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 the1990s 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.
Have there been any two-day first-class matches since 1947? Since 1947 all first-class matches are supposed to be 3 days or more. However, there can be exceptions. A match between NE Transvaal and Australians in 1957-58 was scheduled for 2 days and left drawn. I don't know how it came to be ranked as first-class.
21 May 2022
The Diamond Duck (updated and bumped)
I tried to find early uses of the term ‘Diamond Duck’, as it is most commonly used now, referring to a batsman being dismissed without facing a ball. As a cricketing term, it did not show up in straightforward searches of newspapers before 2011. The following searches were negative: Guardian 1980-2003, Sydney Morning Herald 1980-1995, Times of India 1980-2010, Jamaica Gleaner 1990-2010, British Newspaper archive 1980-2010. Wisden online has a search facility, but that came up negative also.
There is one instance in The Times in 2004, but it defines Diamond Duck as dismissal on first ball of a match.
In my searches, I did get occasional hits for the term, since it is used in Contract Bridge; there was also a mascot for a sporting team in the US known as the Diamond Duck. But nothing in a cricket context in the 20th Century.
I first used the term in an article in 2006. I can't remember where I got it from – perhaps I coined it myself! However, there may be earlier uses of it that others can find.
It is still not universally used in the way that Duck and Golden Duck are. Some people tell me that they recall the term being used in junior or club cricket in the last Century. I can’t say that I remember it myself, and it is worth noting that it is an uncommon event, so would never have been used much. In any case, the fact that The Times defined it differently in 2004 suggests that it was not yet in general cricketing vernacular. It would be nice to find something in print before 2006.
The term was not found in “The Language of Cricket” by Eddowes (1997) or the “Lexicon of Cricket” (2006), or some other volumes that have entries on ducks. It does not appear to be mentioned in Meher-Homji’s entertaining little book “Out for a Duck” (1993). I have an archive of all of Cricinfo’s ball-by-ball texts and match reports from 1999 to 2006, but there is no mention of the elusive “diamond duck”, even when it actually happens, at least in Test matches. Generally, it is described as “run out without facing a ball” or similar language.
There is, however, an occurrence of the term in an ODI report from 2001 (NZ v Zim at Auckland) which describes TN Madondo as the “diamond duckman”. Madondo was indeed run out without facing a ball, so this would seem to be a clear identification, but there is a complication. It was also the first ball of the innings, so it is uncertain whether the commentator was using the ‘Times 2004’ definition, or the more modern definition.
Incidentally, I found an instance of “Golden Duck”, for a batsman out first ball, in a British newspaper in 1963. I was a little surprised not to find it in Australian sources in Trove (which peters out about 1954), nor did it turn up in the Sydney Morning Herald from 1955 until the first hit in 1980. Perhaps others could take a shot at this. It is intriguing that terms like Golden Duck, which seemed so common in the cricketing vernacular in the past, turn up so rarely in published sources.
The first Golden Ducks in Tests occurred in Spofforth’s hat-trick in 1878-79, MacKinnon and Emmett being the second and third victims. The first diamond duck was by W Attewell at the SCG in 1884-85.
ANOTHER UPDATE: Sreeram has found an occurrence of the term diamond duck from 1999. The Age on 28 Oct 1999 refers to “a brilliant run out of Graham Manou for a diamond duck (that is, without facing a ball)”. There is an earlier, somewhat more cryptic, reference in the Sydney Morning Herald in 1997, strangely enough in connection with horse racing, but defined in cricketing terms. Both references go to the trouble of defining the term, suggesting that it is not widely understood.
Sreeram also found an earlier reference to “golden duck”, in 1961, in the Buckinghamshire Examiner. Curiously, this is exactly the same newspaper where I found the 1963 reference.
Every Australian Test team has had at least one player born in NSW. With regards to team representation rather than birth, three teams have had no NSW players, including Perth 1981. The others are The Oval 1972 (two Victorians) and Kandy 1983 (one Victorian).
It is quite striking that there have been 90 Test matches with no Victorian representatives, but only three lacking NSWalers. This lack of Victorians in many teams may be related to the monopoly that AFL has on top athletes and sports in Victoria, which is not the case in NSW.
If I have calculated correctly, Durban 1921/22 had nine NSW players, although one (Carter) had been born in England.
In the Sheffield Shield era, the most Victorians in a Test has been six, most recently Brisbane 1952/53 (including DT Ring born in Tasmania).
Luke Fletcher recently scored a half-century in a County match batting at #11. He followed it up by opening in the second innings and scoring another 50. It is extremely rare for a #11 in one innings to score two 50s in a match, regardless of where they bat in the other innings. The nearest to Fletcher’s double is Robin Jackman in 1982 who scored 60 and 68 for Lancashire against Surrey, at #11 and #3. In the latter innings he came to the wicket as a nightwatchman at 1-1.
I can’t find any cases of a #11 hitting 50s in both innings in first-class cricket. Note, however, that there may be a lot of matches where the second innings batting order is uncertain.
Most non-boundary runs in a Test innings. Several reports give Hanif Mohammad 24 fours in his 337 in Bridgetown in 1957-58, meaning that he ran for 241 out of his 337 runs. However, the old England scorer Geoffrey Saulez examined the scorebook for this Test while in Barbados during the 1970s, and came up with different numbers. He reported 26 fours, 16 threes, 40 twos and 105 singles, which adds up. That would bring Hanif's non-boundaries down to 233 runs.
(It is conceivable that he hit one or more all-run fours, but such shots are rare in the West Indies (about 0.25% of fours in the modern Test game). I don't recall reading about any all-run fours in Hanif’s innings. Sadly, that Bridgetown scorebook is now lost.)
However, Australia's Bob Cowper did have two all-run fours in his 307 against England in Melbourne in 1965-66, bringing his boundary fours down to 18. That would give him 235 in non-boundary runs.
If you include running for partner’s runs, the most running in an innings (more than 10,000 yards) goes to Len Hutton in his 364.
The early Tests involving Bangladesh were very unserious affairs. This can certainly be seen in the two Tests in Sri Lanka in 2002. In the first, Sri Lanka won by an innings and plenty, with Aravinda de Silva scoring 206. The response of the selectors was to drop seven Sri Lankan players, including Aravinda! Some of the replacements were little known; the captain Jayasuriya said he had not even seen them play. Aravinda, for his part, retired from Test cricket, becoming one of only four players who have scored double-centuries in their last Test, and one of only two (with Seymour Nurse) in their last innings.
I am not sure if Aravinda resigned in disgust or was planning to retire anyway; he did continue in ODIs. Sri Lanka easily won the second Test also, but by a reduced margin. Bangladesh did not reach 200 in either Test.
17 May 2022
Sreeram has alerted me to a Twitter post by one “SaikiaArup” which has a ball-by-ball record of the last 13 overs of India’s famous win at Port of Spain in 1976, scoring 406 in the final innings. Made from a radio broadcast at the time. I have extended this by one over (which had a wicket) and posted the data here or here. It encompasses the innings of Brijesh Patel, who scored his 49* off 50 balls. Data like this, from West Indies/India Tests, are few and far between.
It appears that this source has no other similar records; I am still hoping that there must be people out there who scored other matches like this on an amateur basis.
Left, Right, Right, Left
Following a question from Arjun Hemnani, here is data on player v player averages for bowlers of different types against left- and right-handed batsmen in Tests.
f- fast, fm- fast medium (includes medium-fast), m- medium, sp- finger spin, wr- wrist spin. Data since 1999.
I don’t know if there is a great deal to be drawn out of this without over-analysing every little variation. Left-hand finger spinners seem to have quite different records against left and right handers. Left-handed bats do better against pace than right-handers, but worse against right-hand finger spinners.
I was asked a question about most wickets on first-class debut. I came up with 15 by three bowlers, all of them strange cases. The only one in the last 150 years was one Nadeem Malik. Malik took 18 wickets in his three-match career, 15 of them in his first.
A strange feature is that his three matches were spread out from February 1974, then seven months later in September 1974 and one more in 1979. The 1974 matches were for "Lahore Reds" and "Lahore C" and neither team ever played again. The 1979 match was for Lahore Division, a more conventional team.
There is something of a tradition in Pakistan of briefly cobbling together teams and calling them first-class. “Dera Ismail Khan” in that notorious match in 1964 (lost by an innings and 851 runs) is one example. It was the only match played by a team of that name in that era.
