30 March 2009
More columns written for The Age
(19 February 2009)
A couple of previous columns, on the subject of team performances, have mentioned a study of missed catches and stumpings in recent Tests. Here are a few more records, extended to individuals, from late 2001 to the present. Who has benefited most from dropped catches, and who has suffered the most?
Dropped most times: 39 – Virender Sehwag (India). Most for an Australian: 31, Matthew Hayden.
There is something to be said for hitting the ball hard. Sehwag has been dropped almost as often as Tendulkar (22) and Dravid (23) combined.
Most missed chances off a bowler: 74 off Danish Kaneria (Pakistan). Most off an Australian bowler: 51 off Brett Lee.
Danish is adept at creating chances for fieldsmen at short leg, but they accepted all too rarely.
Highest percentage of missed chances (bowler): 42% off Mohammad Rafique (Bangladesh, 37 missed, 52 taken). Highest for Australians: 27% off Jason Gillespie. (minimum 30 wickets)
Spare a thought for Nathan Bracken, who has seen seven catches dropped and only seven taken off his bowling in Tests. Among those let off were Tendulkar (twice), Sehwag (twice), and a sitter off Brian Lara. Bracken’s career may have followed a different path if they had been taken.
Most missed chances in the field: 50 by Adam Gilchrist. Most by a non-keeper: 36 by Rahul Dravid (India).
Gilchrist, of course, saw more edges flying his way than any other player.
Most ‘expensive’ dropped catch: 297 runs. Inzamam-ul-Haq, against New Zealand in 2002, was dropped on 32 by Robbie Hart, and went on to make 329. Honorable mention to Kumar Sangakkara, who was dropped on 0 against Zimbabwe in 2003/04, and made 270.
Highest percentage of missed chances (fieldsman): 41% by Alastair Cook (England; 24 dropped, 34 taken). Highest for Australia: 29% by Shane Warne (20 dropped, 49 taken). (minimum 40 chances)
The figure for Warne was a bit of a surprise, although I do remember quite a few misses late in his career.
Lowest percentage of missed chances: 7% by Chris Read (England; 3 missed, 38 taken). Lowest for a non-keeper: 13% by Graeme Smith (South Africa; 15 missed, 102 taken).
Read, who has played only 15 Tests, is also one of the finest keepers all-time in preventing byes. It’s a shame that in the modern game the very best keepers often can’t hold down a Test place.
Would You Believe?
Andy Blignaut (Zimbabwe) was dropped five times in his 84 not out against India in 2005. Historically, this has been exceeded in the distant past. In 1882/83, Australia’s George Bonnor, in making 87, was dropped seven or eight times, including five off the bowling of AG Steel, and Bill Ponsford was dropped six times (three by Bob Wyatt) on his way to 266 in his last Test in 1934.
Blignaut’s five, however, included a “hat trick” of dropped catches off three consecutive balls (from Zaheer Khan), probably unique in Test history.
UPDATE: Wide Wally points out that Geoff Arnold suffered a hat trick of dropped catches at Old Trafford in 1972. There were two batsmen involved (Stackpole and Francis) so Blignaut is still the only batsman known to be so lucky.
(27 February 2009)
Sometimes – not often – things change quickly in Test cricket. Just four months ago, Australia enjoyed a big lead over South Africa in the ICC team rankings (they led by 22 points, which seems a lot; the calculation method is peculiar, involving various performance weightings). Now the top two teams are, in effect, playing off for the top spot, a rare event. It’s worth looking at how the teams got to this position, in terms of conventional averages.
The chart shows averages for Australia and South Africa over the last twelve years. The measure used is the difference between team batting average and bowling average. This figure can be positive or negative depending on team fortunes.
2008 figures include early January 2009.
Australia has spent the last 20 years in positive territory. 2007 was the best result in decades (although there were relatively few Tests played) but last year was the worst result since 1988, and is only positive if you throw in the Sydney Test in January.
South Africa have also had a good decade, but they have almost always trailed Australia until just now. The sudden turnaround would be hard to believe except that the two teams have actually met during the crucial period, and so far, South Africa are looking like the Number One team in the world.
