Z-score’s Cricket Stats Blog
The longest-running cricket stats blog on the Web
Charles Davis: Statistician of the Year (Association of Cricket Statisticians and Historians)
some remarkable first-class innings, re-scored.
Detailed scores for all Tests from 1877 to the 2000s have now been posted. Almost three-quarters of Tests include ball-by-ball coverage; virtually all others offer some degree of extended detail, beyond anything previously made available online.
In the two-Test series between Sri Lanka and South Africa in 2006, famous for its 624-run partnership between Kumar Sangakkara and Mahela Jayawardene at the Colombo SSC, Jayawardene won the Player of the Match awards in both Tests. It seems very strange, then, that the Player of the Series went to Muttiah Muralitharan.
27 November 2023
Disrupted Careers of the 1970s
In the 1970s, there were a couple of major disruptions to international cricket. The first was the exclusion of South Africa from Test cricket from 1970; the second was the World Series Cricket tournaments from 1977 to 1979, which took a large number of players out of Test cricket until late 1979.
This curtailed the careers of quite a number of important players. But as it happens, some substitute matches were played; there were two five-match series involving a ‘World XI’, in 1970 and 1971-72, and of course the Packer World Series matches. There were 16 five-day Packer ‘Supertests’ played, eleven in Australia and five in the West Indies.
I have gathered the scores of these matches, combined them with players’ Test records and produced the following combined averages…
Some Careers Incorporating ‘World XI’ and WSC Matches (batting)
There are some whose batting averages improved when these matches are incorporated, including Sobers, Ian Chappell and Viv Richards. The most striking is Greg Chappell, who made seven centuries in his 17 ‘other’ Internationals, lifting his average to 55.6 in 104 matches. Among players who have played more than 100 Tests, only Kumar Sangakkara and Steve Smith have higher averages. In terms of average, Chappell has a clear advantage over any of his contemporaries with the exception of Sobers.
Another point of interest is the performances of Barry Richards and Graeme Pollock. Pollock’s performances in the additional matches were somewhat moderate, and this takes his average from 61 in 23 Tests down to 54 in 31 matches. This is a case of ‘regression to the mean’, experienced by nearly every batsman who had an average over 60 after 20 Tests. Richards performed very well in the additional matches, keeping his average above 60, but his total of 14 matches is still too small for reasonable statistical interpretation.
I haven’t prepared a table of bowlers, but two cases are particularly interesting. Dennis Lillee took no fewer than 91 wickets in 18 additional matches, including a famed spell of 6 for 0 at the WACA. This takes him to 446 wickets at 24.2 at this level, which would vault him above Kapil Dev and Richard Hadlee in the wickets list. Lillee’s average of 5.07 wickets per match surpasses any other pace bowler with more than 200 Test wickets.
There is also the case of Mike Procter, who played in seven additional matches and took 41 wickets at 15. This is even better than his 29 wickets at 20 in his 9 Tests. The total comes to 70 wickets at 17.1 in 16 matches; this is still not enough for statistical rigour, but it certainly highlights what a loss to international cricket he was. Procter took 1417 wickets at 19.5 in first-class cricket.
I might add a little table concerning regression to the mean. Among those who played 50 Tests of more, there are ten batsmen whose batting averages exceeded 60 after 20 Tests. Every one experienced a fall in average by the time they played 50 Tests…
Although they lost a little of their early pace, Bradman and Sutcliffe held onto their very high averages quite well, while others lost a lot of ground.
Jack Hobbs had a 20-Test average of 57.1, the highest by anyone who increased his average in his next 30 Tests (to 61.3 after 50 Tests). Viv Richards also increased his average, from 55.6 after 20 Tests to 57.7 after 50.
During the World Series (Packer) season in 1977-78, Dennis Amiss became the first batsman to don a helmet in a senior cricket match. I had thought that this had happened in the aftermath of a serious injury to David Hookes in a match against the West Indies XI on 16 Dec 1977. However, Amiss had already been using his helmet (a modified motorcycle helmet) in warm-up games, starting with a match against West Indies on 24 Nov 1977. On the first day of that 4-day match, Amiss scored 81.
The first use of a helmet in a Test match was by Graham Yallop at Bridgetown on 17 Mar 1978. Yallop scored 47.
That first use by Amiss had been in a match at Football Park, Adelaide; apparently the match was not televised. There is a video online of Amiss, wearing an early helmet, facing Andy Roberts; perhaps that was in a later game.
24 October 2023
A Brief History of the New Ball
More than a decade ago I wrote a review of new ball use in Tests. I am repeating it here, but with a few updates/corrections…
The very early Tests seem to have used a single ball for each innings regardless of length. New balls could be called for if the condition of the ball deteriorated severely: this happened after 207 overs in the Adelaide Test of 1891-92. During the epic Australian innings at The Oval in 1884 of 311 four-ball overs, the ball was replaced after 225 overs, reportedly “a very rare occurrence”. Brodribb in Next Man In (1952) records that in Australia the idea of taking a new ball when 200 runs had been scored was introduced in 1901. However, th`ere is also evidence that a new ball at 200 runs was applied in 1894-95, including the first Test, but was not used in all Tests.
England followed in 1907. The 200-run trigger appears to have been kept in use until 1945.
It wasn’t entirely satisfactory. Sometimes, to avoid a new ball, teams reduced scoring before 200 runs were up. In 1946 the MCC introduced an over limit. Strangely, they settled on 55 overs, an extremely low number that favoured pace bowlers. In Australia, the 200-run limit remained in place in 1946/47, but was switched to 40 (eight-ball) overs in 1947/48. This was of no help to the touring Indian side facing Lindwall and Miller.
In 1949 some common sense returned and the trigger was lifted to 65 six-ball overs or 50 eight-ball overs for the next few years. By 1954/55 this had been abandoned in Australia and the 200-run trigger returned. All the recorded new balls of the 1954 and 1955 series in England were taken over 200 runs, but an over limit seems to have been reintroduced soon after; 75 in combination with 200 runs, whichever came first. By 1962 new balls in England were being taken at 200 runs or 85 overs.