The other bowlers who took 15 are William Brown in his only match for Tasmania in 1857, and John Kirwan for Cambridge Town v Cambridge University in 1836. If you ask me, the first-class status of all of these debut matches is questionable.
Here are some simple but eye-opening stats on the number of ODIs played by each country since 1 Sep last year. Illustrates the collapse of (proper) white-ball internationals.
Papua New Guinea 18
United Arab Emirates 11
South Africa 10
Sri Lanka 6
United States of America 6
West Indies 6
New Zealand 3
England – the men’s team anyway – has not played a single ODI since July.
Australia won the three-Test series in Pakistan by winning in the final session of the series after drawing the first two Tests. This has only happened once before (in any country), when England won in Karachi in 2000 with only 2.3 overs to spare after drawing the first two Tests.
Batsmen reaching milestones at low team totals…
Data is incomplete for many early ODIs, but Jayasuriya reached 50 out of 53 (off 17 balls) in an ODI at Singapore in 1995-95. There were 3 sundries with Kaluwitharana not out on 0.
In Tests, Clifford Roach scored 50 out of 54 at Bridgetown in 1929-30, since matched by Chris Gayle (Port of Spain 2014) and Tamim Iqbal (Wellington 2017). Gayle actually hit a two to go from 49 to 51 out of 55 while Tamim hit a three to go from 49 to 52 out of 56.
David Warner reached 100 out of 122 in his century before lunch at the SCG in 2017.
This question comes up from time to time, so I have added a list to my Unusual Records page.
21 April 2022
First-Class FoW; Some New Stats
Some notes on the recording of Falls of Wickets
The listing of Falls of Wickets (FoW), as a part of scorecards, is such a familiar sight that it might be a surprise that for a long time such detail was often not seen in publications. FoW can be found in original scores going back more than 150 years; however, this data did not always make into newspapers or other published reports, particularly in England. The Times did not routinely list FoW for Tests until 1928. While Wisden followed suit for some Tests from the 1930s, it did not routinely report FoW for county matches until about 1952.
In Australia, the practice goes back rather further, with FoW listed, as a defined section, in Test match scores from the beginning of the 20th Century, but apparently not before.
It is notable that a complete set of FoW for all past Tests was included in Roy Webber’s Playfair Book of Test Cricket in 1951. Webber must have done a fair amount of research to put that together, since the data was largely absent from Wisden; he would have been helped by some Australian newspapers. The book’s successors, now the Wisden Book of Test Cricket, have continued with this.
In the wider realm of first-class cricket, data for older matches still contains gaps. There must have been a lot of research to gather what is now known, but still there are perhaps nine per cent of matches before 1970 for which FoW data is absent or very patchy. This includes most matches before the Test era. FoW data is about 99 per cent complete since 1970, but I might add that not all of f-c cricket matches have full scores anyway; for example, some matches in Sri Lanka in the 80s and 90s, a time of civil war and upheaval, are represented by ‘potted’ scores only.
The practice of naming the batsmen out at each FoW, allowing the easy identification of batsmen in partnerships, is relatively a more recent phenomenon. It has only become widespread in the electronic era from the 1990s, although Bill Frindall was ahead of the game in the 70s with his published Test cricket scores for various series, continued in the Daily Telegraph Cricket Year Books in the 1980s. Researchers have gone back through some types of old matches, notably Tests and Sheffield Shield matches, to make such identifications, although I have found that this data can be a bit unreliable; from my own research, I have made over 400 changes to online identifications of batsmen out in Test matches.
Apart from Test and Sheffield Shield, batsman ID in FoW is something of a blank area in first-class cricket before about 1995. Even quite recently, data can be incomplete, with almost ten per cent missing from online scores since 2013, notably for matches in Pakistan.
Statistics of Major First-Class Partnerships
This is something of an introduction to a piece of research that I have done on identifying the batsmen in major partnerships in f-c cricket. The potential field is vast, so I have been limited to looking at partnerships exceeding 200. Even then there are over 7,000 known partnerships. It would be nice to be extend the analysis to all century partnerships, but with over 65,000 to go through, it would be rather overwhelming. It also becomes more difficult to be sure about batsman ID the smaller the partnership gets. You would think that it would be easy for partnerships over 200, and indeed it is in most cases, but even at that level there are so many uncertain ones that time gets consumed.
Apart from matches for which FoW is missing entirely, there is some uncertainty introduced by the incidence of batsmen retiring hurt. This adds imprecision, but I have decided to include these (if they can be identified), and just list the two major batsmen when it happens. It appears to happen in less than two per cent of major partnerships, and in many of those there was a two-man 200+ partnership anyway.
Firstly, the numbers of partnerships. These are the ones I have identified; the numbers will be mostly but not absolutely complete. Bear in mind that the data I have gathered may not represent all known data.
It is not really clear to me why there are fewer opening stands than subsequent wickets.
Number of Known 200+ Stands in First-Class Cricket
I have searched for the lowest scores made by batsmen involved in a (two-man) 200-run partnerships. The list is led by the freak partnership between the Hazare brothers in 1943-44. Vivek Hazare’s 21 in five and a half hours represents probably the most extreme sustained slow scoring in f-c cricket.
Lowest scores by players involved in 200+ stands.
‘Retired hurt’ partnerships excluded.
These are the final scores by the batsman rather than their contribution to the partnership, although I think they may be the same in each of these cases. There may be others who contributed fewer runs to a major stand, but who ended up with more than 40 runs in total. Generally, it is not possible to identify these.
This list does not include the strangest double-century partnership of all, 246 unbroken between M Nayyar (101*) and K Bhaskar Singh (12*) for Delhi in 1991. There were 180 runs included in this stand thanks to penalty runs for a slow over rate (throughout the match) by Bombay – so it is not a real 200 stand, but I leave it in anyway, because it is so curious.
The dataset allows, I think for the first time, identification of batsmen involved in major partnerships.
Most double-century stands in first-class cricket (individuals).
On reflection, it is not too surprising to see Sutcliffe topping Hobbs. Sutcliffe was a slower batsman; in fact in Tests his average innings length rivals Bradman. This means that there was a higher chance of large numbers of runs being added while he was batting.
Filtering for triple-century stands produces this list
Curiously I get only five for Hobbs and four for Hammond. Both tended to score faster than their batting partners, and so accumulate fewer giant stands.
For 400+ stands, I get 3 for Rahul Dravid and 2 for various others. However, one of Dravid’s included a retired hurt.
For batting pairs, I get the following…
Most major stands by batting pairs (all wickets)
The 200+ stands include the 300+ stands.
I think that the dominance of Holmes and Sutcliffe in this statistic was already known; all of their stands were for the first wicket and so could be very easily identified.
There was a lesser-known pair, FB (Frank) Watson and Ernest Tyldesley, who matched Sutcliffe and Holmes in making four 300+ stands, although they only had six 200+ stands in total. I didn’t find any pairs involved in more than one 400+ stand.
Frank Worrell and Ravi Jadeja have both been involved in two 500+ partnerships during their respective careers.
Most ‘dismissals’ with no balls
since 2000: updated list.
This depends on the accuracy of Cricinfo texts and my ability to search them. Apart from ‘caught off no ball’ and ‘bowled by a no ball’, it includes cases of "lbw to no ball" that may be a matter of opinion or uncertainty.
M Morkel 13
KAJ Roach 13
I Sharma 12
B Lee 11
ST Gabriel 11
Z Khan 7
RA Jadeja 7
Wahab Riaz 6
A Flintoff 6
SL Malinga 6
VD Philander 6
K Rabada 6
Wicket with first ball in Test cricket.
This list has one or two differences with other published lists.
TP Horan 1882
A Coningham 1894
WM Bradley 1899
EG Arnold 1903
AEE Vogler 1905
JN Crawford 1905
GG Macaulay 1922
MW Tate 1924
M Henderson 1929
HD Smith 1932
TF Johnson 1939
KR Miller 1945
R Howorth 1947
Intikhab Alam 1959
RK Illingworth 1991
NM Kulkarni 1997
CJ Drum* 2000
MKGCP Lakshitha 2002
NM Lyon 2011
RMS Eranga 2011
DL Piedt 2014
GC Viljoen 2015
*In Drum’s case it was actually his second delivery, following a no ball to start.
Horan was not playing in his first Test, but it was his first time bowling.