In Australia, the suddenness of the declining fortunes, and injuries to key players, have produced a scramble for new talent not seen for decades. At the peak of its stability from 2000 to late 2002, Australia introduced only one new player (Brett Lee) in 30 Test matches. Combined team experience peaked at 800 Tests during the last Ashes series in 2006/07, but is now down to 281, the lowest level in 20 years since the Adelaide Test of 1988/89 vs West Indies (235 Tests team experience).
Would You Believe?
Australia have introduced five debutants in the space of two Tests (Doug Bollinger, Andrew McDonald, Phil Hughes, Marcus North, and Ben Hilfenhaus). The last time this happened (leaving aside the Packer upheaval of 1977-79) was in two Tests in 1964/65, separated by several months, against Pakistan and then the West Indies. In Johannesburg we have three debutants in one match, a number last seen at Adelaide in 1985/86 vs India (Merv Hughes, Geoff Marsh, Bruce Reid).
(3 March 2009)
In a dramatic week for cricket, there was also action on the smaller stage of Test match statistics. There was an avalanche of giant scoring: in eight days, there were five scores of 600, two of them going on to 700. This is an extreme case of a modern phenomenon: there have now been 44 scores of 600 in this decade, and there have been more totals over 700 in the last five years than there were in the previous fifty.
The soaring incidence of large team scores is shown in the Table.
It’s remarkable that while giant scores have doubled in frequency, overall batting averages have risen only slightly, just four percent higher now than in the 1970s and 80s. The secret may lie in scoring rates, which are about fifteen percent faster than 30 years ago (rising from 46 runs per 100 balls to 53). The effect is to make more time available to get to the really big scores.
This may also be why there are more big fourth-innings run chases than ever, with four of the six biggest successful chases occurring in the last decade. When South Africa was set 454 to win in the Johannesburg Test, commentators were seriously discussing their winning chances, something we wouldn’t have heard much about twenty years ago.
Oddly enough, the top two teams, Australia and South Africa, have not really taken part in the scoring feasts. Australia did start this ball rolling with a 6 for 735 against Zimbabwe in 2003, but since then our best score has been 602, and South Africa’s best in five years is 604. Meanwhile, the West Indies, of all teams, have surged past 700 three times while winning a mere five Tests in five years; one win was against Bangladesh.
It could be that the less successful teams feel more motivated to seize records when they come within reach. When Brian Lara made his epic 400 not out against England in 2004, he took a rather cautious 180 balls to go from 300 to 400, breaking the world record but practically guaranteeing a drawn game. Some have argued that West Indies cricket benefited more from setting the record than it would have from winning the match, and a draw was on the cards anyway. Maybe, but is that what the game is all about?
Would You Believe?
The only team that has not conceded a score over 610 in the last decade is Bangladesh! Ironically, this may be more a sign of weakness than strength. A number of opponents have threatened massive totals, only to declare at scores like 2 for 470 (South Africa) or 3 for 610 (India), knowing full well that this would be enough for an innings victory.
(13 March 2009)
One of many puzzles about international cricket is the huge variation in Test match attendances internationally, seemingly unconnected to either national team success or local passion for the game. In England and Australia, attendances have boomed in the 21st Century, but on the subcontinent, stands are so often empty.
In South Africa, things are fair to middling. Crowds for the big showdown Tests against Australia were unspectacular. Still, the total of 73,000 for the Johannesburg Test was a big improvement on previous tours, up from about 42,000 for both the 2005/06 and 2001/02 Tests at the same ground, but a far cry from the 112,000 recorded in South Africa’s glory days in 1966/67.
On the subcontinent, recent Test match crowds belie the public passion for the game. When Test cricket returned briefly to Pakistan last month, fans stayed away even when Younis Khan, not out 306 overnight in Karachi, was in line to break the world record score; only a few hundred showed up to watch. Tests in Pakistan in the 1950s, which included some of the most boring cricket ever played, used to attract daily crowds of up to 50,000. By the late 1990s, when Australia last toured, this had fallen to 5,000 or so, but now even that seems like the good old days.