There is also some uncertainty about this period in other countries. In the West Indies, 75 overs seems to have been used when the MCC toured in 1960, but 200 runs when India toured in 1962. The known record for use of an old ball is 185 overs at Bridgetown in 1962, but since India scored only 187 runs in that innings, the use of the old ball was not a matter of choice. Some other Tests may have used a combination of runs or overs, whichever came first. In the Australian tour of West Indies in 1965, some new balls came at 200 runs and others at 75 overs, but when England toured in 1968 no new balls were taken before 75 overs, even when the score was over 200.
In India, new balls up to 1965 were generally after 200 runs. There were some exceptions in India and Pakistan when matting wickets were used, with mention of new balls at 150 or 165 runs. The switch to 75 overs was probably in 1965.
In 1965, the runs scored standard in England and Australia fades away and the MCC established a standard in England of 85 overs, or 65 eight-ball overs in Australia (and other countries). This remained in use for many years in these countries, but again other countries had local variations. New Zealand and South Africa followed the MCC standard, but 75 overs seems to have been the norm in the West Indies and the subcontinent.
Finally in 1995, all countries lined up with the same standard, with the new ball available after 80 overs; this remains in place.
Bowlers taking 2 wickets in an over most times (Tests).
There is some uncertain data before 1999, but I think I have covered just about all cases for bowlers like Wasim Akram. The prominence of spin bowlers reflects the fact that spinners are more likely to harvest tailend wickets than pace bowlers.
Best bowling average of career coming in their last Test…
Perhaps surprising, but there are more than 150 such bowlers who have played multiple Tests, including RGCE Wijesuriya, who finished with a bowling average of 294.0 (he took his only wicket in his last Test).
Among those who played 20 or more Tests and took 10 or more wickets there is only
FS Jackson (24 wickets @ 33.3)
RF Surti (42 wickets @ 46.7)
EJ Barlow (40 wickets @34.1)
Bhuvneshwar Kumar (63 wickets at 26.1)
Not including players whose careers are continuing.
Finally, a Complete Picture of the Madras Tied Test
Gulu Ezekiel in India has now sent to me high-quality scans of a complete score of the Tied Test in Madras/Chennai in 1986: a score previously missing, as I have mentioned often enough over the years. Gulu obtained the scans through S. Giridhar and V.J. Raghunath, the authors of From Mumbai to Durban (2016). The original score resides in the office of the Tamil Nadu Cricket Association.
I have completed the re-scoring of the score into ball-by-ball form. (Previously this was available only for the final innings, and much of that was based on an inferior copy.) The re-score mostly went well – but only mostly.
A summary of the findings follow. Note, this is an analysis of the score, not a match review. [The short version – the re-score agrees with the final result, but not with absolute certainty.]
The score format is one innings per page, containing conventional recording of batting strokes, bowling analyses and Falls of Wicket. There is also a table giving the score at the end of each over (this is very useful but, frustratingly, sometimes inaccurate). In the bowling section, ends of bowling spells are marked (again, not 100% accurately) and byes and leg byes are not marked, so their location has to be inferred. Overs are not numbered, so when a bowling spell ends, identifying the next bowler often requires trial and error, guided by the over-by-over score and upcoming strokes in the batting section.
The handwriting changes between the pages; there appear to be two different hands. The scorers are named as G. Ganesh and Chittibabu. The latter is named only for the teams’ second innings. K.S. Mani is also believed to have scored the Test, but his name is absent; perhaps he was making another score.
The pages list the batsmen’s Balls Faced (mostly but not always accurately, and they vary from the Australian published data). BF data cannot be easily gleaned from a conventional score. A linear score is needed; this suggests that the score may be based on additional material of this type. The implication is that the score may be a re-copy from a linear original, or maybe the scorer was writing both types of score, or maybe one scorer was using a linear method while the other made a traditional score.
The score also allows full identification of session scores (lunch and tea). Some of this data was previously unknown; the reporting style at the time, even when quite detailed, tended to dispense with such niceties.
Some comments on the re-scoring each innings…
Innings 1, Australia 574-7 decl. A long innings that eventually yielded a sensible re-score. The only final sticking point was just before tea on the first day, when Boon and Jones were batting. A single by Boon off Yadav does not fit unless moved slightly, and is preceded and followed by leg byes. This preserves the batsmen’s scoring strokes; either that, or the single was credited to the wrong batsman, and Boon actually scored 121 and Jones 211. Some small areas of the page were obscured or indistinct.
UPDATE: Sreeram has suggested a better resolution of this problem that avoids moving strokes around. It does presume errors for a few overs in the over-by-over score section of the scoresheet, but it also produces balls faced figures in the re-score that are more consistent with the scoresheet.
Innings 2, India 397. The score of 397 was preserved, but there were problems. After tea on the third day, two overs by Bright, his 13th and 14th, are marked as ending a spell. As written, these overs do not fit with the batsmen’s scoring strokes and throw the innings ‘out of kilter’. Eventually, I found that if the next over by Bright, his 15th and ostensibly from his next spell, was inserted before his 13th over, then the innings clicked back together neatly (it felt like suddenly solving a Rubik’s Cube). Here is a screenshot of the overs in question. Bright is the fourth bowler.
To make the re-score work, Bright’s Over 15, it appears, should be placed before Over 13. Curiously, Over 15 is written using a different pen to all of Bright’s other overs.
A second problem is seen in Kapil Dev’s innings of 119. While the score does indeed say 119, Kapil’s scoring strokes actually add up to 121 !
The re-score of the bowling also produces 119 runs for Kapil; there are two extra singles in the batting scoring strokes that cannot be found in the bowling, including the last single. In the above screen shot, Kapil’s intermediate scores of 50, 67 and 93 are correct; the problem occurs toward the end of the innings, which should read 2111221 not 211112211.
There are no missing singles from the other batsmen. Re-scoring preserves both Kapil’s 119 and the total of 397, which is important in such a supremely close Test match. However, the existence of errors like this is unsettling.