Kraigg Brathwaite faced 673 balls in the Bridgetown Test just completed. This is the most known for a West indies batsman in a Test, but there is another contender.
can only be estimated, but Frank Worrell faced a similar number of balls at
Kingston in 1952-53. Some years ago I came up with
an estimate of 674 balls for Worrell (237 & 23); Brathwaite faced 673.
Worrell's actual figure is probably in the 650-700 range, but exact figures
were never recorded.
23 March 2022
Warne and McGrath
Sreeram suggested that I should write something about Glenn McGrath and Shane Warne and their head-to-head records against the best batsmen of their time. For those interested here is the updated data. I can't say it is very flattering to Warne, particularly looking at the very best (Tendulkar, Lara).
It's probably avoided by most commentators, but the reality is that top batsmen generally do much better against spin (even the best spinners) than against pace bowlers. Spinners do have something of a disadvantage in that when top-order batsmen fail, it is sometimes before the spinners come on. Nevertheless, the difference is stark. Warne, it should be remembered, was effectively crowned Greatest Bowler Of All Time by Wisden in 2001 – he was the only full-time bowler among the “FIVE CRICKETERS OF THE [20TH] CENTURY” – when his career still had years to run. Yes he was a great bowler, but IMHO Warne was not even the best bowler in his team.
McGrath and Warne: head-to head against top batsmen
Note: averages are strictly player v player; e.g., Kallis scored 167 runs off McGrath’s bowling, with 6 dismissals, average 27.8, and 354 runs off Warne’s bowling, with 7 dismissals, average 50.6.
I have been looking more closely at Barry Valentine’s ball-by-ball work on Ashes Tests from 1920 to 1961, and made a log of all the dropped catches that were mentioned. For the first time, this allows a long-term view of trends in this very difficult area of statistics. Without further ado, here are some numbers, compared to the recent data that I have been collecting over the last two decades. Percentages are calculated using the numbers of catches and stumpings in the relevant Tests.
Historical Missed Chances (Ashes Tests)
Includes missed stumpings, but not run outs.
A couple of points: reporting styles changed in the 1950s, becoming more interpretive and less rigorous in terms of straight out facts. In some series such as 1956 there is a dearth of reports of dropped catches off tailenders; it may well be that some relatively unimportant chances have been left out of reports. For the 1960s, there is currently very little data. I hope this can be improved in time.
While we should not read too much into every bump in the trend, there is a clear trend of improvement over many years.
There is also some more work to be done on Tests from the 70s to the 90s. The percentage in the 70s may have been around 30%, but more work on that is needed. I have some data from Frindall for these decades but need to work on it more.
I have edited the ball-by-ball records of Ashes Tests from 1920 to 1961 to include dropped catches, and included columns in the half-centuries detail files to include dropped catch data. Examples here and here.
Here is a list of most balls faced by a batsman in the fourth innings of a Test…
8 March 2022
The news here – recently dominated by War (Ukraine), Flood (NSW and Qld) and Pestilence (Covid) – has been pushed aside by the very untimely death of Shane Warne. There are already mountains of articles on the subject; I have little to add, except to notice that Warne’s “larger than life” personality features in commentary at least as much as his cricketing achievements, perhaps more so.
A quiet introverted version of Warne would have had less impact on the game, even with just as many wickets. My own opinion: it is possible to be both a very great player and to be overrated.
Plus a couple of little stats…
· Warne’s name appears on standard Test scoresheets (Batting, Bowling, dismissals section and FoW) more times than any other player (Jimmy Anderson is second).
· During his bowling spells in Tests, Warne’s bowling partners took more wickets than he did. That is, 723 wickets were taken by other bowlers when Warne bowled the previous over, vs 708 taken by Warne himself. Run outs excluded.
· I have a theory that Warne was congratulated, in person and by name, more times than any other person in human history. He got a “bowled Shane!” from Ian Healy and others for maybe 50% of the 50,000 balls he bowled in international cricket. Absolutely impossible to prove of course.
A couple of articles of mine from some years ago on the subject of Warne…
A Tale of Two Spinners. (2006)
Some comments on Shane Warne’s Test career (2007)
When Joburg was in the Wars
Here are some more notes on the extraordinary circumstances and events surrounding the Johannesburg Test of 1895-96. Thanks to Robin Isherwood for information.
The tour and its scheduling was impacted in late December 1895 by what is known as the ‘Jameson Raid’, a military incursion from Rhodesia into Transvaal by supporters of Cecil Rhodes. It was hoped that the raid would trigger an uprising against the Boer government in Transvaal, but it failed in that objective. It was undoubtedly a precursor to the Boer War (1898-1902).
Another difficulty for the cricketers, not entirely unrelated to the above, was the deep antipathy between cricket authorities in the Cape and Transvaal. This is evident from both sides, in comments in newspapers at the time, and there was a lack of cooperation between the two groups. Nevertheless, Lord Hawke’s team ventured inland in mid-February, apparently having scheduled a match against Transvaal for the 22nd.
Then on February 19, while the team was playing a scheduled minor match in Bloemfontein, a gigantic explosion occurred in Johannesburg. More than 50 tons of dynamite, on five rail trucks, detonated in a rail yard, killing scores of people and leaving a crater up to 50 feet deep and 250 feet long. Houses hundreds of yards from the blast were destroyed and the shock wave damaged most of the buildings in the city. One witness described “a vast black and gold cloud rising like a colossal mushroom into the blue”, which must be one of the first descriptions of a Mushroom Cloud. The exact circumstances of the explosion were difficult to ascertain, since everyone directly involved had been killed.
The nearby Wanderers Ground was given over to triage and treatment of hundreds of the injured.
Prior to the explosion, it appears that a Test match had been added to the schedule, either replacing the Transvaal match or, more likely, to be played afterwards. It is not entirely clear, but in any case the match on the 22nd could not take place. The team passed through Johannesburg and witnessed the devastation, and arrangements were made to move on to Pretoria instead. The Test match, possibly slated for the 26th, was pushed back to Monday 2nd March.
When the Test started the buildings around the Wanderers were still being used as a makeshift hospital, an extraordinary circumstance indeed. “The Wanderers Hall, which was actually a pavilion, still stank of iodoform and was full of wounded” (Hayward’s History of Transvaal Cricket).
As a sidelight, Robin tells me that referring to the Ground as the “Old Wanderers” is incorrect, since the ground was never known by that name. Known just as the Wanderers Ground, it was closed to make way for the expansion of Johannesburg railway station in the 1940s. The current “Wanderers” further to the north was opened in 1956 and should be referred to as “Wanderers Stadium”. I am making some changes in my database to reflect this.
wicket sequences across two Tests in 1895-96 in South Africa…
Five wickets in six balls.
Six for 2 in 13 balls.
Seven for 2 in 19 balls.
Nine for 2 in 31 balls.
Ten for 4 in 38 balls.
Twelve for 4 in 44 balls.
These figures are gleaned from press reports; the number of balls for six wickets or more are not necessarily exact.
Note that Lohmann bowled in minor matches between the two Tests, so the figures apply to Test (and first-class) cricket only.
There was drama in Johannesburg between the two Tests. On February 19, there was a massive explosion in a railway yard when as much as 50 tons of dynamite detonated, apparently accidentally; more than 80 people were killed and windows were shattered throughout the city.
The English players were in Bloemfontein at the time. The scheduled tour match against Transvaal in Johannesburg on the 22nd was cancelled and a Test was then scheduled, but it did not start until March 2. This was the first Test match played in Johannesburg, a place that had been open farmland less than a decade earlier.
The explosion was not mentioned in Wisden. There is mention of “disturbances” in Transvaal, but this refers to military skirmishes in December that were precursors to the Boer War. The disturbances had delayed the team’s venture into the South Africa inland.
In this tour, RM Poore was asked by the England manager if he would join the England team to replace HT Hewitt, who had returned home. Poore, though, chose to play for South Africa instead.
21 February 2022
Dropped catches report 2021
I have completed an analysis of dropped catches in Tests in calendar year 2021, drawn from Cricinfo’s ball-by-ball texts. There were 289 missed chances found in 44 Tests.
Overall, 23.9% of chances were missed, including missed stumpings (run outs not included), which was almost identical to the 24.1% in 2019-20 (with fewer Tests, I combined the data for 2019 and 2020). The averages move from year to year, but there has been a slight, if irregular, improvement in catching over the 20 years that I have been doing this analysis. From 2004 to 2008 the average was 25.8%; in the last 5 years, 24.4%.