The picture in Australia couldn’t be more different. Test crowds in Australia have been trending up continuously for almost 20 years, almost doubling, as shown in the chart. The chart shows daily averages up to and including 2008/09; considering that more Tests are played now than in past decades, the situation is very healthy, with more people, in total, attending Tests than ever before, even if the daily averages are not quite what they were in Bradman’s day (37,000 in 1936/37). Total attendances at Tests have surged; they had fallen well behind One-Dayers during the 1980s, but caught up again around 1996, and have been more than 50% ahead in recent years.
Daily Averages based on four-year cycle. Not including off-season matches in Darwin and Cairns.
Attendances for One-Day Internationals, by contrast, have been static since peaking around 1988. However, there is no doubt that a gradual decline for One-Dayers from 1996 to 2004 was arrested by the success of Twenty20 Internationals, which have drawn up to 84,000 people. Dropping ‘neutral’ ODIs not involving Australia has also helped averages.
More than 2.38 million patrons have passed through the Test match turnstiles in Australia in the last four seasons, an all-time high, just ahead of the pre-Packer peak of 2.35 million from 1974/75 to 1977/78.
10 March 2009
Some slow-scoring records can be found in cricket record books. There are mentions of batsmen who took a long time to get off the mark, or who went without scoring for extended periods, and occasionally mentions of bowlers who have bowled many consecutive maiden overs. It’s surprising, then, that one basic question is very difficult to answer. What is the longest period in Tests with no score at all?
Here the record books go quiet. It is a question that, even with the aid of original scorebooks, can be tricky, because traditional scorebooks don’t show the real sequence of events. This has to be painstakingly extracted.
Anyway, I can’t answer the question with certainty, but I have come across a couple of extreme cases recently. One was in the Leeds Test of 1958, between New Zealand and England.
The Test was one of the most one-sided ever; England lost only two wickets and won by an innings, even though half the match was wiped out by rain. The New Zealand batsmen looked paralysed from the start, being bowled out for 67 off 59.1 overs. In reply England cruised to 267/2, with May and Milton (on debut) scoring centuries.
It was in the Kiwis second innings that the slow scoring reached extremes. Bert Sutcliffe made a duck off 51 balls, while at the other end John Reid took 39 balls to get off the mark. In total, 81 balls went by without a run off the bat, including the wicket of Miller. There were, however, four byes at some stage, not marked in the scorebook.
It is an interesting sign of the times that this spell did not really come to the attention of Wisden or the correspondent of The Times.
When Sutcliffe was out, Bill Playle came to the crease and took 23 balls to get off the mark. He scored just three runs of his first 140 balls faced (approximate, since there were a couple more leg byes). Playle finished with 18 off about 175 balls, a worthy addition to the pantheon of extreme inertia.
At one point, there were 26 overs from one end, bowled by Laker, Lock and Loader, that produced just eight runs.
The 81-ball pause is the second longest I have encountered. The longest is a spell of 88-92 balls (22 to 23 four-ball overs, can’t be more precise than that) by Alick Bannerman and Bill Murdoch at the MCG in 1882/83. This has been reported elsewhere as a 14-over scoreless spell, but careful reading of newspaper reports confirms that it was longer than that. Once again, though, there were some byes scored during this spell.
The search continues for a completely scoreless period. A relatively recent extreme case is a 63-ball spell in the India v Sri Lanka Test at Chandigarh in 1990/91. I did some research on this Test following an item in Ask Steven on Cricinfo recently. [This is the most recent Test for which complete extended details are not available; for a statistician that is a bit like having a stone in the shoe, especially as Tendulkar played in that match and we don’t know how many balls he faced.]
Anyway, the 63 scoreless balls came in Sri Lanka’s first innings, when Gurusinha scored 52 not out, out of 82 all out. Sri Lanka was on a score of 54, and three wickets fell; it happened during a remarkable spell of 5 for 2 by SLV Raju.
Incidentally, I have gathered quite a bit of previously obscure info on this Test, but unfortunately Tendulkar’s balls faced is still unknown.
So here is a list of the longest scoreless periods I know of. Any suggested additions by readers would be welcome.
Longest scoreless periods (runs off the bat)
* Contained sundries
Thanks to Peter Lyons for the Leeds 1958 scorebook.