Innings 3, Australia 170/4 decl. This innings is written less neatly than the others. While the given score can be reproduced, a few overs suffer from readability problems and require a certain amount of ‘interpretation’ to preserve this. In the score, Greg Matthews is given 49 balls faced. This is impossible: the actual number from the re-score is 25. Once again, an unsettling error. (The Australian report gives Matthews 25 BF.)
Innings 4, India 347. With the improved-quality scan now available, a couple of minor problems with the earlier re-score can be resolved. These occurred in the first few overs, and the innings now hangs together pretty well. The final session is based on Lawrie Colliver’s analysis of the surviving video. A couple of minor observations with this innings:
The scoresheet gives the fall of the 2nd wicket as 159, not 158. The re-score agrees with the ‘official’ 158. The scoresheet has balls faced figures that differ from the Australian version (this is true of all four innings). Contrary to what I wrote last year, Greg Matthews did not bowl his 39.5 overs unchanged. His marathon spell was interrupted by one over, the last before lunch, bowled by Steve Waugh. Strangely, the score gives a total innings time of 342 minutes, but this is not possible. A time of 409 minutes, in published reports, is much closer to the mark.
Finally, a comparison of balls faced figures from the final innings: the Australian source, the score and the re-score. Only the re-score adds up correctly to the number of balls bowled (523).
Madras Tied Test: Final innings balls faced, by source
A link to the revised series page in the Database is here.
A bowler called Shehan Mudushanka (DSM Kumara) of Sri Lanka took a hat-trick with his last three balls in his only ODI, at Mirpur on 27 Jan 2018.
The term ‘Chinaman’, referring to an unusual left-arm googly, is said to refer to Ellis Achong, who played for West Indies in the 1930s, the first Test cricketer of known Chinese descent. However, the term predates Achong, and was used in England and Australia in the 1920s, including by Cardus in 1926. Historian Roland Bowen wrote the following:
"This ball, by the way, was not named after the West Indian Chinese bowler, E. Achong, who used it; the term arose much more crudely in the more arrogant days of the past when it was the custom to make snide remarks about foreigners and Asiatics and orientals. 'Chinamen' as they were called, were thought to be 'wrong 'uns' and this delivery is the left-hander's 'wrong 'un' in cricket parlance, and hence the term.".
I have seen the term mentioned in relation to Roy Kilner, a left-arm orthodox spinner for Yorkshire in the 1920s. He sometimes bowled wrist-spin.
A note in Bill Frindall’s score of the Lord’s Test of 1995: “Female streaker – vaulted both sets of stumps – 2 bouncers”. (West Indies 2nd innings, 99th over: the wicket of PJ Martin fell next ball)
9 September 2023
The Luck of Labuschagne
One thing said about the career of Marnus Labuschagne has related to good fortune: he seems to benefit from dropped catches more than most. I have taken a closer look at this, and come up with some odd results.
I have updated my dropped Test catch data to include the recent Ashes (this data now goes back more than 20 years). So how does Marnus look? We I can confirm that for much of his career his rate of catches and stumpings missed off his batting has been well above average. Said average is typically 25-30 per cent chances missed: Marnus is around 33 per cent. It was more extreme early in his career. Here is a list of Labuschagne’s miss rate by calendar year…
At the end of 2022 the Marnus miss rate stood at an extreme 38 per cent, higher than the whole-career leader Virender Sehwag on 37. In 2023, however, there was a sign of regression to the mean. Marnus has been missed only twice in 16 innings so far this year, and gained a mere two runs from the chances. At Delhi he was dropped on 33 but out soon after on 35. At Headingly, he was also dropped on 33 but did not make another run.
Based on the numbers of runs gained, Marnus’ average would drop from 53.4 to 39.8 if he had never benefited from a missed chance. Of course, this is not realistic: everyone gets dropped from time to time. But if Marnus was dropped 25 per cent of the time (something that happens to quite a few batsmen) instead of 33 per cent, his batting average would drop to about 50, from 53.4.
Marnus has been having an indifferent year by his standards, 718 runs @ 38. The absence of runs gifted through dropped catches may well have contributed to that.
The ‘odd’ data mentioned earlier cropped up when I compared Labuschagne’s career to Steve Smith and David Warner. Turns out that the other two are dissimilar in terms of missed chances. Smith remains even higher than Labuschagne at 34-35 per cent, while Warner is down at 23 per cent. Why is Smith dropped far more often than Warner? In fact, why is Smith dropped so often? Well… I don’t rightly know.
A large part of the difference appears to be dropped catches by wicketkeepers.
Three Australian batsmen - dismissals by keepers
As I said, I don’t rightly know the explanation for this. Part of it might be random variation, part of it might be more batting to spinners by Smith and Labuschagne; the rate of keepers dropping catches standing up to the stumps is greater than when standing back. Perhaps Smith is better at playing with ‘soft hands’, which can make edged chances lower and more difficult.
Another oddity adds to this: Warner has only ever been stumped once in Tests, but I have recorded seven instances of a missed stumping, as described by Cricinfo’s commentaries. Smith and Labuschagne by contrast have been stumped seven times and missed five times.
The drop rate for all three is far more similar when it comes to slips and gully catches – drop rates for all three are in the range 29 per cent to 33 per cent.
The usual caveats apply when doing stats on missed chances. The data is prepared by me; others might get different results, although I hope not significantly so.
Many thanks to Gulu Ezekiel for locating and sending a copy of the entire score for the Tied Test in Chennai/Madras in 1986-87. I have been seeking this for decades – literally. More on this when I have finished re-scoring the score into ball-by-ball form.
I have been doing some repair work on some ball-by-ball records in my Test match database, specifically for the years 2003 to 2005. The problems lay in the identification of extras, almost always restricted to the 101st over of an innings, which was due to an old glitch in a data conversion spreadsheet. I had repaired this glitch for many Tests some years ago, but apparently I had only done so up to Tests in 2003, and had forgotten continue forward. The actual runs scored by teams and individuals was not affected; only the identities of extras was sometimes wrong.
