Z-score’s Cricket Stats Blog Archive 2018-19
The longest-running cricket stats blog on the Web
For comments, or to contact Charles Davis,
at iinet.net.au (no spaces)
Charles Davis: Statistician of the Year (Association of Cricket Statisticians and Historians)
Click on the Date to go to that Blog Entry…
Detailed scores for all Tests from 1877 to the1970s have now been posted. More than two-thirds of Tests include ball-by-ball coverage; virtually all others offer some degree of extended detail, beyond anything previously made available online.
A Bonus Page: some remarkable first-class innings, re-scored.
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.
A further note on the use of air transport: the 1938 Australian touring team to England actually had a clause in their contracts that air transport would not be used. The first use of air transport for a cricket tour came in 1946, when the Australian team flew to New Zealand. Australian tours to England continued to travel by ship up to 1964 (which went part of the way by air); the 1968 team travelled entirely by air to Britain.
At Port-of-Spain in 1960, Frank Worrell was caught on the boundary (one-handed) by Fred Trueman, but umpire Jordan signalled six. It is not clear what the umpire thought had happened, but Worrell ‘walked’ and averted any controversy.
It is unusual to come across ‘significant’ errors in the standard online Test data sources (Cricinfo /Cricket Archive), but here are a couple from 1993…
At Moratuwa (Sri Lanka) Jonty Rhodes in the second innings scored 101* off 204 balls, not 107 balls. The Wisden Book of Test Cricket says 193 balls, but that applies to Rhodes reaching the century, not his final BF.
At the Gabba later that year, Richard de Groen’s innings of 3 is given as 3 balls and 5 minutes, even though his last-wicket partnership with Tony Blain was worth 40 runs. De Groen actually faced 30 balls and batted 55 minutes.
In the third Test of 1993-94 against Sri Lanka at Ahmedabad, Sachin Tendulkar, at age 20, was the Indian vice-captain. Tendulkar took on the on-field captaincy duties on the fourth day when Mohammad Azharuddin was indisposed. Has any player younger than this (20.79 years to be exact) ever taken on captaincy duties, even in an acting capacity?
The youngest confirmed named captain was Tatenda Taibu, who was 21 when he captained Zimbabwe in a couple of Tests. (Rashid Khan was supposedly younger, but I do not regard dates of birth from Pakistan and Afghanistan to be reliable.)
17 December 2020
About four years ago I posted a list of the best head-to-head bowling averages for bowlers, specifically when bowling to the best batsmen – those with batting average over 45. A reader, Arjun, suggested that I try a similar exercise for batsmen against the best bowlers. To do this I had to decide on some criteria for deciding who the best bowlers were. I came up with the following:
- Bowlers with 200 Test wickets, or 150 before 1970, or 100 before 1940.
- All other bowlers with 25 or more wickets and a bowling average under 30.
Bowlers meeting these criteria have taken about 35,000 wickets, representing about half of all wickets taken by bowlers. This divides the bowlers neatly into two similar-sized sets.
Not surprisingly, most batsmen’s averages against the ‘top bowlers’ are lower than their career averages. It is not always the case, however. Typically, the top bowler average is about 87% of career average. I would say this is a smaller effect than I would have expected.
Anyway, here are the batsmen with the best averages against top bowlers…
Minimum 1000 runs against top bowlers
So Bradman’s 99.94 comes down to 75.5 when facing the best bowlers of his time. This is thanks to the efforts of bowlers like Hedley Verity and Alec Bedser, who did well against The Don. Next come a few West Indians, whose averages were little affected by facing top bowlers: I can’t really explain this except to note that Tests in the West Indies in the 1950s were a graveyard for many top bowlers.
Prominent batsmen who did not meet the 1000 run threshold but did very well against top bowlers include Graeme Pollock (68.4) and George Headley (55.7). A curious case is Andrew Jones of New Zealand, whose top bowler average is 56.4, which is 127% of his career average of 44.27, the highest (and most counterintuitive) percentage for any batsman. Wasim Raja had a similar percentage (career average 36.16, top bowler average 46.0).
I’m not sure what more to make of this. Why some batsmen do better against good bowling than bad bowling is puzzling. Perhaps others can make more of this than I can.
The One-Year Wonder
Some notes I made years ago on fast bowler Ted McDonald. I thought I might as well post them…
Warwick Armstrong certainly seemed to prefer his fast men bowling from the same end taking turns, rather than bowling in partnership. Sometimes McDonald had to shoulder a bigger burden: if Gregory had scored runs, Armstrong would cut back his bowling duties for a few hours thereafter.
While they usually opened together, it appears that Gregory and McDonald, as a pair, never took the second new ball together (available after 200 runs in those days). Generally, Gregory would take the new ball, with someone like Kelleway or Hendry at the other end.
McDonald's entire Test career fitted into one calendar year (1921), in which he took 43 wickets. No other bowler with such a short (single-year) career ever took so many wickets; in fact no one is even close.
When Armstrong bowled his famous two overs in a row after an abortive declaration by England at Old Trafford, McDonald probably should have bowled the intervening over. It was reported that the umpires ignored calls from the crowd that the wrong bowler was preparing to bowl. The score presents a puzzle here: I am pretty sure that McDonald bowled only 30 overs, not 31 given in official scores, in that innings.
McDonald twice dismissed batsmen by breaking their bats. At Leeds, McDonald broke A Ducat's bat. The splinter hit the stumps while the ball was caught by Gregory, and the batsman was given out caught. At Johannesburg, McDonald broke the bat of JW Zulch, the splinter hitting the stumps and he was given out hit wicket. I know of no other case of a batsman being out in this way.
Batting at The Oval, McDonald thought he was out bowled (by Parkin) and left the crease, but was recalled by the England captain Lionel Tennyson, who felt that the wicketkeeper had dislodged the bail.
There is a picture of Tennyson batting one-handed against either Gregory or McDonald at Leeds (the other bowlers he faced would have had the keeper at the stumps). Tennyson hit a five off McDonald in this innings, and (amazingly) two fours off Gregory and a six over square leg off Mailey.
The first ever international cricket broadcast?
I came across this snippet from the 1931 England/New Zealand series. It is not 100% clear to me what form this broadcast took…
UPDATE: Peter Huxford in the UK has sent me some links to New Zealand reports of these broadcasts, found on the Papers Past website. It appears that the broadcasts were after-play match reports, and they began with the first Test of the 1931 series. The broadcasts, some of which were relayed through Sydney, were sometimes unsuccessful. It seems that they were attempting shortwave broadcasting; very ambitious in those days considering the distances involved.
It is also evident (from newspaper radio guides) that receivers in New Zealand were able to pick up conventional Medium Wave broadcasts from Sydney and Melbourne at that time. They must have had good aerials, but it was quite feasible, especially if there was ‘clean air’, where frequencies were uncluttered by competing stations. Radio reports of the 1930-31 Australia/West Indies series were received in New Zealand.
It is curious that shortwave broadcasting to Australia was not attempted for the 1934 Ashes series; the ‘synthetic’ broadcasts using telegraphed information were preferred. In 1938 the synthetic broadcasts were still being used; shortwave technology was improving but reportedly still unreliable. (The ABC, incidentally, did not invent synthetic broadcasts – they had been used in 1930 by Australian commercial radio stations.)
asked: John Reid hit four sixes before lunch on Day 1 at Calcutta, 1964-65.
Has anyone else done that ?
5 Gayle at Basseterre v India 2006
4 Blackwood at Kingston v India 2016
4 Hayden at Chennai v India 2001
Blackwood was batting at #5 !
The Reid one is not in the database and was a surprise to me; so there could be others, although it seems very unlikely in older Tests.
The most in any single session is 9 by Hammond and Astle. You can guess in which Tests.
UPDATE: Gayle’s 5 sixes were actually hit in the second session of the match, the first having been rained off.
Update on the 1961 video. (see 18 October)
John Leather writes to say that the BBC did have video machines from 1958 onwards. However, the Ampex machines cost $70,000 each and the tapes were so expensive and few in number that they were used only for short-term storage. Editing these tapes was not practical.
Long-term recording of footage was primarily done with kinetoscopes (also known as telerecording), which converted the image as displayed on a monitor onto conventional film. The 1961 Ashes footage was archived and stored in this way.
continued through the 1960s even as videotape technology
improved. Tapes were wiped and reused even for important shows like Doctor Who
and At Last the 1948 Show, a 1968
precursor to Monty Python. It is fortunate, perhaps, that permanent copies of
Monty Python were kept.
21 November 2020
Timeline of India’s Famous Win, 3rd Test Port of Spain 1976
This was an important match historically. Not only a great win for India, but a turning point for the West Indies, who, after this humiliation, always played for keeps, with the fast bowlers offering no quarter. In the following Test at Kingston, the short-pitched bowling was given free rein, and India eventually crumbled. It was the start of a dominant era for the West Indies.