For 2021, New Zealand had the best catching record, a result that is consistent with widespread impression of that team’s fielding. Although rankings do vary from year to year – sometimes unaccountably – New Zealand, South Africa and Australia have been the top three teams averaged over the last five years, just as they are in 2021.
After several good years, Pakistan slipped considerably in 2021; Sri Lanka was similar. West Indies fielding has been improving. England’s ordinary performance confirms the impression from the recent Ashes series.
A few records from the last 20 years of data…
Batsmen missed most times AN Cook 78 (Sangakkara and Sehwag on 67)
Batsman with highest % misses: J Blackwood 39% (Sehwag 37%, Ross Taylor 36%), minimum 50 chances.
In the field, Cook also missed the most catches with 81. Much of his early career was spent fielding at short leg, which is the most difficult position for taking catches.
Bowlers with most missed chances: Anderson 129, Broad 127 (up to and including calendar year 2021, so not including the last couple of Ashes Tests).
Zulfiqar Babar of Pakistan suffered a rate of 52% of the 58 chances missed off his bowling, including stumpings. Mohammad Rafique had a rate of 44%.
Graeme Smith dropped only 15% of his possible catches in the slips during his career.
Among keepers receiving more than 100 chances, Mark Boucher had a miss rate of 10.3% and AB de Villiers 10.4%. (de Villiers drop rate as non-keeper was a somewhat more typical 21%.) At the other end of the scale, Mushfiqur Rahim missed 31.5% of his 113 chances.
Most misses by a keeper: MS Dhoni missed 66 chances in his Test career (18.3%). Note that keeping to spinners is more difficult, and results in more missed chances, than keeping standing back.
The usual caveats apply with regards to dropped catch data. A chance can be a matter of opinion, and it can be possible to overlook instances when searching the texts. Nevertheless, my search method has been consistent for 20 years (actually it goes back to 2000, but I haven’t analysed all Tests in the early years because the texts sometimes lacked detail).
I mentioned Barry Valentine’s work on Ashes series in my last post. Looking at his work on the 1924-25 series, it became apparent that there were problems with my own analysis, that was carried out quite some years ago and has been online for a few years. I have now corrected the problems and posted new versions of the data.
A core problem was that I was relying on copies of Bill Ferguson’s scores; the originals are kept at Cricket NSW. Although made by Ferguson himself, these scores are handwritten copies made after the event, and there are signs that they were made hurriedly. A good deal of secondary information, including info on separation of bowling spells and session scores, and byes and leg byes, is missing. Fortunately, it turns out that Fergie’s running sheets for this series exist at Lord’s, and copies have even been posted online by the National Library of Australia. Until a few years ago, I was not aware of existence of the 1924-25 linear scores: they are the earliest surviving Ferguson running sheets (earlier ones, going back to 1909, have been lost, although traditional-style scores survive).
Anyway, there are quite a lot of differences between Fergie’s re-copied scores and his running sheets, so going back to the latter has led to changes to some balls faced figures, lunch and tea scores, and ball-by-balls records. Mostly these are not substantial changes, but they are numerous. Even the very first over of the series required changing – the running sheets (0,0,1bye,1bye,1,0,1,0) and the Cricket NSW score (0,0,1,0,1,0,0,0) do not agree. While doing this upgrade, I have also included information on dropped catches, for this series and for 1928-29. For this I have Valentine’s work to thank.
I have also re-done my analysis of the 1958-59 series, again thanks to Valentine’s work. In this case, the score copies that I had obtained many years ago had been very hard to read in places, being multigenerational copies via microfilm. Valentine managed to obtain better-quality copies, and I have used his analysis. Likewise, I have revised the 1954-55 series for similar reasons.
I wonder if this has any parallels: an active first-class cricketer who was murdered during a Test match that he had attended the day before. From Barry Valentine (referring to the last day of the 1st Test of 1920-21)...
"In mourning for the death of Dr. C.J. Tozer, DSO, the flags round the ground were flown at half-mast, and the Australian players all wore black arm bands. Newspapers report that he “was shot dead in a private house to which he was attached in his private capacity”. Claude Tozer had been in the Army Medical Corps wounded in France, and was a leading NSW batsman. In early December at Brisbane he scored 51 and 53 for an Australian XI v MCC. On 20 December he attended the Test. Next day he visited the home of a patient, Mrs. Mort, with whom he had been having an affair, and when he told her was going to marry another woman, she shot him in the head and chest and tried to shoot herself. She was later found not guilty by reason of insanity."
There is a Wikipedia page of murdered cricketers. If I have read it correctly, Tozer was the only one who was an active first-class cricketer at the time of his death.
UPDATE: Sreeram reports the case of Nauman Habib, who was shot three days after playing in a first-class match.
7 February 2022
Over rate in extremis
Some years ago I put together an analysis of the opening partnership of Grace and Scotton at The Oval in 1886 (first day). The stand was worth 170 and when Grace was out at 216, he had also made 170. I recently had reason to take another look at this innings, and I have to admit that I found considerable problems with my analysis.
I have re-done the work and the numbers have changed. (There is no surviving scorebook, so putting together an over-by-over account, from newspaper accounts, is fraught with uncertainty. The results can only be regarded as approximate.) A wider access to old British newspapers has helped, and there are also detailed reports of the day in some Australian newspapers, although they were published weeks after the event – having been sent by mail rather than telegraph.
One peculiar thing about the day is that it appears that the over rate before lunch was very different to the over rate afterwards. Almost impossibly so, I thought, to the extent that I had glossed over it previously. Yet there are multiple lines of evidence suggesting it. The scoring rate was initially exceedingly slow, and the bowlers Giffen and Garrett stormed through their four-ball overs. There were 52 overs in the first 65 minutes, with only 20 runs scored, and 80-82 overs in 112 minutes before lunch, taken at 56 for 0. Giffen was taken off before lunch after bowling 36 overs for 22 runs. These stats are supported by independent sources.
This represents almost 180 balls per hour! This is the fastest over rate that I have heard of (equivalent to bowling 180 six-ball overs in a six-hour day), and bear in mind that the fielders changed end 80 times. Maiden overs must have been completed in a minute or even less.
After lunch, Grace went on the attack and the over rate ‘plummeted’ to about 133 balls per hour. The over rate seems to have been very sensitive to the scoring rate.
the innings was particularly noted for the slow, often immobile, scoring of Scotton. It has been previously recorded that he had
stayed on a score of 24 for 67 minutes; the reconstruction estimates that he
faced about 70 balls. Remarkably, however, it appears that this was not
actually Scotton’s biggest stall of the innings.
Before lunch, Scotton scored a three in the ninth
over, and did not score again until (about) the 60th over, facing
about 80 balls in the interim. Thanks to the extreme over rate before lunch,
this only took about 50 minutes. Ultimately, Scotton
faced about 290 balls for his 34 runs, still the slowest innings of its size
in all Test cricket.
I have posted
the revised reconstruction. There are still many uncertainties – among other
things, the sources sometimes conflict – but I hope that readers can accept
I have begun to include some data on dropped catches into my online material, for selected series – at this point restricted to scores of 50 or more. An example is here. This will only be available for a minority of series for 2000 and 2001, but I anticipate being much more complete from 2002 onward.
I have also accumulated dropped catch data for about 350 Tests in the 20th century, a sizeable number although still only about a quarter of the century’s Test matches. Some of this is thanks to Barry Valentine, who has made some deep dives into Ashes series between 1920 and 1961, using a wider array of sources than I did in my work. Barry has been kind enough to send me copies of his work; the analysis of a single series can run to more than 200 pages. Among the data is mention of dropped catches wherever they can be found.
I have analysed Tests from 2021 for missed chances and will report on that shortly.
In the meantime here is a list of batsmen who were missed most times in an innings, perhaps the first time such a list has been attempted. This list is subject to the usual caveats about dropped catches, and of course it is probably incomplete.
Most missed chances in a batsman’s innings – where recorded
Suggestions for additions to this list would be welcome.
UPDATE: Lawrie Colliver tells me that Clive Lloyd was dropped six times in his 242* at Mumbai in 1974-75. His source lists all six in detail. Actually I think I had read that somewhere but had forgotten.
After Pat Cummins declared in Sydney when Leach took wickets with consecutive balls, I looked for earlier Tests where a declaration was made with a bowler on a hat-trick...