John Campbell of the West Indies has played 20 Tests. Remarkably, considering the Windies often weak recent efforts, Campbell has managed to hit the winning run five times. Only five players have hit the winning run more times in Tests, led by Ricky Ponting with nine. Each of those five leaders played 75 Tests or more.
By my count, Steve Smith has now scored 577 runs off Stuart Broad’s bowling, thus passing 571 by Pujara off Lyon for the most ‘head to head’. Broad has got Smith 11 times so the batting average of 52.4 is respectable for both players.
I have updated the list in the Unusual Records section. I have also added a list of the most ball bowled head-to head, which is (perhaps) surprisingly different. Less surprisingly, the list is dominated by Ashes Tests, since these have always been the most common Tests between two teams.
Lyon bowled only three balls to Pujara in the Oval Test in June, conceding one run.
4 August 2023
Going Out with a Bang
There has been discussion about Stuart Broad, who has become the first player in Tests to both hit the last ball he faced for six, then take a wicket with his last ball bowled. The former feat is much rarer than the latter; only Wayne Daniel of the West Indies is a clear-cut precedent for a six off his last ball. Glen Maxwell is also on the list, although he could conceivably play again, as could Broad (very unlikely, though).
The number of bowlers finishing bowling careers with a wicket is much longer, and probably a lot longer than most people would expect. There are more than 130 names currently. I have attached a link to a list of those bowlers that I know of.
Why is the list longer than might be expected? Remember that most Tests won by runs or by an innings margins will have at least two bowlers who take a wicket with their last ball, and sometimes this will prove to be their last match. For most bowlers on the list, it was not known in advance that this would be their last ball. We remember the relative few who announced a retirement in advance – Lillee, McGrath, Murali, Broad – but don’t notice so much the ones who get dropped or otherwise never bowl again.
The first list is up to 2020 and I think is fairly secure. There is a possibility that a few may yet play again.
There is a second list from 2020 to present. The majority of these will play again; I will leave it to readers to decide who might or might not. Even Broad could conceivably play again (to my annoyance he was certainly bowling well enough to continue).
There is a possibility that some have been missed. For Tests with no scorebooks, there may be some who took a wicket with the final ball of their last over, but the innings continued.
Not all bowlers
did this in their last Test. The strangest is Mark Boucher who played 63
Tests after taking his one and only wicket.
A few years ago I worked out that only Gerry Hazlitt of Australia and Godfrey Lawrence of South Africa took wickets with their last two balls in Tests. Hazlitt took 5 for 1 off his last 17 balls.
I wouldn’t guarantee every last detail of these lists.
Most runs before a new ball was taken
England declared at 524-4 in 82.4 overs against Ireland earlier this year with no new ball being taken.
Highest score at the end of the 80th (six-ball) over is 529/5 by England in the above match against Pakistan.
In the past I have seen lists (including my own) of the most overs bowled without a new ball, but I don’t think anyone has published a list of the most runs scored. As can be seen, there is a lot of variation in the scoring rates in the table, with ‘Bazball’ rather revolutionising this aspect.
Some best-forgotten all-round performances by captains in Tests
Pat Cummins had a lousy match in the Old Trafford Test, but just how bad was it?
Qualification: batted and bowled in the match. Fewer than 10 runs scored in the match and 0 or 1 wickets taken for 120+ runs.
None of these players took a catch. It’s hard to rank the above so they are listed chronologically. Botham is the only captain to score fewer runs and have worse match bowling figures in the same Test than Cummins’ recent effort. (Courtney Walsh once took 0 for 123 in a Test but did not bat.)
Longest innings by Australian batsmen who failed to reach double figures
NB: First Innings only. (Extremely slow innings are more common in match-saving second innings situations.)
Dodemaide was batting at #9, trying to support his batting partner on an extreme turning pitch.
The longest such first innings on record is 8 off 124 balls by Bill Edrich against West Indies at Lord’s in 1950. Terry Jarvis’s 9 in 123 minutes at Chennai in 1964-65 may have topped this, but balls faced is not available.
Other batsmen have been slower in reaching double figures, but I am restricting this to batsmen who did not make it to ten.
11 July 2023
More on the Madras Tie!
I have commented before on the frustration of being unable to find original scorebooks for the two Tied Tests, in Brisbane (1960) and Madras/Chennai (1986). Sustained efforts by myself and others to find these have drawn a blank.
Last year there was a breakthrough. A video of the final session of the Chennai match was obtained by Lawrie Colliver. Lawrie re-scored this ball-by-ball, as reported on this blog.
Now another advance. It turns out that a facsimile of a score of the final innings at Chennai was published in a book in India in 2016! Contacts Gulu and Sreeram alerted me to this. The book is From Mumbai to Durban: India’s Greatest Tests by S Giridhar and VJ Raghunath, who had obtained the score from the Tamil Nadu Cricket Association. When I asked around, I could not locate this book in Australia – the biggest collections in the country were unable to help – but Amazon had one copy available, so I bought it.
The image of the score was in colour but quite small (11 cm x 8 cm), and barely readable. Like most pictures in books, it was printed in halftone, creating a confusing pattern under magnification. I was able to descreen and sharpen a high-resolution scan to improve the readability. Another problem was that the writing was not particularly neat – there was no fixing that. On the upside, there was a section of the score that gave the score after the end of each over, not always readable but quite useful. The following shows a magnified version of a small part of the page, from the high-resolution scan
Eventually I was able to re-score the first two sessions of this innings, which could then be attached Lawrie’s score for the final session. There were various difficulties, and a few uncertainties remain. An example would be the first ball of the innings, which is no more than a blob in the score but which must represent two runs, to preserve the known scoring. The batting section for Gavaskar and Srikkanth is also indistinct. Also note the third ball of Matthews’ sixth over. It is not a 6, but a “C” for caught (Srikkanth).