A Question from Arjun
Nathan Hauritz dismissed Andrew Strauss in all the 6 innings he bowled to him in ODIs. Is it true?
A: This is very well spotted. Not only is it true, but it is possibly unique. Strauss and Hauritz played in 11 ODIs together, but Hauritz bowled to Strauss in only six of those and dismissed him each time.
I have look at ODIs since 2001 (about 58% of all ODIs) and this is the only case of 6 that I found. Other cases of 5: Shane Watson to Chris Gayle, Murali to Chigimbura, and UT Yadav to D Ramdin.
It may also have happened earlier than 2001, but I think that the data is too patchy to make a judgement.
In Tests, Hugh Tayfield bowled to JH Wardle in eight innings and dismissed him each time. This is the most that I have in the Test database. Next is Patrick Patterson to Craig McDermott with 7 dismissals.
Head to Head without dismissal
We are looking here at the number of games where the bowler actually bowled to the batsman. Given that Wasim Akram took far more ODI wickets than anyone else, it is remarkable that he never dismissed David Boon in the 18 innings that he bowled to him…
Wasim Akram bowling to David Boon… 18 games
J Srinath to W Cronje… very uncertain but no more than 18.
JH Kallis to Michael Bevan… 17 games
Chaminda Vaas to Bevan… 15 or 16 games.
S Pollock to Shoaib Malik… 14 games
Mohammad Nabi to Sikander Raza… 14 games
Murali to AD Jadeja… up to 14 games
The uncertainty is due to lack of data for some games. Without ball-by-ball data, it is often not possible to be certain whether a batsman faced a particular bowler.
In Tests, John Gleeson bowled to John Edrich in 19 innings without ever dismissing him. Also on 19: Carl Hooper bowling to Steve Waugh.
With data being incomplete, there could be other pairings exceeding this. However, it is very unlikely. For example, Mudassar Nazar bowled in 25 innings that Sunil Gavaskar batted in, without ever taking his wicket. Ball-by-ball records are missing for most of these Tests. However, available data plus a close look at scorecards suggests that in at least 9 of these innings, Mudassar could not have actually bowled to Gavaskar.
Counting No Balls.
The counting of no balls in bowling analyses (runs conceded) started in 1983-84, but for some reason New Zealand and England did not adhere to it until later. The 1983-84 Test series between New Zealand and England, 1983-84 Sri Lanka and New Zealand, and 1984 Tests in England did not adhere to it. My bbb database for the 1984 Tests was in error in this respect and I have fixed that now.
The adding of a run for every no ball (even when scored from) was much later, starting with Australia v Pakistan in 1998-99. I am pretty certain that all countries changed from that point on. Note that this was a more important change than the earlier one, because it actually changed team totals; it is probably the only such change since sixes were introduced. I have noted earlier that Australia would have won the 1992-93 Adelaide Test under the later protocol, because they scored off more no balls than West Indies. England would also have beaten Zimbabwe in a drawn Test in 1996 for the same reason.
As for multiple no balls being separated into 1 no ball plus byes/leg byes, I have notes of cases occurring from mid-2018. A couple of multi-no balls in mid-2017 were simply described as ‘5 no balls’ and ‘2 no balls’ respectively in the Cricinfo texts, and it appears that they ended up as multiple no balls in the scores. So 2018 seems to be the starting point.
The accounting of no balls and wides against bowlers has a rather complex history. Wherever possible I think that, in scores, the no ball/wides columns against bowlers should count only the actual number of deliveries called by umpires. This is currently the protocol; however, varying protocols have been used in the past. I may at some stage change my online scores to reflect the modern standard (wherever possible but a big job). However, there will be a significant number of other older Tests (about 300) where this cannot be done (bowlers’ no balls and wides were published, but bbb is not available). There are also about 240 further Tests where no information on bowlers’ no balls and wides is available at all.
The Earliest Broadcasts
The first full ball-by-ball broadcast of a cricket match was for the 3rd Test of the 1924-25 Ashes series in Adelaide, from 16-23 Jan 1925. Bill Smallacombe of radio station 5CL was the sole commentator, and he did the whole 7-day match without help.
had only commenced broadcasts a few weeks earlier. There cannot have been
many home radio receivers in Adelaide at the time, but large crowds gathered
around shops and other places that had radios.
had been earlier live radio reporting of cricket matches, including the first
two Tests of that series (Sydney and Melbourne) but they were in the form of
regular updates rather than ball-by-ball descriptions.
Credit to Bernard Whimpress who researched the 1925 Adelaide broadcast.
Last Use of Boundary Fences in Australia
In Australia, some of the Tests of 2000-01 retained the fence boundaries, but (from video evidence) all the Tests of 2001-02 had boundary ropes inside the fences. The ropes in 2001-02 were closer to the fence than they are today.
I recall that ropes were used in early day-night ODIs at the SCG, because the lights did not illuminate the outfield completely.
There was a boundary rope for the Adelaide Test in 2002-03, but it was only about a foot inside the fence. It can be seen in video of Glenn McGrath’s famous catch.
the excellent "test-cricket-tours" website has gone down:
"Suspended" it says. In the past I have linked to this extensively
from my own database. Those links no longer work.
fellow who created this (Michael Ronayne) died a few years ago without
completing it. Most of the 20th Century series were complete, but
there were gaps: Pakistan tours were missing, among other things. Even so it
was an extremely useful source. I placed links to it on each series cover
page on my database website, and saved copies for most series (up to 1991) on
my own computer.
Test series up to 2015 had been posted, although some 21st Century series were missing or incomplete. I don't know if it will come back, but in case it doesn't I have started a scrape of the old copies of the website pages found on the Wayback Machine. There are hundreds of pages to download. I will post these on my database and change the links but this will take some time.
It is a lesson about ‘free’ data on the internet. It can disappear overnight without warning and never return.
I understand that there are some published booklets by the same author that cover the same territory. I must find them and see if they fill any of the gaps.
little snippets...I was watching a ODI from 1980,
and Richard Hadlee, as a fielder, did a 'slide pickup' of a ball that was
nearing the boundary. It is the norm nowadays, but I don't remember seeing it
very often that far back.
film I saw of a Test in 1961 included a 'tag team' return from near the
boundary, one fielder scooping the ball up to a nearby teammate who completed
the throw. Again, something that is routine now.
When Farook Engineer (on 18) was struck on the head by Andy Roberts at Delhi in 1974, the ball landed about 2 yards inside the boundary. Engineer retired hurt and lunch was called; he returned to the crease later and made 75.
18 October 2020
So, who bowled the most no balls?
Identifying the bowler who bowled the most no balls in Tests is not quite straightforward. Information from the major websites is very patchy. There is also the issue of how to count no balls; the protocol has varied over the years.
Nowadays, the number of no balls counted against the bowler is the same as the number of actual no ball calls by the umpires. This was not the case before 1998; before then no balls with runs off the bat were not counted as such, while multiple no balls (with added byes/leg byes) counted as more than one. For comparison of bowlers, I much prefer the modern counting, and this is possible with ball-by-ball records.
As it happens, no balls have become less common for various reasons (mainly umpires who don’t bother to check for them anymore). The greatest numbers come from previous generations of bowlers, after the front-foot rule was introduced around 1964. There are two contenders who are well ahead of anyone else – Bob Willis and Wasim Akram.
I have Willis’ career (17,357 balls) complete in ball-by-ball form. In this database, Willis was called for 932 no balls. In ‘classic’ counting it would be only 763 (no balls with runs off the bat excluded). The 932 is an extraordinary number and sets the bar quite high.
The case of Wasim Akram is more complicated. He bowled 22,627 balls, but I have only about 81% of his career ball-by-ball. There were 768 no balls calls in that data. I also have another 10% of his career with published no ball counts, but no ball-by-ball data. This adds 72 no balls, but these would be classically counted: the figure translates to 81 no ball calls, based on his typical pattern.
There is an additional 9% or so of Wasim’s career with no available data, so some estimating is required. Based on the patterns for the other 91%, the estimate comes to 936 no ball calls for his whole Test career. I am not sure how wide the put error bars should be, but Willis’s 932 would certainly be within the margin of error.
So for now, I cannot distinguish between the two bowlers. Wasim may ease clear if you include his wides – more than 40 to Willis’s 19.
The next bowler on the list appears to be Malcolm Marshall. Once again there is no exact number and some estimating is required; this produces an overall figure of 810 no ball calls.
I was watching some footage of the 1961 Ashes a couple of weeks ago. It occurred to me that this was the earliest footage of such quality that I have seen. By ‘quality’ I mean that the coverage actually captured most incidents of importance, and the viewpoint, in line with the pitch, made appreciation much easier.
about all other film that I have seen from around this time or earlier is of
indifferent quality. Prior to TV, the cameramen sent to the grounds obviously
had very limited amount of film; this meant that they either missed most
important events or only caught the aftermath of dismissals. The newsreels
used various editing tricks to try and paste together a narrative.
