a batsman. However, a flawless, chanceless innings should rank above a "flawed" one. Just that half point or whatever should be deducted for such events to distinguish them from the chanceless classics. [[ Abhi These things are nice to discuss. Who knows when which batsman was dropped. Some drops are notorious such as Lara's 1x (cost - about 485) against Durham or More's miss off Gooch when he was on 30 (cost 303) etc. What about all other misses. 99% of these are not documented. It would be quite unfair to the guys whose innings you asterisk because you know the miss while 99% of the other misses remain "unasterisked". I saw every ball of the Laxman classic. I am sure there were couple of 5x-4x decisions early on the fourth day which could have gone the other way, with no major complaints. However, my point is that these minor blemishes in no way would or should bring down the value of the innings. Frankly I cannot think of one 100% pure virgin-white innings. Oh you could point one to me. I could find a hole if i tried hard enough. Perfection is Tendulkar's overall attitude, demeanour, technical excellence, temperament and the unfailing courtesy towards all. It is not a single error-free innings. Ananth: ]]

  • testli5504537 on August 2, 2010, 4:17 GMT

    [[ Anon: I have not published this mail on Gavaskar's 236 since it was unsigned and no mailid was given, in other words completely anonymous. Even though the point might have been correctly made. Chetan Asher: I would appreciate if you do not send such biased jingoistic mails to this blog. There may be other blogs interested in publishing your mails, not this one. Since I do not want to communicate with you, I am using this space. Ananth: ]]

  • testli5504537 on July 31, 2010, 12:18 GMT

    Ananth - Re your comment that Bradman started with 4 late-order batsmen in Melbourne 1937: O'Reilly, Fleetwood-Smith and Ward were tailenders, but Rigg wasn't - his batting position in that series ranged between 2 and 5. [[ Andrew Since Rigg played 8 tests, scored 401 at 33, he cannot really be classified as a tail order batsman. So I concur with you. Ananth: ]]

  • testli5504537 on July 31, 2010, 12:07 GMT

    Ananth & Rob: re Ananth's puzzlement on why 1-2-3 did not take strike in WI 2nd innings. Wisden says the following. Hoping that the conditions might improve, G. C. Grant altered his batting order and the 2nd day ended 33/4. More heavy rain fell during the night leaving the pitch waterlogged, and not till half past three could cricket be attempted next day. On a still difficult wicket, 3 more batsmen departed for the addition of 18 runs before tea. During the interval, G. C. Grant, in turn, adopted the bold policy of declaring, leaving England 73 runs to get for victory. (Somehow, Constantine did not play this match.)

    Anyway, the strategy of sending tail-enders up the order seems to inspire great 270's. Bradman hit a match-winning 270 in the innings it was implemented while the last test of this Eng-WI series saw Headley hitting a match-winning 270* (see Test #241)! [[ Alex Thanks for throwing additional light on Rob's fantastic discovery. Ananth: ]]

  • testli5504537 on July 31, 2010, 11:00 GMT

    Two comments and both of them came when i checked India vs WI 1994 match which figures in your third list.

    1) Is there anyway to eliminate innings where bowlers batted in the top 5 ? Then we get a more realistic picture of lower order revival. Though a great innings, Bradman's actually wasn't a lower order revival with bowlers as we understand it. In the afore-mentioned match, Manjrekar walked out at 7 with India at 88/5. Mongia had come out at 4.

    2) Instead of the innings highest scorer, the more meaningful is the highest scorer considering only partnerships from 6-10 ? A batsman could have got out 6th after a century and the majority of the scoring can be done by tailenders after that. [[ Yogesh 1. My gut feel is that out of the 1431 innings, very very few would have been because of the reverse batting policy. Almost all the instances would be cases of top batsmen failing and low order batsmen rescuing. So simplicity is sometimes the best method. 2. I thought of doing this. However what happened was that many a top innings (such as Clem Hill's 188) which was the cornerstone of revival got shunted out. I understand what you are trying to stay. I should get the top scorer out of the batsman at the other end of the fifth wicket and batsmen 7-11. This is possible for recent matches. However for well over 100 years there is no clear data on who got out at the fall of each wicket. Ananth: ]]

  • No featured comments at the moment.