However, in the final analysis the re-score holds together quite well with minimal fudging. I can say with ‘partial’ confidence that the final score of 347 is accurate, and while apparent errors can be found, there no clear errors that affect the final score. The rumours of errors in the scoring are not supported, at least in this innings. The teams would have known throughout the day how things stood in relation to the target (presuming that the scoreboard was up to speed).
My contacts are
endeavouring to get hold of a better copy of the score, including the other
innings. There could even be other Tests from that source, since a number of
Chennai Tests from the 1980s and 90s are not represented by surviving scores.
Strangely, the score appears to give an innings time of 342 minutes, but this is not possible. A time of 409 minutes, found elsewhere, is much closer to the mark.
I have posted an updated article on this final day here.
A full version of the Remember that the original image is only 11 cm by 8 cm.
The updated data page, with the ball-by-ball re-score, is here.
Most runs in an over without a boundary:
There were a couple of cases of 15 runs in the days of eight-ball overs. At SCG in 1936-37, O'Brien and Fingleton hit 2,0,3,1,3,0,2,0,1,3 off Sims, including two no balls that were scored off. That would count as 17 runs in modern counting.
At SCG in 1963-64, Lawry and O'Neill hit a Pollock over for 0,3,2,3,3,0,2,2.
For s six-ball over, there was Tavare and Botham off Holding at The Oval in 1984 3,1,wide,3(no ball),0,2,3,1 = 14 runs or 15 by modern counting.
This is a ‘where known’ record.
In the Test at The Oval, Pat Cummins bowled the first ball and also hit the winning runs. There are only 15 other Tests where the same player did both, and the Cummins instance is the first where both balls went for four.
In the Women’s Ashes Test, England lost the match in spite of having a double-centurion (Beaumont 208) and a bowler with 10 wickets (Ecclestone). There is no parallel for this in Men’s Tests. At Old Trafford in 1896, Ranji scored 62 and 154* and Tom Richardson took 13 wickets, but England lost by 3 wickets.
28 June 2023
Oval Declaration Goes Pear-Shaped
In the Oval Test, Ben Stokes declared England’s first innings at 393 for 8 on the first day – a first in a five-day England/Australia Test – even though his best bat, Joe Root, was on 118 and in full flight. The theory is that putting the opposition in late in the day (to face 10 overs in this case) stresses the batsmen and creates an advantage greater than the runs foregone. So I’ve done a little study...
Question: does having to start a team innings just before stumps harms that team's performance?
I looked at 1000 recent Tests, match innings 2 or 3 only. From this, I filtered out 2 sets:
A. Teams that started the innings 10 overs or less before stumps, in days that lasted more than 75 overs. (188 instances)
B. Teams that started the innings in the first 10 overs of a day (239 instances).
I calculated the scores, runs per wicket and 1st wicket partnerships for the two sets and compared them.
Conclusion: there is no statistical difference between these cases; any differences are negligible. In fact the two datasets are so similar that it is almost spooky.
There is no statistical disadvantage in being forced to start an innings late in the day. However much openers dislike having to start late in a day, it does no real harm.
I did not distinguish between innings starting as a result of team dismissal or declaration, but I don’t think that it would make any difference.
Anyway, England’s ultimate losing margin of two wickets speaks for itself.
A couple of comments about the colour of cricket balls…
Red balls are centuries old, and probably predate any published scores.
There was an experimental night cricket game on a football field under lights in England in August 1952, using a ball painted white, in a Testimonial match for Jack Young. Apart from that, white balls were first used in a World Series (Packer) One Day match at VFL Park in Melbourne on December 14, 1977.
There were a couple of experiments with 'mustard' coloured ball in day/night Sheffield shield matches in 1994-95. A pink ball was first used in a day/night Shield match at the Adelaide Oval in March 2014 (South Australia v NSW). (There had been a pink ball, possibly painted, used for a charity promotion in a women's daytime match in 2009).
There is an
article that says that a white ball was used in a World Series one-day match
in Adelaide in November 1977 (prior to the Melbourne match). It was at the
end of a 4-day ‘make work’ match for Packer players who were not involved in
Packer’s main weekend match. The match finished a day early, so they played a
fill-in one-dayer on the final day. However, the article gets both the ground
(it was at Football Park not Adelaide Oval) and the date wrong, and newspaper
reports from the time do not mention any white ball or night play. The
reports do mention, however, that fielding circles were used for the first
time in that match.
I was away on holidays in central Asia for a while (returning with a broken ankle!). I will still post from time to time, perhaps with decreasing frequency.
I am continuing to post updates to my Test match database, which has reached 2005. Who knows how far it will go, even 2010 seems a long way off.
9 June 2023
The Fastest Pair
Newspaper reports have the first innings ending at 4:25 and the second innings at 6:25. Ward would have been out for a second time about 10-12 minutes before the latter, making his ducks about 108-110 minutes apart, including the break between innings. (The newspaper clock times would be to the nearest 5 minutes only.)
It appears that after Ward was out the second time,
the last 25 runs were scored very quickly in about 3 overs. The hat-trick
over was WWW044 and Matthews' next and last over was 230400. Kelleway took
the last two wickets, in the 3rd last and the last overs.
I also did a search for the shortest interval between two dismissals of the same batsman, regardless of score. The record here was 11 minutes’ play, plus 16 minutes between innings by Percy Sherwell in the 5th Test of 1910-11 (SCG). Sherwell batted #10 in the first innings and was last out. He kept the pads on and opened when the follow-on was enforced, and was first out. There is also the case of Joe Darling at the SCG in 1901-02. Darling was 9th out in the first innings and first out on following-on. His elapsed time was probably closer to 30 minutes than Sherwell’s 27 minutes (there is no surviving score for this Test). Farook Engineer, in a Test in 1964-65, was also out twice in a very short span, probably just over 30 minutes.