I don’t know how the 1961 footage, which originated with the TV broadcast, was recorded. Videotape existed in those days but I don’t know if it was being used in Britain at that time. The footage that I have seen looks like it was recorded by putting a film camera in front of a TV monitor. There are fragments of earlier such footage: some of Laker’s 19 wickets at Old Trafford, and England winning the Ashes in 1953 can be found online. They look like TV material. The 1961 material is much more extensive.
have the ABC coverage of Tests in Australia from 1958 to 1963, about an hour
per Test. This was taken by a single film camera; it misses a great deal in
terms of highlights but is still most interesting. For some of these Tests,
the film was processed and edited in a great rush each day and flown by
special courier to other capital cities, where it was shown as highlights
about 10pm on local TV. It was not possible to transmit TV from one city to
another in Australia (using coaxial cables) until the early-to-mid 60s.
[UPDATED see 21 November.]
some old films of the 1962-63 Ashes, an odd observation: at Adelaide, there
was a boundary rope maybe 10 metres in at the River
end of the ground. Very unusual for those days.
There have been no recorded instances anywhere since 1996; such strokes are much more unlikely now that grounds have been shrunk down.
Incidentally, there was only one hit for six in the whole 1958-59 series. That was by Fred Trueman off Richie Benaud at the SCG, and he was out next ball.
At Brisbane in 1970-71, Rod Marsh was caught off Colin Cowdrey, but a no ball was called because three fielders were behind square leg. Lou Rowan was the square leg umpire.
Wisden 1984 says that the first use in a cricket match of a full electronic scoreboard that could show replays was at the Victoria v England XI match in December 1982.
At Lord’s in 1956, Richie Benaud was struck on the pads by Brian Statham. There was a loud appeal, and umpire Lee immediately give Benaud out lbw. However, the ball, still in motion, rolled onto the stumps and the bails fell. Benaud was recorded as out bowled, as the Laws allow for the ‘bowled’ dismissal to take primacy here.
Also from 1956, spinner Tony Lock actually opened the bowling on the final afternoon at The Oval “…rubbing the ball first into the ground to remove the shine.” (Belfast News-Letter). Australia only had to bat 2 hours to save the match, but were in such a state of confusion against Laker and Lock that they almost lost, finishing on 27 for 5 off 38.1 overs when bad light stopped play.
The great double-century drought: when Greg
Chappell scored 204 at the SCG against India in 1981, it was the first Test
double-century in Australia for more than ten years, since Keith Stackpole
had made 207 against England at the
’Gabba in 1970. There had been more than 1,950 innings played in Australia in
the meantime. Moreover, there had been more than 560 innings since anyone had
passed 150, the score made by Derek Randall at the SCG in 1978-79.
I remember noticing the dearth of big scores at the time, and wondered where they had all gone. It just seems to have been one of those things.
The practice of altering playing hours and extending lunchbreaks in Pakistan, to accommodate Friday prayers, appears to have started in the series against India in 1978-79. I have notes that in two of the Tests (1 & 3), the first session of the Friday was “extended”, and it may well be that this applied to the other Test too. I don’t know how long the lunch breaks were.
It probably did not apply in the previous series against England in 1977-78. In the Karachi Test in that year, the first Friday session was 10:00 to 12:00. Lunch was in fact extended that day, but this appears to be due to the teams being presented to General Zia. Play restarted at 1:05.
20 September 2020
The Old Stump Scramble
I remarked a little while ago on the old, odd, practice of players ‘souveniring’ stumps, bails and balls at the end of a Test match, usually after an unseemly scramble. In 1946 Keith Miller even grabbed a stump while the ball was still in play, and ran the winning run with stump in hand. At Trent Bridge in 1948 Sid Barnes, thinking the match won after he hit a boundary, seized a stump and bolted for the Pavilion. He, and the stump, had to be hauled back out of the dressing room because Australia still needed one run to win. It took some time to restore order and complete the match.
I wondered when this practice ended, and found a reference in Trove. In September 1952, the Australian Board of Control ‘asked’ that captains instruct players not to do it anymore. For the upcoming South Africa series, umpires were instructed to collect the stumps and bails and return them to ground authorities. The authorities were then permitted to distribute the stumps and bails to the players, equally to each team.
So the on-field scramble had ended (in Australia) with the last Test of 1951-52 against the West Indies. A report from the final Test of that series says the souveniring had only been half-hearted anyway, perhaps because the series had long been decided.
The Board’s action followed the lead of the M.C.C., which had stopped the practice in England that year.
I am not so certain when the habit started. I couldn’t find and mention of it for the 1920-21 and 1924-25 Ashes series, but some souveniring went on when England won the Ashes at The Oval in 1926. Arthur Mailey, last out, stuck the ball in his pocket, while Herbert Strudwick grabbed the “last” stump. There are references to players taking stump souvenirs in the 1928-29 series.
Other countries also took part. There was the “usual scramble for souvenirs” at the end of the Bombay Test of 1948-49 (India v West Indies). In that case, the scramble may have tricked the umpire into calling stumps early, with India needing six runs to win. The scramble was reported in the 1951-52 Tests between New Zealand and West Indies, but I found no mention of it in newspapers reporting the 1952-53 New Zealand v South Africa Tests.
[UPDATE: Ashru has sent me a photo of a stump scramble when New Zealand won its first ever Test, against West Indies in 1955-56. It would have been quite unusual by that time, but it was a very special occasion as far as New Zealanders were concerned. On (special) occasion, stumps have been souvenired since then. I think Botham grabbed a stump when England won the Ashes in 1981, and there is that famous footage of Shane Warne ‘dancing’ with a stump in 1993.]
Largest innings without facing a maiden over (where known)
Warner’s innings is the highest for a team batting first. The Walters and Paynter innings involved 8-ball overs, so maiden overs were harder to bowl. Walters faced one over where he did not score off the first six balls, as did Paynter. The highest score with just one maiden is Ben Stokes’ 258 at Cape Town in 2015-16.
The most maiden overs found in a single innings is 36 by Dudley Nourse (208) at Trent Bridge in 1951. The number is a little uncertain because of a high number of unmarked byes and leg byes in the score. Bob Simpson faced 33 maidens and Ken Barrington faced 31, in the same Test at Old Trafford in 1964.
The number of maidens faced by Hanif Mohammad in his 337 is not known, but would probably well exceed the above figures.
The Lord’s Test of 1963 is famous for its finish, a draw with England nine wickets down, six runs to win, and with Colin Cowdrey at the non-striker’s end with his arm in plaster. The match would have had a much different finish, however, but for an oddity in the scheduling of Tests in England in those days. The standard hours were 11:30 to 6:30, but on the last day hours were shortened to a 6:00 pm finish, apparently to make it easier for touring teams to reach their next location. And so it was at Lord’s – the fifth day was shortened by half an hour and play finished at 6:00. [Correction: the final day playing hours were 11:00 to 6:00.]
I have been gathering more information on bowling ends and umpires’ ends for Tests before 1972, this time for Tests in England. Much of the data has come from scouring the online (subscription) British Newspaper Archive. Information is now virtually complete for all Ashes Tests since 1948. (For 1948, Barry Valentine’s analysis was very useful.) I have also gathered the necessary info for series involving West Indies in 1963, 1966 and 1969, and a few other Tests (including Edgbaston 1957 and Lord’s 1960). From 1971 onwards, the information is substantially complete, thanks largely to Bill Frindall, who started recording these things in his scoresheets from about that time.
The information is incorporated into ball-by-ball files and is being uploaded progressively. For Tests in Australia, the upload is complete for Tests (those that have bbb) since 1911-12. I have bowling end and umpire info on more than 60% of all Tests; this will end up online eventually I hope.
I hadn’t realised this before, but Lord’s is one of very few grounds to have an east-west pitch (Pavilion to Nursery end). The standard north-south is probably not possible at Lord’s, because the famous slope would play havoc with bowlers. Old Trafford used to have an east-west pitch too (Stretford to Warwick Road) with the Pavilion off to the north, but it was reoriented after 2010. The ends are now called the Anderson end (Pavilion) and the Statham end.
The MCG was originally an east-west ground. It was used as such in the Tests of 1877, but stumps had to be called at 5 pm (in March, very late in the season) because the batsmen were looking into the sun. Shortly after that, the pitch was realigned to north-south, and we ended up with a ground with huge square boundaries and much shorter straight boundaries. Football is still played on a (roughly) east-west axis on the MCG.
Eden Park at Auckland has a pitch running southeast to northwest.
When illness struck the team during the Bangalore Test of 1988, New Zealand took to the field with five fielding substitutes, including Jeremy Coney – retired from Tests but reporting for Radio New Zealand – and TV commentator Ken Nicholson. Only three bowlers were fit to bowl in India’s second innings.
We’ve all seen the old films of players rushing to souvenir stumps when a Test match finished. In the 5th Test of 1946-47, Keith Miller went one better. Three runs were needed to win when Colin McCool drove a ball from Dennis Compton. They ran two, and with an easy third run on offer, Miller seized one of the stumps at the bowler’s end. With the ball still in play, Miller ran the winning run with stump in hand. Miller handed the stump to Compton as they left the field.