3 March 2009
A couple more columns, as written for the Age.
Powerplays are one of the complications of One-Day cricket that have persisted without being well understood. They are an extension of fielding restrictions that originated in the Packer World Series matches in the 1970s, and by 1983 were in use in ODIs. The restrictions were increased for the first 15 overs in 1991/92. Powerplays, offering some flexibility to the extended restrictions, were introduced in 2005, and now in 2009 one Powerplay is at the discretion of the batting captain.
After all the fanfare, it is surprising to learn that Powerplays don’t seem to have a great effect. There has been little creativity in their use. In 2007 and 2008, Overs 16 to 20 almost always had Powerplays, but Overs 21 to 25 did not. Yet the Powerplay overs produced the same average scoring as overs 21 to 25, about 22 runs. Not much to see here. Overs 10 to 15 were a little more active under Powerplay, but any observable effect was no more than about five runs per innings, on average.
Now we have “Batting” Powerplays, mostly used after the 40 over mark. Is there any positive effect on scoring? I have compared scoring rates during a dozen of these new 2009 Powerplays (first innings only) with a similar set of matches from 2008 (excluding rain-affected, short matches and ‘minnow’ matches). The 2009 and 2008 samples had simiIar overall scoring, close to 250 runs.
The Powerplays at Overs 16 to 20 have been abandoned, resulting in a slight reduction in scoring, by about 3.5 runs in these five overs. This is an effect, but it is not great. Is there an increase in later overs that makes up for this?
*The 2008 matches all had Powerplay in Overs 16-20, but none in Overs 41-46.
The Table compares the scoring rates with and without the new Powerplays later in the innings, alongside results from 2008. There is no noticeable positive effect on scoring. The 2009 Powerplays produced 6.4 runs per over, against 6.8 when not used, (and 7.0 in 2008). Wickets fell at about the same rate during the Powerplays than without them.
These are early days with the new system, but overall, there is little solid evidence here that Powerplays influence matches significantly. Perhaps there will be more effect as teams adapt to the new system, but nothing much has happened yet.
12 February 2009
Fiftieth anniversaries are especially notable when they celebrate rare events. In Adelaide fifty years ago last week, Richie Benaud’s team won back the Ashes, something that since then has happened only twice in this country (1974/75 and 2006/07). It may be worth taking a look at changes in the game since that time.
The 1950s are often seen as a period of decline, but in 1958/59 Test cricket was still a big deal. Four years had passed since any Tests in Australia. In that time, Australian teams had played Tests in every Test nation except the nearest, New Zealand (who had had to make do with a Second XI tour). The MCC tour was a major undertaking for the touring players, who still travelled by ship (air travel was now routine but expensive) and who could count on seven continuous months away from home.
The six-day Tests were the last in Australia to use the traditional timing of five-hour days. Even though over rates were declining, a day’s cricket produced up to 550 balls, about the same as six hours work for today’s generation.
Remuneration for the Australian players could only be described as semi-professional, generally less than £100 ($200) per man per match. Team payments for the whole series would have come to around £5,000, compared to gate takings of £185,000. No Aston Martins for girlfriends in those days. The Tests were the first to be televised in this country, and although broadcast was only local (no live intercity broadcast was possible), the ABC paid the princely sum of £5,000 for the series, to televise the last session of each match.
Crowds averaged about 29,000 per day, as against 38,000 per day for the 2006/07 series.
The series is notorious for slow scoring, although Australia’s series in South Africa a year earlier had been even worse. England, though much fancied, did not exceed 300 for the series, while the Australians, bitter over ‘doctored’ English pitches in 1956, enjoyed grinding their opponents into the dust. Brisbane was the worst. Trevor Bailey made a failed attempt to stonewall the match to a draw with an infamous 68 off 427 balls; his half-century off 350 balls remains the slowest known in all Test cricket, and at one stage he hit only one boundary in 293 balls.
The cause celebre of the day was chucking, which by mid-series was becoming an obsession for the massive English press corps. There would be ramifications: Ian Meckiff, who destroyed England at the MCG, would eventually be drummed out of Test cricket, though not for another five years.