A few comments on long sixes, to add to my 2016 article…
Possibly the longest six at the SCG was by Alan Davidson for NSW v MCC in 1954-55. Hit forward of the left-hander’s square leg, it struck high on the roof of the old Brewongle Stand. In an article in The Cricketer in the 60, Ray Robinson gave a figure of 113 yards, plus 40 feet off the ground, measured with a steel tape. Perhaps 125 yards or more, given a 45 degree descent angle. However, there is a description of this shot in the Sydney Morning Herald from that time, and it does add that the wicket position was slightly on the western side of the ground, and so favoured Davidson. Checking with Google Earth does suggest that Robinson’s figure may be a slight exaggeration. I would say 115 metres, with 120 metres unlikely.
The longest six in the 2023 IPL was 115 metres by Faf du Plessis. It is intriguing that the seven longest sixes in the IPL (>117m) were all hit prior to 2014. What has happened to the 120-metre hits?
Ashley Chandrasinghe batted all day for 46 runs in 90 overs on the first day of the Sheffield Shield Final in Perth, raising the question of who has scored the fewest runs in a complete day’s play. It may be in the Sheffield Shield (no one seems to have suggested any alternatives) but in England Hashim Amla of all people went from 0 to 37* on the 4th day of Hampshire v Surrey quite recently in July 2021. Wisden doesn't mention any interruptions, and there were 96 overs bowled. Amla faced 278 balls.
I get 13 instances of left-right combinations in all 10 partnerships of an innings, most recently the West Indies 2nd innings against India at St Lucia in 2016. Curiously it first happened in 1884 at The Oval, when there were fewer left-handed batsmen. On that occasion Scotton, a left-hander, batted through most of the innings with the right-handers, with the only other leftie, Emmett the tailender, coming in for the final partnership when Scotton was out.
In the final Tests at Ahmedabad, Ravi Ashwin, a spinner, bowled the first ball of Australia’s second innings to Matthew Kuhemann, normally a specialist spinner. In the I thought this must be unique until I found Pallekele 2018 where Jack Leach opened facing the bowling of Dilruwan Perera, in a very similar situation.
There were also a few early Tests where it is possible, but bowling styles can be difficult to unravel. Hugh Trumble opened facing Len Braund in a Test in 1902 (4th Test SCG), which seems to be fairly clear-cut.
In the Christchurch Test against Sri Lanka, New Zealand scored 257 runs without a break in play on the final day, after the first two sessions had been washed out. This is the most runs in a session by a single team, even if it stretches the definition of a 'session'.
The session lasted about 230 minutes, which is just a little short of the 241 minutes for the final session at Melbourne in 1998-99.
At the Oval in 1884, there were 259 runs after lunch on the final day. Back then there were no formal tea breaks on most days, including this one; there was a change of innings though, which lasted about 17 minutes.
The most in a more conventional session is 249 by South Africa against Zimbabwe in 2005, a 155-minute session. The most in a 2-hour session is still 236 by Australia at Johannesburg in 1921.
The Christchurch Test was only the second to reach a decisive result on the last possible ball, the other being the first Test at Durban in 1948-49. By coincidence, England also needed 8 runs off the last over (although it was an 8-ball over) of that match, and they won the match by 2 wickets with an extra off the last ball. It was a four-day match.
Two matches have been won with one ball to spare - Port of Spain 1934-35 (West Indies def England by 217 runs) and Leeds 2014 (Sri Lanka def England by 100 runs). The two Tied Tests also finished with one ball to spare. Two other Tests have been left drawn with the scores level at the end of the match.
3 April 2023
Bowler v Batsman ‘Hat-Tricks’
There are probably more cases. The first that I found, Ray Lindwall dismissing Alec Bedser, was spread across 2 series, so Bedser played quite a number of innings in between without facing Lindwall. Most recent is Kyle Jamieson dismissing Roston Chase in 2020-21.
Some of these 'hat-tricks' do not involve any golden ducks. Even those that do are complicated. In 2000 Andy Caddick got Curtley Ambrose three balls in a row that he bowled to him, the last two being first ball, but in between those ducks Ambrose played innings of 36* and 1 without facing Caddick.
Surprisingly perhaps, Ajit Agarkar is the only batsman in Test history to make 3 consecutive golden ducks. He went on to a fourth consecutive golden duck, and five dismissal in six balls, but he does not appear on the list because different bowlers were involved.
At the other end of the scale, here is the updated list of the most runs scored by a single batsman off a single bowler. Pujara off Lyon now leads, a sign of the relatively high frequency of India/Australia Test matches in recent times. Although the number of runs is very high, the contest can be regarded as rather even, given that Lyon has now taken Pujara’s wicket 13 times. Steve Smith may have a chance to challenge this total in the upcoming Ashes, although he will have to bat well even if Broad get to play most of the Tests.
Individual Batsman v Bowler: Most Runs, all Tests
All the above figures are reasonably exact. The highest total where estimates are necessary is Javed Miandad v Kapil Dev, estimated at 470.
The relevant section of the “Unusual Records” section has been updated.
I have started working again on the Test Match Database, continuing (slowly) Test series from 2003. I hope one day to get the database up to about 2010.
Cheteshwar Pujara has now scored more thqn 560 runs off the bowling of Nathan Lyon in Tests matches, taking the #1 position for a batsman off a single bower. I will update the list when the current Test in Ahmedabad is over.
I have completed
a rather large project – to scan and digitise all the paper scores for Test
matches that I have collected, over 20 years or more. There are scores from
more than 1550 Tests in the collection; quite a few of these Tests are
represented by more than one score, and there are 30 more represented by
ball-by-ball records from other sources.
Ball-by-ball records from Cricinfo and other online compilers are not
I plan to make copies so the material can be preserved in other hands. There are even a few scores in there for which the originals have gone missing since the copies were made, including the oldest one of all, the 1880 Test at the Oval. When I visited the Oval a few years ago, that score could not be found.
7 March 2023
A Quick Visit to DRS
While I have it in front of me here are some broad stats on DRS, up to Jan 2023...
Tests with DRS: 447
DRS called on: 4795 times
Umpire decision overturned: 1285 times = 2.87 per Test
303 not out decisions overturned by DRS, out of 1935 reviews.