When Ernie Toshack took 5 for 2 off 19 balls against India on a drying ’Gabba pitch in 1947-48, the only runs he conceded were off a thin edge that all but bowled Sohoni. Sohoni was out next ball.
Peter Petherick of New Zealand scored a place in the records by taking his first three Test wickets as a hat-trick, at Lahore in 1976. Petherick was also a “world class” Number 11 batsman. In the following Test at Hyderabad, Petherick hit his first ever boundary in first-class cricket, on the way to 12 not out. Prior to that he had played in 13 first-class matches and had scored just 17 runs.
‘handled the ball’ incidents are somewhat contentious, but when Mohinder
Amarnath was the first to be dismissed this way in an ODI, at the MCG in
1986, the batsman actually walked before the umpire could give a decision.
Amarnath had blatantly used his hand to swat away a ball that he had played
but was heading for the stumps. He knew right away that he had done wrong,
and when the Australians appealed he turned and headed for the pavilion.
Cricketing Understatement of the Century?
As Trevor Chappell prepared to bowl the notorious underarm ball against New Zealand in 1981, TV commentator Bill Lawry, perhaps taken aback, commented
“…possibly a little bit disappointing”
Richie Benaud was rather more forthright when presenting the highlights that evening:
“…a disgraceful performance…should never be permitted to happen again”. “We keep reading that the players are under a lot of pressure; perhaps they might advance that as an excuse…not with me they don’t. It was a very poor performance, one of the worst things I have ever seen on the cricket field.”
At Perth in 2015-16, David Warner and Usman Khawaja put on 302 in 63.4 overs without a maiden over being bowled. There were 91 consecutive non-maidens including parts of the previous and subsequent partnerships. The first day of 90 overs included only one maiden, the first over of the day. Warner did not face a maiden over in his innings of 253.
A couple of snippets from India’s tour of New Zealand in 1989-90, courtesy the “Test Cricket Tours” website:
• When selected for the New Zealand tour, pace bowler Vivek Razdan had already toured Pakistan with India. Even though he was now going on his second Test tour, he had played only one first-class match in India at that time. He had played two Tests (in Pakistan), but was not selected for the Tests in New Zealand. He faded from selectors’ favour and played for a few years in Indian domestic cricket.
• “Bedi’s reign as India’s first
‘cricket manager’ or coach was brief. His uncompromising approach became
renowned after making furious threats to throw members of the team out of the
plane into the sea (!) after losing to New Zealand, and after eight months he
23 August 2020
Shipperd Strikes Back
When Mark Greatbatch took 462 minutes (341 balls) to reach 100 in the Perth Test of 1989-90, it was the slowest century in terms of time in Australian first-class cricket history. As it happened, the previous record had only been set three weeks earlier when Greg Shipperd took 449 minutes (343 balls) for Tasmania v Western Australia on the same ground. I don’t know if Shipperd regarded this as a challenge, but only six days after Greatbatch’s marathon, Shipperd re-took the record with a century in 481 minutes (412 balls) against Victoria at Launceston. This remains the slowest century in Australia in terms of both time and balls faced. Outside of Test cricket, only two batsmen are known to have faced more than Shipperd’s 412 balls in reaching a century.
Shipperd’s earlier record came during an innings of 200 not out (in 708 minutes, 571 balls). I wonder if anyone making a double-century has taken longer over the first 100. In Tests, the slowest century to be turned into a double was by Grant Flower, who took 437 minutes and 340 balls to reach his century on the way to 201, against Pakistan at Harare in 1995, in a rather notorious match.
Here is an apparent error in the official umpire listings. For the 5th Test of 1951-52 (Aus v WI) the umpires are named as HAR Elphinstone and MJ McInnes. These names are given online and in the Wisden Book of Test Cricket. However, the official score from the SCG for this Test names HAR Elphinstone and RJJ Wright (name given as R Wright).
Searching Trove for "umpire McInnes" for the dates of the Test produced no hits, but Wright is named in newspapers from the time.
Cricinfo and Cricket Archive have been informed.
Another correction: some sources say that Thomas Flynn, who umpired some Tests in the 1890s, was born in Kyneton, Victoria in 1869. This is not the case. The actual year of birth appears to be 1849, and although he lived in Kyneton he was born in Melbourne or in Tasmania. (There was a younger Thomas Flynn born in Kyneton in 1869, but he was the nephew of the Test umpire.)
As such, Flynn cannot be counted among the youngest Test umpires.
Flynn died in Townsville in April 1931, aged 82.
Fewest Scoring Strokes to Reach Test 100.
Not surprising to see Astle and Gilchrist on such a list, but note that their innings are not their most famous high-speed centuries. McCullum’s record-breaking century off 54 balls heads the list, though. Beyond the above list the field is quite crowded, with almost 40 innings coming in at fewer than 35 scoring strokes. Also, this is a ‘where known’ record.
I have been doing some work to identify bowling ends in Test matches in Australia, and the umpires at those ends. (specifically, the end used for the first over of each innings). This information is generally not found in scorebooks prior to 1980, with the exception of Frindall’s scores (from 1968). Even Fergie’s scores lack the information, with the odd exception of the 1911-12 Ashes.
I am only looking at Tests for which ball-by-ball scores are available. Once that info is available, you only need to find a single incident in a particular innings for which the end is known, and everything else falls into place. Very often this can be gleaned from photographs. Prior to 1970, nearly all Test match photos published in Australian newspapers were taken from the northern (Pavilion) end of the grounds, and this can be confirmed by the direction of shadows.
Where possible, I have added identification of umpires for the first over of each innings. This is not always so easy for the old days, as umpires were not the celebrities that they seem to be regarded as today, and were only infrequently mentioned by name in connection to specific incidents. However, with the power of Trove it can be done for some innings. Moreover, once you have an umpire ID for one innings of a Test, the rest of the Test falls into place, assuming that the umpires followed the protocol of standing at one end for the first two innings of a match, and changing ends for the third and fourth innings. I have checked data where possible, and I think that Australian umpires have been sticklers for this protocol for a very long time. Ray Webster tells me that the practice pre-dates Test cricket.
I have completed this work from 1911-12 onwards in terms of bowling end, and from 1924-25 in terms of umpires, up to the late 1960s. The 1970s was already done, but with gaps that I will try to fill. There are several Tests for which I cannot find the umpire information, even with Trove. Trove also largely ends in 1954, so other sources have to come into play. The Sydney Morning Herald is available up to 1995 through the Victorian State Library, and Google newspapers has a partial but substantial collection of The Age. Trove still has the Canberra Times after 1954, and various foreign papers (The Times, Guardian, Times of India) are available online. I have video copies of Test highlight films, made by the ABC, for the 1958-59, 1960-61 and 1962-63 series, about 45 minutes per Test and all very useful.
For Tests in England, end ID and umpire ID is substantially complete from 1968 onwards, thanks largely to Frindall’s scores. I may be possible to push this back in time a little, but it may be more difficult than the Australian work. Other countries will be more difficult still, at least for years before 1980.
The updated ball-by-ball records will be posted progressively. I have done 1911-12, 1920-21 and 1924-25 so far.
I have been making some cosmetic changes to the presentation of data for 50s and centuries in my online database. It looks a little less cluttered. An example is here. I will post these progressively. The original versions were posted (a frightening number of) years ago and there may be some added information. I haven’t kept track of changes.
dropped (catches) most often in 21st Century Tests: 78 for Alastair Cook, 67
for Sehwag and Sangakkara. Sehwag is an interesting one since he played fewer
innings than the others. Perhaps it has something to do with the fact that he
hit the ball so hard.
Ben Stokes recently reached #3 in the ICC rankings even though his career batting average was only 38.5.This is the lowest average for a #3 batsman even when rankings are ‘backcast’ to 1954. (IMHO, the ranking system does not work well before then, because Tests were too infrequent). The previous low for a #3 was an average of 38.8 by Gundappa Viswanath in September 1972.
Lowest batting average by a batsman when he reached #1 batsman in ICC Rankings…
KR Stackpole (1972) 39.96
IM Chappell (1973) 40.90
GR Viswanath (1975) 42.19
GA Gooch (1993) 43.05
(since 1955). Gooch was also #1 in 1991 when his average
HJ Tayfield bowled to JH Wardle in 8 Test innings and dismissed him each time. Next best is Craig McDermott bowling to Patrick Patterson in 7 innings, dismissing him each time.
Charlie Macartney scored 231 and the Australians 330 runs in a two-hour session (lunch-tea) against Nottinghamshire at Trent Bridge in 1921. Don't know if that is a record for first-class cricket, but it must be right up there for a two-hour session. When Australia scored 721 in a day against Essex in 1948, the session totals were 202, 292 and 227.