Eng v WI (1), Bridgetown, Barbados 1935
SAf v WI (3), Bridgetown, Barbados 2001
Aus v WI (3), Bridgetown, Barbados 2003
Aus v SAf (2), Melbourne (MCG) 2005/06
Most bowlers faced by a batsman in Tests: Tendulkar faced about 300 bowlers, and Murali bowled to about 390 batsmen. In both cases, there are maybe 10 opponents included, but for whom data is uncertain.
Batting through a day most times…
If you exclude days with less than 50 overs (300 balls) and exclude the final days of Tests with a definite result, then the leaders are
AN Cook 15
G Boycott 11
KC Sangakkara 9
L Hutton 9
MA Atherton 9
This stat favours openers, partly because openers are the only batsmen who can bat through Day 1 of a Test.
After a bit more research into 12th men, I have found that Upul Chandana of Sri Lanka has passed Andy Bichel and now leads 20 to 19. There are also several other Tests played by Sri Lanka around that time where Chandana may have carried the drinks, but no 12th man names can be found.
Official Australian scores still usually have names for 12th man in them, but for most teams the idea has become obsolete.
In a number of Tests from 2013 to 2015, Australian teams had eight players born in NSW. There are also a number of earlier instances.
At Perth in 2002, Australia had 5 from (born in) NSW and one each from Qld, Vic, SA, WA, Tas and Northern Territory.
That is from a file I made in 2016, which I have not updated.
12 January 2022
Scott Boland’s Flying Start
Scott Boland must think this Test cricket game is a bit of a lark. So much for taking several years to adapt to the highest level of the game.
I have made several list showing some of Boland’s noteworthy achievements already, including a list of bowlers who were fastest to take their first 10 Test wickets, in terms of balls bowled.
Fewest balls for first 10 Test wickets
I don't have Hirwani's Test ball-by-ball, but I figure that after taking 8 in his first innings in 18.3 overs, he took wickets in each of his first two overs in the second innings, leading to the estimate of 120 balls. Perhaps readers have more info on that.
I'll admit that I hadn't heard of Sipamla. The 239 balls comprise his entire career to date.
I expect that the data is largely complete. If anyone can think of candidates for the Top 10 who are not there, let me know.
Fewest balls to reach six wickets in an innings, after first coming on to bowl.
Boland also equalled the record for fewest balls to reach five wickets in an innings, after first coming on to bowl - 19 balls. Shared with Toshack 1947-48 and Broad in 2015.
Six wickets in (the space of) fewest balls
Appropriate updates have been made to the Unusual Records section.
The First Boundary Hits
I mentioned last month the matter of the introduction of boundary fours in first-class cricket, an innovation that predates Test cricket. There is also the related concept of five runs or six runs for hits over the boundary. Here are some notes on this subject: some of this information has been provided by Shane Hicks.
The notes are by no means a full study, but are presented as is. It is not a particularly easy area to research, but it does seem that, while some very early examples have been found, it has been hard to find more. So it seems that the awarding of five or six for big hits in the 1860s was unusual.
· In the 1860s, boundary fours began to be recorded at larger grounds (with defined boundaries) in England. A curious exception was Trent Bridge, where most boundaries were awarded 3 runs and the batsmen changed ends.
· There is an interesting phrase in a report of an 1865 match at the MCG, saying that Ned Gregory's hits to the fence scored "four, as per agreement". The use of that phrase suggests that it was a novel idea.
· There is a reference in the Nottinghamshire Guardian in 1865 to Richard Daft hitting a six out of the Trent Bridge ground. Daft also hit boundary threes and boundary fours in the same innings.
A batsman named Coates was awarded six for a “brilliant hit outside
the fence beyond long off” in an intercolonial match at the
· At an intercolonial match between Victoria and New South Wales at the MCG in early 1870, a report makes reference to hits to the "pavilion fence" counting 3 and the "ring fence" counting 4.
· There is a reference to Charles Bannerman hitting a ball over the chains for five in an intercolonial match in 1874-75 (MCG). I haven’t found any earlier references to such fives in Australia. In this case, the captains agreed before play started that hits over the fence would count for five; this suggests that it was not standard practice. I noted two hits for five at the MCG in 1870 by Wardill, but it appears that both were all-run.
Alan Davidson is one of quite a number of bowlers who took a wicket with their last ball in Test cricket. But Davidson also took a wicket with his last ball in a non-Test first-class match, and it was Garfield Sobers out bowled in a Sheffield Shield match. I reckon that such a double (Test and non-Test f-c) must be very unusual.
That final Test was a few weeks after the Shield match.
When wicketkeeper Ridley Jacobs was injured in mid-over at Antigua in 1999, the bowler, Jimmy Adams, donned the pads and took over. This meant that he could not finish the over, so Carl Hooper finished it for him.
In the Old Trafford Test of 1999, every player for New Zealand went into the Test with a first-class century to his name.
The Australian tour of England in 1884 was so lucrative that each player made 900 pounds clear profit from the tour. This was at a time when £50 per year was a very good salary and many in the working class could earn less than £10 per year. The Governor of the Bank of England had an annual salary of £400 in 1880.
The tourists in 1882 – the year of the original Ashes Test – made between 600 and 700 pounds per man.
At Chittagong in 2008, Daniel Vettori had New Zealand’s best bowling in both innings, top scored with 55 in the first innings and was 3 runs off top score in the second innings (76 to Redmond's 79).
When did the concept of four runs for a boundary begin? My impression is that the idea of a boundary hit began in the 1860s. WG Grace said that in his earliest years all hits had to be run, and that around 1860 he once hit a six and a seven, all run, in the first over of a match.
By 1865 boundary hits were scoring four in Australia. There is an interesting phrase in a report of an 1865 match at the MCG, saying that Ned Gregory's hits to the fence scored "four, as per agreement". The use of that phrase suggests that it was a novel idea.
I would be interested if readers have any other information on this question.
Most bowlers go from "x99" to "x(+1)00" wickets in the same match. At about 326 days, Nathan Lyon took far longer than anyone to go from 399 to 400, previously Chris Broad with 76 days.
Most for other milestones...
99-100 W Rhodes 548 days.
199-200 CS Martin 288 days
299-300 RJ Hadlee 79 days
Historically, there are more than 40
cases in Tests of a batsman facing a hat-trick ball as the first ball of his
career, as Alex Carey did in the Gabba Test. At least three were out to their
first ball, thus completing the hat-trick (TA Ward, FT Badcock and WW Wade).
Ward also faced a hat-trick ball as his second ball in Tests and was out
again, completing TJ Matthews’ second hat-trick at Old Trafford in 1912.
From 2013 to 2015, several Australian teams had eight players born in NSW. There are a number of earlier instances.
At Perth in 2002, Australia had 5 from (born in) NSW and one each from Qld, Vic, SA, WA, Tas and Northern Territory.
Since 2017, when India started using DRS, Virat Kohli has been out LBW 17 times. One of those came via a successful bowling review. Of the other 16, Kohli reviewed 14 unsuccessfully. He chose not to review the other two; ironically one of them would have been given not out (he was on 204 at the time).
He has also made two successful batting reviews, where the initial LBW decision was overturned.
The 14:2 batting LBW review ratio is the worst among batsmen with 10 or more LBW reviews. David Warner is 8:1.
Mayank Agarwal’s 150 in the Mumbai Test against New Zealand is the highest score ever achieved in the face of a bowler taking all ten wickets (Ajaz Patel 10 for 119) in all first-class cricket. Previously 149 by Les Ames in 1934, when W Jupp took all ten in a county game.
Bangladesh lost a Test against Pakistan in Dhaka even though they did not commence the first innings until after lunch on the fourth day. This is the latest ever start for a first innings in a completed Test.
15 December 2021
The Curious Case of Frank Irving.
You don’t often come across someone like Frank Irving, a man who, despite playing a role in four Australian cricket tours to England, has virtually disappeared from the annals of Australian cricket.
They were historic tours too – the first four Test tours by Australian teams, in 1880, 1882, 1884 and 1886. So what role did Irving play? Until recently, no one has identified names for scorers for these touring teams: it appears that no official scorers were appointed. But now some detective work by various people on an ACS (Statisticians) chat page has produced convincing, if sometimes circumstantial, evidence that Frank Irving was involved in these tours as a scorer.