Also five years in the future was an answer to the call for ‘brighter cricket’: the one-day game. Ultimately, this would help transform Test cricket, too, as the shotless stonewallers gradually disappeared from the game.
Slowest-Scoring Ashes teams (five- or six-Test series)
Australia in 1956, 29.9 runs/100 balls.
England in 1958/59, 33.3 runs/ 100 balls.
England in 1978/79, 34.4 runs/100 balls.
Australia in 1884/85, 34.5 runs/100 balls.
England in 1953, 34.7 runs/100 balls.
Fastest: Australia in 2001, 71.3 runs/100 balls.
Would You Believe?
In the entire 1958/59 series, only one ball was hit for six (by Fred Trueman). The 2006/07 series, by contrast, produced 27 sixes.
12 February 2009
A couple of columns, as written for the Age.
(22 January 2009)
Top cricketers profess unconcern at such trifles, and highbrow commentators scoff, but player ratings and rankings remain a favourite topic among fans. The ICC now has its own historical player rankings system, an extension of a well-known system that has gone under several names over the years, and like many things associated with the ICC, it is quite contentious, and secretive in its workings. While the system works well for snapshots of current form, its extension to historical data is dubious.
The rankings have attracted a lot of heat because Matthew Hayden is placed well above Sachin Tendulkar, Hayden even making it into the Top Ten batsmen of all time. The method in this is obscure and difficult to discover, but it seems to be associated with career peaks rather than averages. There seems to be bonus points for scoring well in winning sides, something that happens far more often for Australia than for other nations.
Hayden does have some claims to greatness. He is only the sixth Australian – and our first opener – to complete a career of more than 20 Tests with an average over 50. He held, briefly, the world record innings, a real curiosity considering that huge scores from Hayden were rare.
Ultimately, few sober judges would rank him above Tendulkar. Hayden does have the slightly better average in the 21st century, but Tendulkar also averaged in the mid-60s from 1993-99. The fact that Tendulkar had a more difficult set of opponents should clinch it, even if this is not really Hayden’s fault.
I have done my own calculation of averages adjusted for strength of opposition, location of run-making, and changes in scoring standards. Most current batsmen lose out a bit, partly because of smaller grounds and the new super bats. Hayden suffers more than most, his average falls from 50.7 to 44.8 on adjustment, Tendulkar’s from 54.3 to 51.3.
If this process is carried out for all batsman (something I did for my book The Best of the Best in 2000) an historical ranking can be produced. There is no space here for the detailed explanation found in the book, but here is a list of Top Ten all-time batting rankings, updated to 2009. The ratings are calculated from the adjusted averages, also using a factor for length of career (in years).
A Better Batting Ranking?
Three other batsmen might qualify for this list, but they played too few Tests for a reliable statistical fix. They are George Headley (5.30), Graeme Pollock (5.06), and Golden Age maestro Stanley Jackson (4.76). Matthew Hayden ranks 39th on this list.
Would You Believe?
Hayden played 109 Test innings after his world record 380 against Zimbabwe; this included 15 centuries – a remarkable success rate – but he never got past a score of 153 again.
(29 January 2009)
A Sundries column in November discussed the statistics of dropped catches, which have been gathered in detail for Tests since late 2001. These showed that while Australia has done well when it comes to taking catches, the record is probably not as good as may have been thought; in general, South Africa has dropped fewer catches over the last seven years.
A corollary to “catches win matches” might be “catches dropped by winning sides don’t get noticed”. During the winning years, the standard reaction to Australians such as Shane Warne dropping catches was surprise, as if it’s rare (it’s not). In the final innings of the Sydney Test, Australians dropped four catches, but most was forgiven when they won with ten balls to spare. Still, Matthew Hayden dropping a sitter with just overs to go was memorable, and it points to an interesting observation about Hayden’s later career. Since 2002, there has been a very clear correlation between Hayden’s batting average (going down), and the number of chances he dropped (going up).