On Australian TV, they have been showing some ODIs from the 1980s. An observation: displaying players' names on their cricket shirts started in 1988-89 (Australia/West Indies/Pakistan tri-series).
Most runs in the 50th over of ODI innings: MS Dhoni 315, (data complete), CZ Harris at least 243 (data missing for five matches)
In Tests, Decision Review outcomes are close to 60:40 in favour of batting teams, partly because bowlers are more likely to initiate reviews, and most reviews are unsuccessful.
Where umpires' decisions are overturned, it is a bit tighter; the outcomes favour the batting side 55:45. This means that wickets are slightly more harder to get than in the days before DRS.
29 July 2020
The First Test umpire
A couple of snippets of information missing from the very first Test in 1877 are the bio dates of one of the umpires, known as Richard Benjamin Terry. My wife Ann, a skilled hand at genealogy, has tracked him down and found that he was born Benjamin Terry in Bulwell; then a village, now a suburb of Nottingham. He was born on 25 Nov 1852 and was baptised was on 5 Dec 1852. Known as Ben, he spent a few years in Australia from 1876 and played three first-class matches. He returned home and moved to Scotland, played some more (non f-c) cricket and became a bookmaker. He died aged 57 on 10 July 1910 in Edinburgh.
are that Terry, aged only 24, was umpiring at the bowler’s end for the
first ball of that first Test in 1877 (unconfirmed). The pitch was oriented
east-west in those days, and the first ball was from the east end. Terry
remains one of the youngest ever Test umpires, although George Coulthard in a
Test in the following year was younger still.
standing at square leg for the opening over of the England first innings in
the inaugural Test in March 1877. The Melbourne Argus reported that Hodges’
final delivery [from the west end] was turned to leg and, as the batsmen set
off for a run, it was noticed that a bail had been dislodged. On appeal, Terry claimed that he had not
seen how that had occurred, nor did Reid at the bowler’s end, with the result
that Hodges was denied a wicket. It
was further reported that the batsman later confirmed he had made contact
with the stumps.”
Online and later published sources give Terry’s given names as Richard Benjamin. However, no original sources, including birth or death certificates, include the name Richard. That name should be deleted.
the sudden flurry of activity, there have now been 72 Tests at Old Trafford
since five-day Tests were introduced to England in 1948 (the three-day Test
of 1949 being the only exception). Most have been affected by weather to a
greater or lesser extent. I have made a list of Manchester Tests that lasted
into the fifth day and experienced no apparent weather interruptions. It is
Special mention should go to the 1995 Test against the West Indies, which lasted only four days. It had no weather interruptions apart from a short suspension of play due to “sun glare” from a row of greenhouses adjacent to the ground. Tea was taken 21 minutes early, but no net time was lost.
The 1955 Test finished with South Africa winning with nine balls to spare. The 1964 Test, a famous (perhaps notorious) marathon draw, went the distance with 551 overs bowled. Modern Tests almost never go beyond 450 overs. Bob Simpson was on the field for 548 of those 551 overs.
Although it was a four-day Test, 1934 is also deserving of a mention: "From first to last, the sun blazed down, the heat being at times almost unbearable" (Wisden). The next Ashes Test at Old Trafford (1938) was rained out without a ball being bowled.
Q. Shannon Gabriel's highest FC score is 20
not out. Does he have the lowest high score of all the players with minimum
100 First class matches?
A. Jim Griffiths played 177 matches for Northamptonshire from 1974 to 1986 with a highest score of 16.
Also Eddie Row, 103 matches, HS 16 and James Shaw (long ago), 115 matches, HS 18*. More recently, Ethy Mbhalata, 129 matches, HS 19 from 2002 to 2016.
None of these players played Tests.
I just came across a note from the Lord's Test of 1979. John Lever batted left-handed for one ball of his innings of 6 not out off 8 balls, facing Bedi, the second ball of 4 that he faced in that over (0001). A strange incident, but recorded clearly in Bill Frindall's score. The only other known cases of batsmen batting both left- and right-handed in the same innings are Salim Malik v West Indies in 1986, and Talat Ali at Adelaide in 1972-73. In both cases the batsmen were suffering from broken bones. Colin Cowdrey, with a broken arm, was prepared to bat left-handed at Lord's in 1963, but he did not have to face a ball.
George V was the first British Monarch to attend a Test match, Australia v South Africa at Lord's in 1912. The King arrived just before tea, and play halted while he was being seated. The players were 'presented' to the King, off the field, when tea was called a few minutes later. It was a Wednesday.
George V also attended Saturday play of Ashes Tests in 1921 and 1926. Players were presented to the King, off-field apparently. George V attended the 1924 Test v South Africa at Lord's but apparently not the 1929 Test or the 1928 Test v West Indies.
When the King attended the Lord's Test of 1930, play was interrupted for 10 minutes and the King met the players on the field. Ponsford was out 2 minutes after play resumed.
Players have been presented to Elizabeth II in quite a number of Tests. One of these, in 1977, was at Trent Bridge rather than Lord’s.
Reaching ODI 100 with a four most times...
V Kohli 12
S Dhawan 10
LRPL Taylor 9
HH Gibbs 8
KC Sangakkara 8
One other century by Gibbs is not recorded.
AB de Villiers 7
EJG Morgan 6
HH Gibbs 4
RH Sharma 4
Combined... de Villiers, Kohli on 14.
In Tests, Wasim Akram bowled at least 20 unsuccessful hat-trick balls, and two successful ones. Murali, who never took a Test hat-trick, bowled at least 17 unsuccessful hat-trick balls.
These figures do not include taking of the last two available wickets with consecutive balls, precluding any hat-trick.
The figures are uncertain because there could be some unidentified instances from the 1990s. These would be few in number.
In the Tests that I have ball-by-ball up to 2019, there are 1195 hat-trick balls and 38 hat-tricks, or one hat-trick in every 31 attempts.
Damien Fleming’s hat-trick in 1994-95 must one of the most unusual in first-class cricket. His third victim was Salim Malik, whose 237 is far and away the highest score by the 3rd wicket of a Test hat-trick; the next highest is 1. I suspect that it would be a record for first-class cricket also.
Only four century-makers have been involved in a Test hat-trick. Apart from Malik, the others were all the first victim, the highest being Javed Miandad with 163.
The highest score by the second victim of a hat-trick is 52 by PM Walker, in Geoff Griffin's hat-trick in 1960.
26 June 2020
Barry Valentine has produced a ball-by-ball record of the 1945 "Victory" Tests, which is available on the ACS website for those who are interested.
I get a rather fulsome mention.
The series has some interesting aspects. It began only a couple of weeks after the German surrender, while the Pacific War was still raging. One of the matches started on the day the Atomic Bomb was dropped in Hiroshima.
It is interesting that Bill Ferguson was around to score the series. I can only presume that he was in Britain for the duration of the War (he had scored the Tests of 1939, on a West Indies tour of England that was cut short by the outbreak of war). Fergie accompanied the 1945 Australian Services team to India for some matches and the team then toured Australia before disbanding. I have read that the players did not much enjoy all this touring, since they were keen on getting home to their families, but they were aware of the importance of this tour in re-starting first-class cricket.
I have completed the task of making notes on all cases of batsmen retiring hurt in Tests, recording the cause of each. There are 342 cases in my list, a handful of which are absent from online sources. The Wisden Book of Test Cricket was useful for descriptions of many cases, but many others required deeper reading.
The data probably justifies a longer article, but I will just summarise here some data on the bowlers responsible for ‘retiring’ batsmen. This is something I have touched on some years ago, but the data now is more certain.
Batsmen retire not out for a variety of reasons, many of which do not directly involve a bowler: pulled muscles, illness, cramp, collisions, or previous injuries are among them. The majority, however, retire after being struck by a bowled ball – 269 out of the 342. The bowlers involved are nearly all pace bowlers (the cases off spinners generally involve batsmen edging balls onto the face/head).
The bowlers inflicting retirement on the most batsmen are listed. I have taken care to identify the bowler responsible for the injury, which is not always the bowler active when the batsman actually retires. There are some cases where the batsman retired some time after being struck.
Bowlers Causing Most Retired Hurts in Tests
While Courtney Walsh leads in this category, Colin Croft stands out with six retirements in just 27 Tests. It is also fair to say that Croft was regarded as the most feared and dangerous bowler of his time.
Other bowlers with high ‘strike rates’ include Silvester Clarke (27 per 100 Tests), Azeem Hafeez (17), Harold Larwood (14), and Neil Adcock (12). However, these bowlers only caused three retirements each during their careers.
The absence of active bowlers from the list is a sign of the decline in batsmen retiring hurt, which in the past decade has been less than half the rate of the peak years 1975-85; this is attributed to improvements in protective equipment. There has, however, been a recent increase in cases due to the increased concern about concussion and its long-term effects. The availability of full substitutes is a corollary.
Which century partnership has one partner contributing highest percentage of the runs in Tests?