Irving was a newspaperman (specifically, a compositor) with an apparent penchant for long sea voyages; he travelled to England five times in his short life, and can be placed in England during all four tours. He acted as a reporter on those tours for the Advertiser in Adelaide, even though he was and is largely unknown for such work in Australia. Nor is his name found in connection with scoring in Australia. Yet it seems that, after faithfully following the teams around in England, he came to act as scorer in some historic matches, including the original Ashes Test in 1882.
I asked some of the luminaries of Australian and South Australian cricket history; they were unfamiliar with Irving. His name does not appear to be mentioned in the 1882 Ashes Tour book by Charles Pardon or Clarence Moody’s 1898 history of South Australian cricket, and eluded Michael Ronayne in his detailed tour summaries covering that period. The Trove newspaper archive does record his name, but only fleeting references have been found so far; likewise the British Newspaper Archive.
Currently it appears that Irving was not appointed to tour with the team in an official capacity, and he did not actually sail with any of the teams. He seems to have settled into the scorer duties rather unofficially.
Note that players George Bonnor and Jack Blackham are the only others known to have been involved in all four tours.
I would like to add some evidence from my score collection that links two of the tours. I have copies of scores from the 1882 and 1884 Tests from the respective Australian tour scorebooks, and I am convinced that they were written by the same person. Examples of the writing are shown below. (The 1880 and 1886 Australian tour books cannot be located; the surviving 1880 score is from the Surrey CCC and is in a different hand, presumably that of the local scorer.)
The styles of numerals are also identical in the two scores. I would add that the method of recording bowling figures, in both cases, is unusual. In the box for each over, the four balls are arranged in a square pattern, with the first ball of the over in the bottom line, like so…
whereas scores made by others from the period are what I would call more conventional…
This helps tie together the scorers for 1882 and 1884. The evidence is also clear that he was scorer in 1886 (see below). The evidence from 1880 is less clear, but he definitely accompanied the team.
While the Manchester and Oval Test scores from 1884 are complete, it may well be that Irving did not actually score the second Test at Lord’s. The Lord’s score is in the same hand as the other Tests but it is incomplete, lacking detail in the bowling section. It is probably a re-copy made by Irving.
There is another Test in the back of the 1884 scorebook – the first Test of 1884-85 at Adelaide. It was played in December by the 1884 tourists, who were still together, against the English team that had just arrived in Australia. This score is not in the same hand as the Tests in England. This suggests rather strongly that Irving did not accompany the Australian team on their return voyage. Indeed, passenger manifests show that Irving did not return to Adelaide until late December, after the Adelaide Test match.
The following compiles information available about Irving, as gathered by various ACS members.
The critical observation that got this investigation going was by Harry Watton when he picked up the reference to Irving in the December 1902 Cricket magazine.
Frank (Francis) Irving
Born 1855 Adelaide
Joined Advertiser in 1876.
March 1879: Identified as a printer who was going to England via New Zealand and USA as a journalist for the Advertiser. His activities in 1879 are unclear; there is no direct evidence that he was a scorer in 1880, but he certainly accompanied the team.
Returned from Plymouth to Adelaide departing 2 October 1880.
Listed as a passenger on the steamer POTOSI sailing from Adelaide bound for London on 13 March 1882, departing three days before the Australian team sailed from Melbourne. (South Australian Register 14 Mar 1882).
Identified as Australian scorer against Cambridge P&P at Portsmouth (Hampshire Telegraph 19 Aug 1882), a week before the ‘Ashes’ Test.
Frank Irving age 26, listed arriving in Adelaide in Dec 1882 from England on the HAUROTO. This was not with the team, which had returned via America in November.
In a 1902 interview in Cricket magazine, Surrey scorer Fred Boyington said that Irving was his co-scorer in the 1884 Oval Test.
A Mr F. Irving, age ~30, is listed as a passenger on the AUSTRAL departing London on 11 Nov 1884, bound for Adelaide (and on to Melbourne and Sydney). He is the only F. Irving listed travelling to Adelaide in the years 1883 thru 1885. Newspapers have him disembarking in Adelaide on Dec 24.
Irving was a press representative on the tour (as he had been in 1884), the Adelaide Express reporting on 19 April 1886 that Frank Irving is ‘about to visit England [to] represent the Advertiser in connection with the Australian Eleven’. He is reported in the South Australian Chronicle of the same day to have made three previous visits in that capacity (i.e., 1880, 1882 and 1884).
A column in the London Evening News of 17 July 1886 refers to ‘[m]y old friend Mr. Frank Irving, the Australian scorer’, and reports an anecdote of Irving’s from the recent match between Yorkshire and the Australians on 12-14 July, when the ball had whistled past the scorebox.
F. Irving age ~31, occupation compositor, arrived back in Adelaide from England aboard IBERIA on 7 Jan 1887.
1887: captain of Advertiser cricket team.
No known connection with the touring team, or evidence of travel.
Leaves for England again 11 Nov 1889?
Died (intestate) 6 May 1890 Carlisle, Cumberland, England, just prior to a planned return to Australia. He is listed in a ship manifest departing London on 9 May but his name is crossed out.
Irving’s ancestors and relatives hailed from Dumfries Scotland and the surrounding area, just over the border a few miles from Carlisle.
21 June 1890: a short obituary appeared in South Australian Chronicle (“four trips to England since first journey with the Australian eleven”)
ACS members who gathered the above information include Harry Watton, Neville Flood and Sreeram Iyer.
The Test Match Database has now reached the 21st Century. It is now well and truly overlapping with other online ball-by-ball records, but I will continue with it for now. For some Tests, there is material in my scores (for example, session scores) that would be rather hard to extract from other online sources. I don’t know how much further I will take it; who knows what 2022 will bring?
I came across a 1982 newspaper article by Bill O’Reilly. He was talking about his early recollections of cricket broadcasting on television, which he had first seen in 1938 in England. The match was Middlesex v Australians on 30th May 1938 (O’Reilly says it was the 28th, but that day was rained out). This is before the Tests that year, and must represent one of the very first cricket TV broadcasts.
The television set (super-expensive in those days) was on loan to the hotel where the Australians were staying. Most curiously, O’Reilly said he was in the team playing that day, but he had stayed behind in the hotel to deal with some correspondence! He had expected the team to bat all day, but they started losing wickets. O’Reilly could see from the broadcast that the pitch was dodgy, so he hurried down to the ground, and was just in time to bat and make a duck.
I took a look at the question of a bowler taking a wicket with his last ball of a Test and first ball of his next, and found only 25 cases (from the database covering about 85% of Tests). Waqar Younis and Richard Hadlee did it twice.
The low number might seem surprising when you consider that there are well over one thousand cases of wickets with consecutive balls in general play, but remember that there is never more than one opportunity per Test match for a bowler to do take two in two in different Tests (and usually no chance at all), whereas the same bowler may well have multiple opportunities for two wickets in two balls during the course of a Test match.
There are no known cases of three in three across two Tests (a quasi-hat-trick) except for the extraordinary case of George Lohmann, who took a hat-trick to finish a match in 1895-96, and then a wicket with his first ball of the next Test, to secure four wickets in four balls (he made it five in six two balls later).
Overall, the chances of taking a wicket with a hat-trick ball are only 1 in 30, so, given the low incidence of two in two balls across two matches, it is not surprising that three in three is so rare.
There is an interesting case of Mervyn Dillon, whose last two balls at Port of Spain in 2002 were a run out and a wicket (in two different overs), followed by a wicket with his next ball at Bridgetown.
Waqar Younis extended one of his wicket pairs to three wickets in four balls; Rashid Khan of Afghanistan has done the same, and that was across his first two Test matches (9 months apart).
1 November 2021
The Test Match Database Online has reached a crossover point with the Cricinfo Ball-by-Ball Archive texts. Beginning with the 1998-99 India v Pakistan series in January 1999, Cricinfo began to archive their bbb logs. I am fairly certain, from memory, that they were doing some ball-by-ball descriptions of internationals before that – the earliest record goes back to the 1996 World Cup final – but for some reason they never preserved them (sad emoji). If I have checked correctly, the only Test that we have from Cricinfo, from that period, is a single day from the 1997 India series in West Indies.