511 OUT decisions overturned out of 1638
181 not out decisions overturned out of 608
161 OUT decisions overturned out of 273
Overall, there are more umpire decisions overturned in favour of batsmen than bowlers. However, the net effect is fairly small, averaging about one less dismissal every 2.4 Tests.
although DRS can have a major effect on individual innings, and on close
matches in critical situations, its broad effect on statistics is not great.
My opinion of the pitches being
produced in India for the Test matches, particularly the Indore pitch…
Pitches don’t have to be like this. In ODIs in India, teams regularly get 350 in 50 overs (4 times in 6 matches in this season alone). It shows that respectable pitches can be made in India, although in some cases it goes too far the other way, in favour of batsmen. There is also plenty of big scoring in Indian first-class cricket – some of the pitches must be ok.
Test pitches have long favoured spinners in India; that’s all right, but this season it is going much too far. When Hayden scored plenty of runs in 2000-01, he played lots of sweep shots as I recall; it negated the spin. Now that is impossible because of the variable bounce.
Anyway, India was hoist on its own petard at Indore. Perhaps we will see an end to such extremes in the future.
I took a screenshot of a closeup of the pitch after the first day at Indore, from the TV broadcast. Martin Briggs wondered if it had been taken by Neil Armstrong on the moon. More like Mars, I think.
Most ducks by a team in a first-class match…
In the famous MCC v Australians match at Lord's in 1878, lasting only one day, MCC (33 & 19) suffered 13 ducks. This was topped by "Muslims" versus "Europeans" at Poona in 1915, with Muslims (21 & 39) suffering 14 ducks.
4 March 2023
A Player of the Series in 1951
At the end of the 1950-51 Ashes series, Len Hutton was given an award as best player, calculated using a runs/wickets/catches formula. The award, sponsored by the Vok liquor company, was worth one thousand pounds. Such a sum was big money in cricket in those days, when Australian Test players were lucky if they could make ten pounds per day. (The winner of The Open golf championship in 1950 won £300). Apparently, when the existence of the award had been announced earlier in the series, the players unanimously decided to share the prize equally, and this is what Hutton did.
Still, the award as it stood was effectively a Player of the Series award. I don’t know of any similar (single player) awards being made until John Edrich won Player of the Series in the 1968 Ashes; and even then the concept did not come into regular use until the late 1970s (Ian Botham in 1978). There were other Player of the Series awards in England in the late 60s, generally one for each team. Curiously, England returned to the double-award approach in 1986, and this continues, but in other countries the award is usually to a single player.
A small breakthrough. I found a full score for the 4th Test of 1881-82 (Melbourne), and it was online! Years ago, someone had told me that this Test score was included in the 1882 Australian England tour book, which survives at Lord’s, and he sent me a copy, but only the first innings of the match. I assumed that that was all that was available. It turns out that the whole match score exists in that source, and a microfilm version was copied for a project for the National Library of Australia. Digitised versions are online here. And here.
This is the oldest surviving Test match score from Australia. I have re-scored the whole match now and posted corrected versions of the match data. As with many early scores, re-scoring this one was a bit of a struggle, but I hope the rendering is reasonable.
The 1882 team had already assembled in Melbourne by the time of that 1881-82 Test, and they sailed only three days after the match. Perhaps because of the imminent departure, the match was left drawn even though it had been scheduled as timeless. It was the last drawn Test in Australia until 1946-47.
Harry Brook’s diamond duck in England’s one-run loss at Basin Reserve had me looking for cases of ‘first ball’ diamond ducks, that is batsmen run out without facing a ball, and from the first ball bowled after they came to the crease. Most Diamond duckers were at the wicket for a few balls before being run out, so the list is quite short…
“First Ball” Diamond Duck
Brook is the first batsman to score a century and diamond duck in the same match. Brook has now scored 809 runs off 819 balls, and is well on track for the fastest first one thousand runs in Tests, in balls faced. Tim Southee is currently the fastest with 1132 balls, ahead of Colin de Grandhomme second.
Four instances in ODIs of a pair of batsmen reaching a century off consecutive balls…
RS Dravid SC Ganguly Ind v SL, Nagpur 22-Mar-1999
SR Waugh MG Bevan Aus v SAf, Docklands Stadium 16-Aug-2000
G Gambhir V Kohli Ind v SL, Kolkata 24-Dec-2009
KC Sangakkara TM Dilshan Sco v SL, Hobart 11-Mar-2015
It is curious
that there are no known instances in Tests, given that there are two cases of
a century and double century off consecutive balls. At Durban in 1999, Gary
Kirsten reached 200 off CJ Adams and Mark Boucher reached 100 off the next
ball. At Adelaide in 2012-13, Michael Clarke reached 200 off Imran Tahir, and
Mike Hussey reached 100 off the next ball.
batted both left-handed and right-handed in a Test innings…
Talat Ali, Adelaide 1972-73 (left-handed to right-handed)
had hand/arm injuries.
A reminiscence triggered by Australia losing two batsmen retired hurt in the Melbourne Test against South Africa…
I was at the ground the day that the three West Indians went RH. There were 53,000 at the SCG, what a day. My brother and I got there an hour before the start of play, but had to sit on the steps of the stand, which was already full. The noise was immense, peaking as Thomson bowled the ball; it sounded like baying for blood. I greatly admired the concentration of Lawrence Rowe, scoring 67 without helmet or much of the modern protection. Thomson (3-117) was dangerous but inaccurate.
Funnily, my memory says that Thomson inflicted all three injuries, but, on checking, Holding was actually hit by Greg Chappell's bowling! Holding retired on the last ball of the day. The next morning Holding wanted to return, but was disallowed. Julien instead returned to the crease, which was not permitted at the time, but the umpires overlooked that.
I also went to the fourth day of that match. We sat side-on to the wicket. When Thomson (6-50) was bowling, I simply could not see the ball once it left his hand.