Sanath Jayasuriya (253) scored 89 out of a partnership of 101 in 2004; there were 11 extras, and his partner Dilhara Fernando made 1. You would expect that (at 87.1%) this would be the record, but Mike Hussey scored 88.8% of a stand of 107 with Glenn McGrath in 2005-06 at the MCG.
Dennis Lindsay scored 71.0% (157 runs) of a 221 stand with PL van der Merwe in 1966-67. For 300+ stands, Wasim Akram is unchallenged with 70.3% of the 313 with Saqlain Mushtaq at Sheikhupura in 1996 (220 out of 313)
Honourable mention: Dennis Compton 164 in a stand of 192 with Trevor Bailey at Trent Bridge in 1954 (85.4%).
Here is a rare reference to the first 1877 international as a 'test match'. This was in a New Zealand newspaper; I haven't seen the term 'test match' used in relation to this match in Australian sources, although it was used occasionally in reference to other matches or other sports around that time. This reference was found by Peter Huxford. Peter has also found the phrase ‘test match’ in Australian newspapers in the early 1880s; while they sometimes refer to actual Test matches (as per the modern canon) the references are rather scattered and irregular.
In my recent searches, I have seen the use of the term 'test matches' randomly in the NZ papers for even provincial matches.
What is the highest score by a batsman off his first 100 balls in a Test innings?
An interesting question with a strange answer: Brendon McCullum - twice.
For exactly 100 balls, McCullum reached 145 off 100 balls against Sri Lanka at Christchurch in 2014 on the way to 195. For less than 100 balls, it is 145 by McCullum once again, this time off 78 balls at the same ground, against Australia in 2016. He was out next ball.
Ross Taylor (138) was on 137 off 100 balls at Hamilton in 2010 against Australia, and Roy Fredericks reached 134 off 97 also against Australia, Perth 1975-76, on the way to 169.
In ODIs, Shane Watson scored 185* off 96 balls against Bangladesh in 2011. For a batsman facing his 100th ball, AB de Villiers had 172 against Bangladesh in 2017.
Highest Test partnerships without extras:
Two Test partnerships of 222, both for the 4th wicket:
Hazare (89)/Manjrekar (133) Leeds 1952 v England
(82)/Kohli (140) Hyderabad 2017 v Bangladesh.
In the Pakistan v New Zealand series in 2018-19, Pakistan had five 5WIs to New Zealand's one, and four centuries to New Zealand's two, but New Zealand won the series 2-1.
First use of StumpCam:
Stump Cam was an innovation of Kerry Packer's Channel Nine network. Allan Annual 1989-90 (Allan Miller) records the first use in the 1989-90 SCG Test v Pakistan. (My own memory was that it was older than that, but I think I was confusing it with the Stump Microphone which goes back to the early 80s.)
first overseas cricket televised live in Australia was part of the final Test
of 1972. I watched it, and remember Rod Marsh swinging his arms round and
round as he ran the winning run. In retrospect it is strange that they didn't
televise more of that series, given that the technology existed.
was so keen in those days that many times I sat up late into the night
listening to the radio broadcast.
was the first overseas series shown in full. Colour TV had just started in
The first use of the term "test match" was for some matches of the Stephenson tour of 1861-62. The term was coined by Tom Wills. Those matches were against odds and certainly not regarded as Tests today.
A quick look through Trove for "Test match" gave no hits for 1881-82, but a number of hits for 1882-83 including the first Test. It still was not a common term. There were mentions for 1884-85, but few and far between. The Shaw/Shrewsbury tour book uses the term test match (once) in relation to the 3rd Test, but not the first two Tests. In fact, the touring team did not regard the first two Tests as authentic.
The term became more common from 1886-87 onwards, particularly from 1891-92.
A quick search of The Times and the Guardian produced no hits for 1890 or 1893, but a number of hits for 1896. The surviving scorebook for the 1890 tour does not seem to use the term Test match, but the 1893 scorebook does so.
Australian newspapers were using the term regularly when reporting Tests in England from 1884 onwards, but not in 1882.
See update above - 26June.
22 May 2020
Here is a list of players who played Test cricket with a notable disability or chronic condition, for all or most of their careers. Sporting injuries are excluded.
Credit to Michael Jones for initiating the list; others have contributed. Readers are invited to suggest additions.
(I am not sure about Murali, who rather benefited from his condition)
Highest average Test partnerships. Unbroken stands counted as ‘not outs’.
DG Bradman 71.1
H Sutcliffe 56.7
RT Ponting 55.9
JB Hobbs 52.5
ML Hayden 51.8
Younis Khan 51.7
Shoaib Mohammad 49.9
SJ McCabe 49.8
DPMD Jayawardene 49.8
GC Smith 49.7
minimum 30 Tests
Factors: batting average, batting averages of team mates, openers and higher order bats favoured, because they bat less with tailenders. Slower batsmen do a little better than expected (note Shoaib Mohammad). Their partnerships tend to be longer, and when batting with a fast batsman, are worth more runs.
22 April 2020
The nervous 90s and Beyond
Here is a graph of the frequency of dismissals over a range of scores (75 to 130) in Tests. Strange to think that I first made a graph of this more than 20 years ago, to look for scores where the number of batsmen getting out was unusually high or unusually low. Well, I have updated it now.
The relationship between score and number of dismissal is an exponential decay curve, best shown on a log-linear scale. An exponential line has been fitted to the data; this comes out as a straight line on a log-linear graph. It can be seen that this line fits the data quite well across the range of scores used.
Of particular interest is the frequency of scores before and after the magic score of 100. In short, there is a deficit of scores in the range 85-99, and an excess in the range 100-115, and more specifically a deficit from 95-99 and a surplus from 100-105. These observation can be quantified as departures from the fitted curves, in a table
This data shows that the “Nervous Nineties” is a myth, or rather, that whatever nerves occur are often beneficial for the batsman. In general, a batsman has considerable lower chance of dismissal in the 90s than just after reaching 100. In the range 94 to 99, the chances of dismissal are depressed by about 10 per cent. From 100 to 105 the chances are elevated by about 13 per cent. It is curious however that the number of dismissals specifically on 99 is very close to the expected value.
I mentioned a while back that I have a new format for ball-by-ball files in my database, limited to one over per line. I have now completed the uploading of the new format for Tests prior to 1940. The work will continue, but at a measured pace. I still find it necessary to check each bbb record for problems before uploading.
21 March 2020
Highest Innings Starting and Finishing in Same Session
Nearly all of these were post-tea sessions. You will see from the batting times that at least four of these innings benefited from extended session times. In the old days, this could happen due to the flexibility of the tea break, which could be called very early if there was a change of innings. More recently, post-tea sessions are often extended to meet minimum over requirements.
have done a reconstruction of Shahid Afridi’s famous century in Nairobi in
1996 (102 off 40 balls), and posted it here. The reconstruction uses the
ball-by-ball record in the Cricinfo Archive, supported by
other sources. Cricinfo lists every ball faced by Afridi, but does not
include his partner Saeed Anwar, and it does not
have the first wicket partnership of Anwar and Saleem Elahi. There is some
quite good video on YouTube that allows most of these gaps to be filled,
resulting in a reconstruction that covers the first 20 overs of the innings.
While a few of the overs remain speculative, the reconstruction reproduces
all known stats for the first 186 runs, including Anwar reaching 50 off 47
balls with 8 fours.
14 March 2020
What is the most common number of runs conceded by bowlers? A bowler taking, say, four wickets in an innings will rarely concede less than 15 runs or more than 150, but somewhere in between the frequency will peak. It is quite interesting to look at the data that emerges from many Tests…
Most Common Runs Conceded by Bowlers
Data for Test innings since 1907. For 3 wicket data, bowlers bowling fewer than 10 overs excluded. More than 10,000 bowling returns were used in the calculation.
The most striking thing is that, by and large, the typical runs conceded hardly changes with the number of wickets taken. Nor does the spread of results, expressed as Standard Deviation. The data for nine wickets is the exception, for reasons unclear; note that the only bowlers taking 10 wickets conceded 53 and 74 runs respectively (average 63.5), rather similar to the data for 3 to 8 wickets.
For wicket counts lower than 3, the medians are lower, but the results tend to be muddied by a large number of ‘small’ analyses from single short spells. This data is not shown.
Graphing the data for each level produces something akin to a set of Bell Curves, although somewhat skewed by the fact that the data is bounded at zero but unbounded at the high end, giving each curve a ‘long tail’.
I said ‘akin to’ a Bell Curve but there is something else at play here. It appears that the data does not smoothly fit a Normal Distribution. The upslope seems more linear than bell-like. Better statisticians than I might be able to work it out.
1 February 2020
I have done some work on tidying up my session-by-session database, and fixing a few anomalies. In doing so, I came across an individual century in a session that I believe was previously unrecognised. At Leeds in 1924, England was 122 for 1 at lunch on the first day against South Africa. Patsy Hendren came in shortly after lunch and was 104 not out at tea, out of a score of 303 for 6.