The early Cricinfo texts were typed commentaries and were not designed as rigorous scores; there are gaps and anomalies. Where necessary (and possible) these have been adjusted with reference to surviving scores or other published information. In a few cases, such as the final day of the Asian Test Championship Final in Dhaka 1999, the gaps and problems are substantial, and there is no detailed backup. In such cases, I have ‘recreated’ a bbb version that is consistent with surviving scores and reports. These cases are few and I hope I will be forgiven for doing this for the sake of completeness.
There are actually six Tests from the early period that are missing entirely from Cricinfo, five of them from New Zealand; I have managed to obtain alternative scores for all of these and so maintain continuity of the ball-by-ball record.
I have also obtained alternative scores for a significant number of other 21st Century Tests, but this collection is not comprehensive. Of the first 100 Tests from the ‘Cricinfo era’ starting in 1999, there are 36 Tests where I rely entirely on Cricinfo for bbb records.
The archived Cricinfo texts gradually improved in detail and reliability, especially after 2002.
It would be wonderful if someone at Cricinfo could unearth some ancient backup tapes from 1997 and 1998 and find some more bbb logs. 1998 is not so essential as I have all the Test matches from that year already (some ODIs are missing), but there are considerable gaps in my data for 1997. I understand that there has been a search of this kind, sadly without success.
I am not sure how far into this new era that my online work will go. But I have been at this for nine years now and I don’t really know how to stop.
The Kolkata Test of February 1999 at Eden Gardens is one of only seven Tests that Pakistan has played at that cavernous venue. The match quite possibly attracted the highest attendance of any Test match; however, the numbers were never accurately counted. I am told that the ground had 90,462 seats at the time, so any numbers in excess of that must have been standing room only. The available numbers are estimates, and it must be said that the estimates vary widely.
I have collected (with assistance from others) some mentions of crowd numbers from reports at the time, illustrating the variations in estimates…
Ironically, the last overs of the match were bowled in a virtually empty stadium, following a roughhouse clearing out of the final day crowd by police, in response to serious unrest that suspended play for three hours and 20 minutes. There had also been serious unrest on Day 4.
Even though it appeared in Wisden, I consider the estimate of 465,000 to be highly improbable. A figure around 400,000 seems a reasonable compromise from the conflicting numbers. This number is in the same ballpark (if I may use the expression) as the estimate of 395,000 for the Test against Australia at the same Ground in 2001, and the 390,000 estimate from 1981-82. I don’t have figures for Tests at Eden Gardens after 2002, but it has been apparent that Test attendances have been declining in India, even though interest remains substantial. At the same time, Indian authorities have been distributing Test matches among a wider range of smaller venues, and Kolkata has hosted only 12 Tests in 22 years since the match in question.
The largest accurately measured total attendance for a Test remains the 350,534 over six days at the MCG back in 1937. The most for a five-day Test is 271,854 for the MCG Boxing Day Test in 2013-14 (including a ground-high 91,062 on the first day), and the highest average daily attendance is 81,450 for the equivalent match in 2006-07, which only lasted three days.
I don’t think I have discussed before, on this blog, the origins of the early Test match “canon”, that is, the list of matches regarded as official Tests.
It is widely held that the originator of the list was a South Australian journalist named Clarence P. Moody. Beyond that, it remains a bit of a mystery how Moody’s list, drawn up in the 1890s, became accepted as gospel. With limited debate, and no input from English sources, it has become set in stone, so to speak. Moody was something of a cricket statistician. He was also a friend of George Giffen, who may well have suggested the creation of the list. However, there was never any official imprimatur on Moody’s work.
I went looking for Moody’s original list. Various printed sources and online articles said that it came from Moody’s 1898 book on South Australian Cricket. I managed to borrow, from Roger Page, a facsimile copy of this book (originals are rare and valuable) and was rather surprised to find that it contained no such list. However, an Introduction to that facsimile edition pointed me to an earlier (1894) Moody work, Australian Cricketers 1856-1893-94. Fortunately, the Trove Archive has this book online, and there can be found the original list, on pages 80 and 81.
Clarence Moody’s 1894 list of Test matches
One noteworthy aspect is that Moody restricted himself to England v Australia Test matches. In that respect, the list is indeed identical to the accepted canon, but there are several other matches, played in South Africa in 1888-89 and 1891-92, that are not included. So how did those Tests get into official lists? I don’t know but I would like to find out. (I am told that the South Africa matches were listed as Tests by Ashley-Cooper in a Cricketer Annual in 1930-31, but I don’t know if there are earlier references.)
I do know that there has been plenty of doubt and dispute over the Test status of some matches, and also some matches that did not make the list but might have. In the case of 1888-89, even the first-class status can be questioned, since there was no first-class cricket in South Africa at the time, and a number of the Englishmen were themselves not first-class cricketers. (A favourite stat: JEP McMaster played Test cricket, but was out to the only ball he ever faced in first-class cricket.) In his 1951 collection of Test match scores (The Playfair Book of Test Cricket), Roy Webber certainly expressed a sceptical view, but he also said that “little purpose seems to be served by omitting ‘doubtful’ matches”. I agree, but I think that caveats need always to be expressed when records from those matches crop up (such as Briggs 15 wickets for 28 at Cape Town, 14 of them clean bowled).
A comment from the Sydney Sportsman in 1901 seems pertinent here…
It should be said that Moody’s list did not require deep scholarship. Once the matches of 1877 are accepted into Test cricket, most of the rest falls into place. However, the fact that the list came to be used as a reference, in the face of the disagreements expressed above, makes it important.
I won’t go into the detail of the claims of certain matches for Test status. But here are a couple of observations:
- The touring English team in 1884-85 did not regard the first two matches, now regarded as Tests, as authentic. As far as they were concerned there were three Tests in the series. The canon lists five.
- The 1887-88 match was by no means regarded as a proper Test match even in Sydney. The Sydney Morning Herald described the Australians as the "non-representative Australian XI".
An Earlier Wagon Wheel
I have written (somewhere) that the first known batting Wagon Wheels were found in Test match reports in the Daily Express in 1905. Sreeram has now pushed the date back a little further, finding similar diagrams in Manchester Guardian reports for the first Test of that series (innings by Hill and Tyldesley). Unfortunately, the Guardian reports do not say how the diagrams were made or by whom, although they do claim copyright. The Express does suggest that theirs were drawn up in the newspaper office, based on information telegraphed or telephoned from the ground.
It has been claimed/reported that Wagon Wheels were invented by Bill Ferguson, who in 1905 was the Australian scorer, making his first of many tours. However, the diagrams in the British papers are different in style to Ferguson’s, with concentric circles for each run value. No actual Ferguson-style Wagon Wheels earlier than 1911 have been sighted.
In his autobiography, Ferguson provides a whole range of examples of his Wagon Wheels, but he is curiously vague about their development. The earliest diagram in that book is from 1912 (he didn’t call them “Wagon Wheels” by the way; I wonder when the term originated).
Sreeram has added further evidence by finding another 1905 Guardian Wagon Wheel, this time from a county match, Lancashire v Yorkshire in June (Tyldesley 134). It even names the bowlers for each shot.
Ferguson at this
time was elsewhere scoring for the Australian team, so he could not have
contributed to this. I think that it is now fair to say that Ferguson
adopted, rather than invented, the idea of a Wagon Wheel.
In the Test match at The Oval, Rohit Sharma (127) was dismissed by the first ball with the new ball, after a partnership of 153 with Pujara. This is the second highest partnership ended by a brand new ball, after a partnership of 172 between Mark Richardson and Stephen Fleming at Colombo PSS in 2003.(where known)
Something curious about cricket watching in Sri Lanka…during England’s Test match in Colombo in 1992-93, the attendance on the Saturday was only about 1,000. On the same weekend, two interschool matches in Colombo attracted crowds of over 10,000.
(source: Sunday Times)
Here is a
neat little stat. When West Australian Des Hoare
batted for the first time on debut at Adelaide in 1961, his first ball was
from Lance Gibbs, who had just completed the hat-trick.
As far as I can see, Hoare is the only batsman in Test history whose first ball in Tests was the first ball after a hat-trick (the double hat-trick ball). Others on debut have come in after a hat-trick, but it was either the second innings, or they didn't face the double hat-trick ball. One case was CA Absolom in 1878-79, who came in after Spofforth took a hat-trick, but the hat-trick came off the last three balls of an over, and Absolom faced a ball in the next over, before Spofforth bowled again.
Longest sequence of missed chances off a bowler without a catch being taken (data since 2003)...