The year-by-year fluctuations in Hayden’s fortunes can be seen clearly in both stats (see table). Hayden was superb in both departments in 2002 and 2003, but he had a relatively poor 2004. He averaged only 43 with the bat, while his percentage of dropped catches soared from 18% to 29%. Both stats recovered nicely in 2005, but in 2008 his average fell away to 32 while he dropped seven of the 17 chances (over 40%) of the chances that came his way. Through this period, dropped catches by Australians in general stayed fairly constant at between 21 and 28%.
Fielding in the slips can be tough, and even a good slipper may drop one-quarter of the chances on offer (the current best is Graeme Smith at 13%, the worst Alastair Cook at 42%). At his best, Hayden was well below 20%, significantly better than average, but when he began missing over 40%, it just added to the writing on the wall.
This pattern echoes the decline of Adam Gilchrist, who at his peak from 2002-2004 dropped catches at a rate of only 12%, but by 2008 the rate was rising to over 20%. Gilchrist himself cited his declining form behind the stumps as hastening his retirement.
Would You Believe?
Australia’s One-Day batsmen are suffering a century drought. There were only three tons by Australians in ODIs in 2008 (Adam Gilchrist, Ricky Ponting and Shane Watson), the quietest year since 1995. Ponting’s concern that the big match-winning innings are not happening often enough seems justified. Only one of the last 27 half-centuries by Australians has been converted into a century.
2 February 2009
234 or not 234, that is the question...
I have mentioned a few times in the past (can’t be bothered looking for the links) that there are some oddities about surviving scores of the Bradman/Barnes partnership of 1946 (405 runs, which remains the highest Test partnership in Australia). The story, oft-told, is that when Bradman was out for 234, Barnes threw his wicket away when on the same score a few minutes later, saying that he thought his innings would be remembered longer if his score was the same as the Don’s.
So what are the oddities? There are two surviving scores of the innings that I know of. One is a fantastically detailed re-copy on large-format paper by S.G. Miller, presumably prepared from the scorer Bill Ferguson’s running sheets. Miller’s scores of this series generally look very accurate, but in the case of the Barnes’ innings, the scoring strokes only add up to 233! Time to re-write history? Not so fast. When the bowling analysis is re-scored, the missing run is found, and the score is back to 234.
There is more. I have a copy of Fergie’s “traditional” scorebook that he copied in the evenings of the Tests from his running sheets. Taking a deep breath, I re-scored this one too, and this time Barnes came out with a score of 235! However, there are two ‘anomaly points’ in this score, which when cross-checked against Miller, look very much like errors. When corrected, Barnes recovers his 234 once again.
Fergie’s running sheets would certainly help here, but if they exist, I don’t know where they are. But even the running sheets can have problems, not so much with the scores but with regards to the balls faced. It appears that Ferguson could be a little careless in recording dot-balls, and it is quite common to find in his running sheets the wrong number of balls (usually too few, but not always) in the over. He seems to have ‘corrected’ these anomalies when making the re-copied scores, but whether it was through a superior ability to interpret his own writing, or just a ‘best guess’ in some cases, is unclear. I favour the latter.
Support for this comes from comparing Miller’s version with Ferguson’s. In the Australian innings, there are no fewer than 63 separate overs where the order of the balls differs in the two versions. Even the very first over of the innings is recorded as 20020200 by Miller and 02000202 by Ferguson. It is curious that with all these differences, the final tallies of runs come out the same (apart from the couple of anomalies mentioned earlier).
It is further complicated by a note in the Miller score that, after the tea break on the fourth day, the batsmen returned to the wrong ends, and so “crossed”.
All this assumes some minor importance because Barnes’ innings may hold a record as the slowest double-century ever by an Australian. The Miller re-score seems to settle on a tally of 608 balls to reach 200, which is exactly the same number recorded by Bob Simpson in his 311 at Old Trafford in 1964. The Ferguson score for Barnes (re-scored) comes to 601 balls.
As it happens, both these tallies have been exceeded (marginally) by Glenn Turner in his 259 at Georgetown in 1972, which is reported (though not recorded on the scoresheet) as 611 balls for the first 200. Just how trustworthy this figure is could be debated; I have encountered problems with scores from the West Indies in the past, and the huge Turner/Jarvis partnership, probably the longest in Test history, has reported figures for length that vary in media reports.