Even though I have a scorebook copy for this Test, I had missed this one, mainly because the scorebook is in a mess and contains so many anomalies that my analysis has remained incomplete. Nevertheless, the century in a session is clear from newspaper reports that are now available on the British Newspaper Archive online. Hendren finished with 132 off about 160 balls.
I was also able to compile a list of batsmen making 99 runs in a session, which I think should be complete.
99 Runs in a Test Match Session
BB refers to total balls bowled in the session.
In some of these cases, the batsman benefited from considerable extensions to the session, but I have included them anyway. That first day of Test cricket way back in 1877, featuring Charles Bannerman, was effectively just two sessions. Play did not start until 1:00 pm, and ended at about 5:00 pm, possibly because the batsmen were looking into the late afternoon sun (in mid-March) on the east-west pitch (the pitch orientation was changed to north-south a few years later). A lot of balls were bowled in that second session, but it is nice to put Bannerman on the list.
Trumper, in his famous 214* at Adelaide, made 98 between lunch and tea and then 99 between tea and stumps. Both sessions were only about 90 minutes long.
Kapil Dev made his 99 in the final session of a dull draw. He had been one not out at tea, and was allowed to complete his century, but not his session century, before the match was called off prior to the scheduled close.
Some adjustments and addition have been made to session data in the database. Most changes are minor, although there has also been some data added. Affected series are:
1907 Eng SAf
1912 Eng Aus
1924 Eng SAf
1953 WI Ind
1955 WI Aus
1958 Ind WI
1959 Pak Aus
1976 WI Ind
1978 Ind WI
1984 Pak Ind
Another block of Tests, from 1982 to 1986, has now been completed and added to the database. Most of this period is based on surviving scorebooks, but there are still gaps. The last series in this block, Pakistan in Sri Lanka in 1986, is particularly short on extended data, perhaps as much as in any other series in Test history. If anyone has any detailed info on this series, by all means contact me!
10 January 2020
From time to time I hear of teams that choose not to use the DRS (Decision Review) when doing so would have resulted in the reversal of a decision. I haven’t seen any stats on this, so I set out to acquire some, by sorting through Cricinfo texts for the last couple of years.
I looked through the texts for 90 recent Tests, searching for occurrences of the word “review” or similar on balls where no review occurred, and phrases that indicated that the bowling team should have reviewed a “not out” or a batsmen should have reviewed an “OUT”. This is not an exact process; it is probable that instances were missed. We should also recognise that any checking of decisions in the absence of an official review is informal, so we cannot be absolutely certain that the 3rd umpire would have upheld the review (overturned the decision).
All ‘hits’ from the search process were checked, and confirmed or rejected, by careful reading. There were also a number of instances where recognition of the failure to review was only noted several balls later, when the DRS data became available. Nevertheless, I would think that I identified a large majority of cases, enough to compare teams.
Overall, there were 67 definite or probable cases of ‘failure to review’ – where a review would have been successful – in 90 Tests. In 13 of these cases, the failure was that the team had no reviews left; they had failed in too many reviews earlier in the innings. In the other 54, the team had an opportunity to use the DRS but declined to use it. Overall, 55 failures involved lbw decisions – the rest were catches.
The breakdown by team is as follows…
The figures mean that Australia (for example) had 14 ‘failures’. Two of these were due to running out of reviews, so Australia chose not to review on 12 occasions when decisions would have been reversed in its favour.
If you were under the impression that Australia has tended to blow its reviews, there is good support for that. The above figures include the 2019 Ashes, during which Australia failed to review six times when a review would have gone in their favour; they also had one case of running out of reviews. This was exacerbated by Australia having 13 consecutive bowling reviews go against them in the Ashes. Australia’s only successful bowling review in the series came against a tailender, and that was just a few overs before the end of the series.
did a little better in batting reviews in the Ashes, getting three decisions
out of 18 overturned. That is still a very unimpressive success rate.
I have posted a list of review failures here. If anyone can add to it (for Tests in the last 2 years) please let me know.
A contact in India, Gulu Ezekiel, has sent me a copy of an interview with Col Hoy, published in an Indian cricket magazine (Cricket Quarterly Jan-Feb 1978).
It contained an interesting item about the Brisbane Tied Test (umpired by Hoy) that I didn't know: Hoy says that the scoreboard at the ‘Gabba missed a run during the last over, and showed the West Indies winning the match. The operators had missed a bye off the fourth ball of the over. The scorers, who were probably a bit snowed under at that point, had not called the scoreboard to correct the error.
It is not mentioned in Fingleton's book or the newspaper reports that I have on hand. I wonder if anyone has read about this elsewhere.
The reaction of the West Indies players at the end suggest that at least some of them thought they had won the match.
Sreeram has noted that Keith Miller had once hit the first ball of a Test day for six (Adelaide 1946-47, Day 4, off a no ball bowled by Doug Wright) and asked if I knew of any other cases. To my surprise I was unable to come up with anything apart from Chris Gayle hitting the first ball of a Test for six against Bangladesh. So Miller is the only known overnight not out to do this.
also tells me that during Ben Stokes' century at Leeds he scored 70
consecutive runs off the bat (61* to 131* plus a wide) scored between
Archer's last four and Leach's single.
6 December 2019
Here's an odd coincidence...
The sharing of the strike can be an important factor in some innings. Most large innings fall in the range 45-55%, but there are some outliers.
I figured out a way to easily calculate % strike received for major individual innings, without rearranging my data. So I calculated this stat for all the centuries and half-centuries that I could, over 3500 Test centuries in all (out of 4100).
Here is the coincidence, for Test centuries:
Lowest % strike: 36.3% AC Gilchrist 101 Port of Spain 2003.
Highest % Strike: 66.3 % AC Gilchrist 113 SCG 2004/05
There have been over 770 century-makers, so seeing the same batsman at both extremes is strange indeed.
One factor involving Gilchrist is that innings with few balls faced tend to have a wider spread in terms of the strike, and Gilchrist faced fewer balls in his centuries than just about anyone. Longer innings tend to regress toward the mean; it is very hard to farm the strike for extended periods.
Highest % Strike: Centuries
The figure for Sinclair is only approximate.
Lowest % Strike: Centuries
The extremes for half-centuries…
Ashraful’s 67 was an extremely fast innings; domination of the strike is much more likely over short periods.
Intikhab’s innings was during a famous 9th-wicket partnership at The Oval in 1967, which is also represented, from the other perspective, in the century by Asif Iqbal. I also remember watching Asif farm the strike at the WACA in 1978-79; he was the most skilled batsman in this respect that I have seen.
Dropped Catches Report, at last
After a long layoff, I have managed to update my database of missed chances (catches and stumpings) that I have been maintaining since 2001. (Based on searches of Cricinfo’s texts. These are wonderful; however, the searches are tiresome work and I wish Cricinfo’s commentators had a way of ‘tagging’ chances. It would make it so much easier.)
There is enough data in the update to make a historical comparison of wicketkeepers in this century. The results are interesting, I think.
The Best Wicketkeepers of the Century: Fewest Missed Chances
Minimum 50 chances as wicketkeeper (32 wicketkeepers qualified).
Catches and stumpings are only counted for those matches where missed chance data is available (not necessarily total career). In the case of Rashid Latif, that makes the numbers rather provisional, because only 18 Tests out of his 37-Test career have data. This includes a couple of Tests from the 1990s where data was logged by Bill Frindall. I took a close look at Rashid’s stats because Rashid himself asked me about them.
For Adam Gilchrist, some early matches are missing.
For most players data ends in May this year, except for Tim Paine whose data includes the recent Ashes. Paine’s figures are remarkable; we will see if he can sustain this (Gilchrist and Boucher were also in single digits at the same stage of their careers, but both faded a little in later years)
It is also interesting that Matthew Wade, who was Australia’s keeper in between Nevill and Paine, had a much higher drop rate of 17%. Wade, of course, is a much better batsman than either of the others and is now back in the team as a specialist batsman. I did calculate once that the extra runs conceded through Wade’s missed chances (compared to Nevill) almost exactly counterbalanced the extra runs that he scored.
At the far end of the scale, about half a dozen wicketkeepers have missed over 25% of their chances. Mushfiqur Rahim missed over 30%.
After a lot of thought, I have decided to change the layout of the ball-by-ball records of matches in the Test Match Database. Previously, I presented data with two overs per line. The saved space and was quite neat in presenting overs at each end in a side-by-side configuration. However, I finally decided that this layout was just too difficult to read. I had thought that readers could figure out the complexities if they really wanted to, but it was all a bit too difficult.
The new layout presents one over per line, rather like linear scoring. There are also line breaks where wicket(s) occur during an over, and at the end of every session, so that the exact score at these events is clearly displayed. An example of the new layout is linked below.
The new layout uses more pages in the pdf format, but I hope it is more user-friendly. Eventually, I will redo all the old ones, about 700 of them (!)