In 2019, Joe Root had a sequence of seven chances missed off his bowling, without any catches being taken, spread over three Tests. Same thing happened to Nathan Hauritz in 2010, again over multiple Tests.
It hadn't occurred to me until a question on Ask Steven, but I found that never before had a team, behind on first innings, declared its second innings after lunch on the 5th day, and won. This is what India did to England at Lord's.
The most threes in a Test career were hit by Steve Waugh (397) and Ricky Ponting (381). The top six positions are held by Australians who were active (at least in part) during the 1990s, before the much-lamented ‘shrinking’ of the grounds and the advent of Super Bats. Mark Taylor hit 354 threes or 14.1% of his runs; this is the highest percentage of any major batsman.
Threes have always been more common on the larger Australian grounds. The top 13 positions are held by Australian or English batsmen.
16 September 2021
I have been working on a list of “Unusual Dismissals” in Test matches. This is largely drawn from notes that I have made over the years in my study of Test matches and scorebooks, rather than specific research. Some of the instances have been thanks to suggestions by others.
The stimulus to finally put this together was the dismissal of Nauman Ali at Harare earlier this year, stumped by Chakabva off a wide. As far as I know, this was unprecedented in Test matches, although it happens from time to time in limited overs games.
Any suggestions for additions would be welcome. The general criteria include: a dismissal must have occurred, and there must have been something very unusual about the dismissal itself, not just the situation (for instance, run out for 99 is not included unless there was something strange about the dismissal). Also excluded is where one fielder drops a catch and deflects to another: there are actually many examples of this in Tests. Bowlers deflecting a shot to effect a run out is also excluded, unless it was a dropped catch.
While checking through the list of most overs in a day by individual bowlers, I noticed again the absence of any modern names. The list is effectively set in stone, and is dominated by Tests where over rates were high and days lasted six hours. Most instances came from the 1940s and early 1950s, when some captains, rather lacking in imagination, would put spin bowlers on and just keep them bowling. This certainly applied to John Goddard of the West Indies.
So I have prepared a separate list for Tests since 1998.
Most Overs in a Day: individual bowlers since 1998.
Incomplete overs counted as one
There are ten cases of 39 overs. Even in this list, the near absence of instances in the last ten years is notable, with just one appearance, by Jadeja. There are only three instances from the last ten years in the Top 30. By contrast, Muralitharan appears ten times in the Top 30, and even then his Tests before 1998 are not included.
There are worrying signs that over rates are on the way down again. It has become unusual for teams get through 90 overs in the allotted six hours, often not even in the allowed extension to 6.5 hours. The great majority of innings in Tests this year have recorded over rates of 80 balls per hour or less (40 out of the last 46 innings). In the 1960s, rates of more than 100 balls per hour were commonplace; in the 1940s and earlier, it was 120 balls per hour or more.
Fielding sides are mostly to blame, but not entirely. Batsmen nowadays often go through elaborate preparations and are quite frequently not ready to face when a bowler is trying to get through an over more quickly than usual.
There has been a general checking and updating of various records in the “Unusual Records” section. Not everything has been checked, but those sections that have been checked have been appropriately labelled with a date.
A scoring curiosity: in the second Test between Zimbabwe and New Zealand at Harare in 1997-98, it appears that when no balls were scored from, an extra run was added (for example, a no ball hit for four added five runs to the total). This was not the established protocol at the time: the practice came into general use about a year later. There were no instances in the first Test of that series (none of the no balls were scored from), but there were several instances in the second Test.
The change became permanent in Test #1424, Pakistan v Aus, Oct 1998.
In the Zim Test, it is particularly interesting in that the match was very close. In the final innings, the 8th wicket fell with 4 balls left and 11 to win. At that point the tailenders decided not to go for victory and played out the four balls. But if the scoring protocol had been normal for that time, the target would actually have been 8 runs (and 7 to tie), not 11. I wonder: perhaps they would have had a go at that.
From time to time a batsman scores a century on the first day of a Test even though his team batted second. Bowlers have also been known to take five or more wickets on the first day, bowling second, but it has become rather rare in these days of covered wickets and slow over rates. The most recent case is Glenn McGrath at Lord’s in 2005. After Australia was out for 190, McGrath took five wickets for two runs from his 4th to 9th overs, immediately following tea. He finished the day with 5 for 21.
Fred Price was a wicketkeeper from Middlesex who played one Test in 1938. He played 402 first-class matches without ever bowling, the all-time record for a complete career.
Kumar Sangakkara played 529 List A games (and also 267 T20) without bowling. However, he bowled in f-c cricket.
T20: Eoin Morgan has played 333 T20s to date without bowling.
Combined totals: Steven Davies, who has played for England and various counties, has played 581 games (240 f-c, 188 List A, and 153 T20) without bowling. He bowled one over in an Under-19 Test. He is still active.
17 August 2021
In response to an enquiry, I put together a list of bowlers who have bowled underarm in Tests.
Gerald Brodribb actually wrote a book on underarm bowling. I don’t have a copy, but there is a surprising amount of stuff on ‘lob’ bowling on the internet (Cricket Country and elsewhere) that is probably derived from Brodribb, at least in part. I have scoured this and come up with the following (some is from notes in my database). Most of these were bowling lobs, and the underarm part is presumed. Hornby is an exception, bowling "grubbers".
Underarm bowlers in Tests
T Armitage Eng v Aus (1), Melbourne (MCG) 1876/77
AN Hornby Eng v Aus (1), Melbourne (MCG) 1878/79
WW Read Eng v Aus (1), Melbourne (MCG) 1882/83
WW Read Eng v Aus (3), The Oval 1884
A Lyttleton Aus v Eng (3), The Oval 1884
G Ulyett Eng v Aus (2), Melbourne (MCG) 1884/85 one ball only
AE Stoddart Eng v Aus (4), Melbourne (MCG) 1897/98
GHT Simpson-Hayward Eng v SAf 1909/10 Five Tests
Simpson-Hayward is regarded as the last of the regular lob bowlers.
Test players recorded as bowling underarm in first-class cricket but not in Tests
For the most part, these players did so only once, or on rare occasions.
Yuvraj of Patiala
Dilip Vengsarkar is said to have bowled an over of lobs for West Zone against the MCC in 1984-85. However, these appear to have been ‘donkey drops’, not underarm.
I read also the Hornby was also ambidextrous with regard to bowling. Apart from him, I have not seen any references to bowling ‘grubbers’ (as opposed to lobs) in Tests.
Dismissal Frequency by Ball of Over
Is there any pattern to dismissals according to which ball of an over is being bowled? I decided to have another look at this. For all dismissals in Tests from 2003 to 2020, the numbers are
This distribution is largely random. There is a slight shortage of dismissals on Ball #1, which seems to be associated with a pattern in tail-end dismissals. There is an excess of 8th-10th wicket dismissals toward the end of an over, perhaps because of failed strike-farming attempts.
For the first seven wickets, where the innings continued after the dismissal, the numbers are
This is a fairly random set of numbers. None of these numbers is even 1 percentage point away from the mean of 16.67%.
13 July 2021
I have updated the Hot 100 list – the fastest-scoring Test batsmen. It’s been a couple of years since the last update.
Most batsmen tend to score at a characteristic rate which varies less, over time, than batting average. The upshot is that the list changes only gradually, apart from new players making an appearance, so it doesn’t matter too much if updates are infrequent.
It is interesting though that David Warner, still very prominent in the list, has been ‘calming down’ to some extent in recent years. He sometimes plays defensively now (only sometimes). His career scoring rate as of 2019 was 74.5; this has now dropped to 72.7. His scoring rate in the interim has been 62.5 runs per 100 balls, and that includes his triple-century against Pakistan.
The full lists are at the usual link. I have also prepared the list below, restricted to fully-recognised batsmen, just to see how the list looks without the lower-middle-order all-rounders and wicketkeepers who are prominent on the full list. The list is filtered simply by restricting to batsmen with an average batting position of less than 6.1. (All innings for such batsmen were included, even below #6 position, as long as the career average position was 6.1). The runs qualification has been raised to 2000 career runs for modern batsmen, although it remains at 1000 runs for earlier times.
Given these qualifications, Virender Sehwag’s lead becomes very striking indeed. Shahid Afridi was faster still as a top-order player, but his Test record is rather patchy and he never reached 2000 runs.
Fastest-scoring Test batsmen: ‘recognised’ batsmen