Thomson also took a great outfield catch on the first day,
9 February 2023
Dropped Catches Report for 2022
I have done a
search for Test match dropped catches (and missed stumpings), along the lines
of previous years, using the Cricinfo ball-by-ball texts. This study now
spans back more than 20 years. For 2022, the analysis was assisted by Garry
Morgan, who kindly sent me his list of dropped catches.
most striking thing about 2022 was the poor catching of Australia, rising
from 18% missed in 2021 to 31% in 2022, which is by far the worst single-year
performance recorded for Australia in this century. I had picked up 52 misses
by Australia, so many that I went back and checked them all, but all of them
seemed reasonable. The only caveat is that 34 of the 52 misses were regarded
as difficult chances, which is a higher proportion than normal (which is
around half). There were 26 % drops off Australian batsmen in the same set of
what they say about Marnus Labuschagne seems to be true; he gets dropped more
than almost anyone else. Labuschagne has been dropped off 38% of the chances
that he has offered, more than anyone who has played over 30 Tests, except
for Taufeeq Umar (also 38%). Virender Sehwag saw 37 % of his chances dropped;
that at least could be explained by the fact that he hit the ball harder than
anyone else. The explanation for Marnus eludes me. If the drop rate for
Marnus had been normal, he would lose three or four runs off his batting
There is a ‘tussle’ at the top of the bowler’s list, with Stuart Broad retaking the lead over Jimmy Anderson. Broad has had 136 chances dropped to Anderson’s 132: surprising, perhaps. Note that this is 21st Century data; we don’t have complete numbers for the likes of Muralitharan or Warne. Next after Anderson is Harbhajan on 99 misses, but again data is not complete here, since Harbhajan started out in 1998. However, it is possible to extrapolate the missing data to produce an estimate of 115 missed chances in total for Harbhajan.
Who attended the most Test matches in person?
In 2003, Richie Benaud said he had attended 486 Tests. We don't know the exact number after that but the total will be around 550. That is around 25% of all Tests up to the time of his death, or 31% of Tests that were played while he was actively attending them.
No one is particularly close to Benaud in number of Tests. Pakistani journalist Qamar Ahmed has over 450, while John Woodcock, who died last year, was also well over 400.
Bill Ferguson scored 204 of the first 382 Test matches. His scoring career extended from WG Grace to Garfield Sobers. There had been 443 Tests played when he died.
At the Gabba, South African bowler A Nortje bowled two deliveries in one over that went for 5 wides. I can only find one precedent (including 4 wides): Neil Johnson of Zimbabwe against England at Trent Bridge in 2000, in the 104th over.
Johnson is only debited with 2 wides in that innings, but both went to the boundary.
Thinking about king pairs, I wondered if there was a bowler’s equivalent, that is a bowler taking a wicket with his first ball in each innings of a Test.
Turns out to be a very short list and rarer than a king pair.
MC Bird Eng v SAf (1), Johannesburg 1909/10
Z Khan Ind v Ban (2), Dhaka (Mirpur) 2007
AR Patel Eng v Ind (3), Ahmedabad (Patel) 2020/21
Zaheer Khan’s effort was as an opening bowler, inflicting a king pair on Javed Omar.
The last time a Test finished in 2 days and had more than 33 wickets, as in the Gabba Test v South Africa, was Port Elizabeth 1895-96 (40 wickets). Two other Tests lasting 2 days had 40 wickets: The Oval 1882 (the Ashes Test) and Lord's 1888. The Oval 1890 had 38.
One of the Tests of 1912 had 37 wickets on the first 2 days, but it went into a third day.
These Tests were all limited to 3 days.
Seeing two batsmen retiring hurt on the same day (Warner and Green) in the South Africa MCG Test had me reminiscing about the day that three West Indians retired hurt in 1976…
There were 53,000 at the SCG, what a day. My brother and I got there an hour before the start of play, but had to sit on the steps of the stand. The noise was immense, peaking as Thomson bowled the ball; it sounded like baying for blood. I greatly admired the concentration of Lawrence Rowe, scoring 67 without helmet or much of the modern protection. Thomson was dangerous but erratic.
Funnily, my memory says that Thomson inflicted all three injuries, but, on checking, Holding was actually hit by Greg Chappell's bowling! Holding retired on the last ball of the day, mid-over. The next morning Holding wanted to return, but was disallowed. Julien instead returned to the crease, which was not permitted at the time, but the umpires overlooked that.
I also went to the fourth day of that match. We sat side-on to the wicket. Even though we were in the front row on the fence. when Thomson (6-50) was bowling, I simply could not see the ball.
31 December 2022
What is a Fair Qualification?
When comparing player careers and many other stats, it is routine to place qualifications on the data, particularly minimum matches or runs or etc. List the highest career batting averages, and Kyle Patterson (144.0 in two Tests) is at the top, ahead of Andy Ganteaume, but apply a reasonable minimum and the list comes out rather differently.
But what is a reasonable qualification for such stats? I have seen lists of highest batting averages that variously have players like CS Dempster, SG Barnes, RG Pollock, and GA Headley right near the top (below Bradman of course). On the other hand, I have seen qualifications so high that even Bradman is excluded.
I looked at the question by taking the population of recognised batsmen (about 90 of them) who have played 80 or more Tests and looked at their career progression in terms of ‘average average’ and standard deviation at different stages of their careers. The aim here is to look at how the spread of the stats levels out as careers get longer – it is not necessary to include shorter careers with this aim in mind.
Batting average in general is relatively steady, but with an upward trend probably due to those great players who had very long careers, and were still reaching their peak after they had played 50 Tests.
interest is the standard deviation of the batting average. Early on, averages
are quite variable, but as careers progress, there is ‘regression to the
mean’, the lesser batsmen tend to improve (those who don’t get dropped and
don’t play 80 Tests) while the high flyers come back to the pack. After ten
Tests, TT Samaraweera had an average of 83.0 and Mushfiqur Rahim 16.2. They
finished their careers with averages of 48 and 38 respectively.
This has the
effect of excluding some notables like Graham Pollock (60.97 in 23 Tests) and
George Headley (60.83 in 22 Tests). I think that that these careers were just
a bit too limited to get a clear fix on their abilities. Chances are that
they would have regressed towards the mean if they had enjoyed longer