I have just reached Test # 1000 in my database!
I have added some bits of data to certain post-War series in my database. Some if this data came from Ashru. Other data concerns batting milestones, particularly times for half-centuries. Series affected include Eng v NZ 1949, Aus v Win 1951-52 and v SAf 1952-53, and series in India and Pakistan in 1954-55 and 1958-59.
14 November 2019
One of the most fascinating innings from the ‘Golden Age’ of Test cricket is Jimmy Sinclair’s 104 against Australia at Cape town in 1902. It was one of the fastest innings of its day – it would even be the fastest century of all time if some reports are to be believed. In truth, though, the record-breaking claims are very dubious.
I have studied Sinclair’s tour de force in the past, and some years ago posted online my reconstruction, based on contemporary newspaper reports. Recently, Robin Isherwood sent me a copy of another over-by-over analysis of the innings, made many years ago by R.H. Curnow. Curnow also based his analysis on newspapers, perhaps a more extensive set than I had access to. I have posted the resulting over-by-over score here. In short, the two versions are substantially in agreement with regards to Sinclair’s innings and its statistics, although there are differences in detail.
One contentious aspect of this innings is that some newspapers state that Sinclair’s innings lasted an hour or less; this would make it the fastest-ever Test century in terms of time. However, my analysis and Curnow’s agree that there were far too many overs bowled for this time to be possible, and 60 minutes is in clear conflict with times given for other milestones in the innings, stated in the same reports. One report said 80 minutes rather than 60, and this seems to be correct. The error may have arisen if the dismissal of Shalders was used as Sinclair’s starting time (leading to a time of 60 minutes), when in fact Sinclair had come to the wicket at the dismissal of Smith about 20 minutes earlier. Reports saying that Sinclair reached 50 in 35 minutes are similarly almost certainly wrong; the real figure is 55 minutes, in all probability.
I wondered if there had been a 20-minute tea break, but no report mentions any breaks in the innings. In those days, there was usually no tea break if a change of innings occurred after lunch, which was the case here.
A remarkable aspect of the reporting is the detailed account given in the Cape Argus. Amazingly, the report, covering the entire innings, was published on the same day as the innings (Monday Nov 10, 1902) even though Sinclair’s innings did not end until 5:40 pm! The Argus was an afternoon paper with multiple editions, and apparently they held the final edition open until the cricket report could be completed. Reports were sent from the ground to the office by bicycle courier.
I have a photocopy of this report, sent to me by Ross Smith many years ago; unfortunately it is sometimes hard to read, and I haven’t been able to get a better copy. I presume that Curnow had access to a clear version. Anyway, here is my interpretation of some of the time features of the innings, based on reports from five newspapers:
4:15-4:20 pm, over 31. CJE Smith out at 81/2. Sinclair in.
4:25 pm, over 36. South Africa 100 in 95 minutes.
4:30-4:35 pm, over 38. Shalders and Twentyman-Jones out. 115/4. Sinclair 26 off ~22 balls.
4:50 pm, over 44. Llewellyn out 136/5.
5:15 pm. Sinclair 53 off ~50 balls, 55 minutes. Over 49.
Overs 51-52. Sinclair hits 34 runs in 2 overs.
5:30 pm over 55. South Africa 200 in 160 minutes.
5:37 pm. Sinclair 100 in 80 minutes, 70-75 balls. Over 57
5:40 pm. Sinclair 104 in 83 minutes, 75-80 balls. Over 58, stumps called.
Uncertainties about balls faced are unavoidable, because dot balls are mostly not mentioned in reports, even though we have a good over-by-over account. In overs where singles or threes are described but the specific ball numbers are not, dot balls are distributed in what seems a reasonable fashion. It seems fair to assume that Sinclair faced fewer dot balls than his batting partners, given that he was making far more scoring shots.
Fast Centuries, Slow Times?
I was looking at some Tests from earlier this century when I came across some odd stats for a century by Adam Gilchrist at Port of Spain in 2003. Gilchrist reached his century off 104 balls, impressively fast as usual, yet he it took him 208 minutes. He received only 36.5% of the strike during his innings; in particular, he received little strike late in his innings, while batting with Darren Lehmann (160) and Brad Hogg (17*). Gilchrist faced only 31 out of the last 120 balls of the innings, which was declared closed when he reached his century.
I decided to take a look at centuries with the most extreme ratios of minutes to balls faced. Gilchrist is the leader here.
Test Centuries: Highest Ratio of Minutes batted to Balls Faced
At the other end of the scale we have innings from long ago, when over rates were much higher…
Lowest Ratio of Minutes batted to Balls Faced (where known)
[Note that I only have the requisite data on about 70% of early centuries.]
One point that I would add is that while balls faced is rightly recognised as the best way to compare the speed of innings, minutes batted should not be ignored. The latter is an important element of the spectator’s experience. A two-hour century will generally be more memorable than a three-hour century, other circumstances being equal.
Generally, it is very hard to maintain a severe imbalance in strike over a long period, but evidently there are exceptions. I don’t know if Gilchrist’s century is the most extreme in % Strike, but I may report on that later.
A surprise to me, worth recording...
In the 2002-03 Champions Trophy (ODI) in Sri Lanka, lbw decisions were frequently referred to the 3rd umpire. Shoaib Malik was the first batsman given out lbw this way, on 12 Sep 2002.
Back then there were fewer hi-tech aids, and the 3rd umpire was simply making his decisions from conventional replays.
Many catch decisions were also referred to the 3rd umpire; almost all ended up 'not out' because the available vision was inconclusive (in the days before HD TV) and the batsmen got the benefit of the doubt. There were complaints about this and about the delays it caused.
The lbw experiment was shelved after this series. The more sophisticated DRS was trialled in 2008 and introduced in Tests in 2009.
Tim Paine recently scored his first first-class century for 13 years (125 matches). This ranks pretty high in the longest intervals between centuries, but not at the top.
Meyrick Payne of Middlesex, like his near namesake a wicketkeeper by trade, scored a century in 1907 and his next in 1927. For a career uninterrupted by War, Arthur Sims went 17 years between centuries. His second century, in 1913-14, was notable for a world record partnership of 433 for the 8th wicket with Victor Trumper.
Fred Titmus went 293 f-c matches between centuries, from 1965 to 1976 (age 43). He had made his f-c debut in 1949.
Some years ago I did a study of some of Bill Frindall's scores that recorded shots that went off the edge (as Frindall saw it). I logged the edge shots from 27 Tests. FWIW, there were 1443 runs off the edge out of 25,156 total runs off the bat - about 5.7%.
Rohit Sharma made 176 and 127 in the recent Test at Visakhapatnam, repeating exactly the scores of Herbert Sutcliffe at the MCG in 1925. It is only the second time that a century in each innings has been repeated exactly. The other was Inzamam making 109 and 100* at Faisalabad in 2005, matched exactly by Azhar Ali at Abu Dhabi in 2014.
Only two batsmen have made higher scores in both innings than Sharma (and Sutcliffe): Brian Lara with 221 & 130 in 2001 and Greg Chappell with 247* & 133 in 1974.
There was also Andy Flower 142 & 199, if you reverse the innings.
Most dismissals by a fielder/bowler pair in first-class cricket: I get 356 for Ames/Freeman. Next is FH Huish/C Blythe on 320 and Hunter/ Rhodes on 307.
The above figures include a large proportion of stumpings. For catches alone I get 252 for George Dawkes off Les Jackson for Derbyshire. I also get 250 catches for Edward Brooks off Alf Gover (Surrey).
before 1984 only)
17 October 2019
Long-time correspondent Ashru has reminded me of an unresolved anomaly in the score of the Trent Bridge Test of 1950, and pointed out that Brodribb discussed this incident briefly in Next Man In (1952).
Day 3 of this Test ended when rain interrupted, after Reg Simpson had hit the first ball of an over for three. When play restarted after a rest day, there was confusion over who should bowl and who should face. First Ramadhin, then Valentine, were told to bowl, before the scorers (Ferguson and Wheat) ruled that Ramadhin had to finish the over. Unfortunately he then bowled to Simpson again, so the wrong batsman was facing anyway.
The surviving score does not resolve matters satisfactorily. It seems clear from the score that only Ramadhin and Valentine bowled between tea and stumps. The overs are not numbered in the score, but Valentine must have bowled the odd-numbered overs, starting at Over 37, and Ramadhin the even; this preserves the correct sequence of scoring strokes for the batsmen, which otherwise goes haywire under any other bowling order. There were no extras in the session.
The main problem in the score is that, after Ramadhin bowled Over 48 to Washbrook, the three by Simpson follows immediately, off the first ball of Over 49, apparently with Ramadhin bowling again. There are no other available overs in the recorded score to insert after Over 48. The scores published in newspapers next morning reproduce exactly the bowling figures in this scenario, recording 6.1 overs for Ramadhin and 14 for Valentine.
Tea-Stumps Day 3, Trent Bridge 1950