There is even one more candidate for slowest 200. When I re-scored Dudley Nourse’s 208 at Trent Bridge in 1951, I got 597 balls for the 200. However, this score lacked markings for byes and leg byes, which were relatively numerous, so uncertainties exist about the strike in some overs. There is also a likely error in the score, since the analysis came to 209, not 208, runs for Nourse and 123, not 122, runs conceded by Bedser. I didn’t work out the probabilities for the possible range for the balls faced – a more complicated calculation than it sounds – but Nourse could conceivably have faced up to 607 balls for his 200. (By the same token, it could well be lower than 597 balls also).
Some statisticians might write this off as a loss due to the uncertainties. I prefer to think that this is useful information; we know these four innings were quite similar statistically, and that they represent the extremes in slow double-centuries, which have largely (and thankfully) vanished from the modern game. If we can’t quite say for sure which is the slowest, we have still learned something.
30 January 2009
A couple of columns, as written for the Age.
(8 January 2009)
Mike Hussey has had a career like no other. Ignored until past his 30th birthday, he was finally selected after Australia’s chastening in the 2005 Ashes. Almost immediately, Hussey showed Bradmanesque tendencies, averaging almost 85 in his first 20 Tests. Lately, though, his career has reflected Australia’s declining fortunes. There is an interesting graphical comparison to be made with the careers of other top line batsmen.
The chart summarises how the careers of 115 recognised batsmen who have played over 100 innings have progressed. These players have an “average average” around 44 runs. The “Low Range” and “High Range” lines show the band of averages where the great majority of the top batsmen can be found (technically, ± 2 standard deviations). Of course, individuals often have careers that fluctuate up or down. Some players improve their averages after 50 innings, some decline: the numbers are about equal, and generally they stay within this band.
“Typical” Averages are calculated for recognised batsmen with more than 100 innings.
The outstanding exception, of course, is Don Bradman, whose career is so far out of the band that it looks statistically impossible. Mike Hussey, at one stage, looked almost in the same class. For most of his career he has been second only to Bradman. Just recently, however, he has returned to the pack, and his average at the 58-innings point of his career (59.3) is now below a few of the greats, like Garry Sobers (61.5), at the same stage. Surprisingly, even Adam Gilchrist had a higher average after 58 innings (60.7).
Statisticians term the experience of Hussey (and Gilchrist later on) “Regression to the Mean”. The Indian and South African bowlers have certainly had something to do with it, but it is probably too much to expect that a batsman who averages just 42 at Sheffield Shield level could maintain a Test average of over 80. Ricky Ponting, who averages over 63 in Shield cricket, is an interesting comparison here. Bradman averaged 110.
Would You Believe?
The sight of a player batting with a broken hand or arm, a la Graeme Smith in Sydney, is very rare, but it happened in consecutive Tests in England in 1984. At Headingley, Malcolm Marshall braved a broken thumb to help Larry Gomes reach a century, even hitting a boundary one-handed. Then at Old Trafford, Paul Terry did the same for Alan Lamb; Terry’s example is perhaps the most extreme since his arm was actually in a sling. Other cases include Salim Malik against the West Indies in 1986, and Colin Cowdrey (who didn’t actually face a ball, but saved the match by coming to the crease for one over) at Lord’s in 1963.
(15 January 2009)
Although Twenty20 cricket is still the new game in town, there are already some interesting patterns emerging in the stats.
There have been 55 T20s between Test-ranked nations, 25 won by the team batting first and 27 by the chasing team, with three ties. This ‘balance’ is superficial, and disguises some deeper inequalities. In day/night games, there is a major advantage to batting first. Nineteen games have gone to the team batting first, and only ten to the chasing team, six of them in the dry air at Johannesburg. At other grounds the winning ratio is an unhealthy fifteen to four.
The Melbourne International is a case in point. Australia basically won that game in the first eight overs, before the lights went on. A pattern also seen rather too often in the 50-over game.
In daytime games, the opposite holds. Teams batting second have won seventeen games, and lost only six. Why is this ratio so lopsided? Maybe teams batting first overreach in the early overs and lose too many wickets. When batting second, you have the luxury of tailoring your strategy to suit the target.
The average first innings score is around 163 with 7.4 wickets lost. Over-by-over scoring is shown in the chart.
*Teams batting first in games between Test-ranked nations.
After a couple of settling overs, batsmen take advantage of fielding restrictions to score at nearly nine runs an over in Overs 3-5. Rates drop when the field spreads in Over 6, gradually rising again to about 10 runs per over in the final two overs. Naturally, the frequency of wickets increases. Early on, there is about a 25% chance of a wicket falling in a given over; this rises gradually to 50% in the 16th over, and jumps to 70% in the final over.
On average, the last five overs produce 15-20% more runs than the first five. In 50-over games the ‘gradient’ is greater: the last 10 overs are typically 50% faster than the first 10.
Even when teams have wickets to burn in the final overs of Twenty20, the average scoring rate rises only to about twelve runs per over at the death. There are many exceptions, of course, but twelve an over seems about the average to expect from all-out attack. It will be interesting to see if these patterns change as the game matures.
Would You Believe?
The only Australian to be dropped from the Test team immediately after scoring a Test double-century was Jason Gillespie. Gillespie is the only Australian to finish his career with a top score (201*) more than ten times his batting average (18.7).
20 January 2009
In discussing over rates on this blog, I have once or twice alluded to a record number of overs bowled in a day of Test cricket, 162 overs bowled on the second day of the 1946 Lord’s Test (England v India). The number, based on a press report, needs correcting. Now that I have re-scored the scorebook for this Test, I have found that this was a slight exaggeration. The correct number is 157.4 overs.
This is still the most cricket in a day that I know of, in terms of balls bowled (942), ahead of two days in the 1947 Trent Bridge Test (153.5 on the second day and 155.2 on the fourth).
The 1946 record was assisted by the fact that, at that time, English first-class cricket often ran to 6.5-hour days on the second and subsequent days. On that day at Lord’s, there was also a change of innings that coincided with the tea break, so both teams shared the bowling without any extra interruptions.
Most Overs in a Day
Some of these figures may be revised with further research.
It is also interesting that five of the eight highest daily over counts occurred in 1946 or 1947. One wonders why these over rates were higher than in most pre-War Tests , which occurred under the same playing conditions but which generally did not exceed the low 140s. Put it down to post-War enthusiasm after seven years in cricketing purgatory.
Five-day Tests limited to six hours per day became standard in England in 1950, after a series of three-day Tests in 1949 (against New Zealand) were all left drawn. The last entry in the table, at Trent Bridge in 1950, actually has the highest over rate, because it was a six-hour day. This is a tribute to the young West Indian spinners Valentine and Ramadhin, who were assisted by slow scoring (263 runs off those 146 overs). Valentine, it is said, could bowl a maiden in 90 seconds. More than 50 maiden overs were bowled that day.
10 January 2009
A couple of new columns, as written for the Age.
First, a slightly longer article here, on the “problems” of winning the toss and batting first.
(26 December 2008)
The alarming lack of depth in Australia’s bowling in the Perth Test produced some interesting statistical aberrations. Australia’s fourteen wickets in the match included eleven from Mitchell Johnson, an unprecedented haul from such a meagre total. Australia’s other bowlers, between them, took 3 wickets for 509, a combined analysis never seen outside the most extreme innings defeats. The ‘best’ of them, Peter Siddle, took 1 for 128 in the match, which is the worst return for any Australian “second-best” bowler, the previous record being set in 1893, with 1 for 94 by Charles Turner in an innings defeat at the Oval. This sort of thing is seen occasionally from the likes of Bangladesh, but will be unfamiliar fare to Australian fans.
Meanwhile, Jason Krejza, with 1 for 102 and 0 for 102 in his second Test, became the first bowler to concede the century in each of his first four bowling innings (Arthur Mailey came close in 1920).
Johnson’s first innings of 8 for 61 is the best return in a losing side by any Australian bowler, previously 8 for 65 by Hugh Trumble at the Oval in 1902, a match lost by only one wicket. Kapil Dev leads the international field here with his 9 for 83 in losing to the West Indies at Ahmedabad in 1983.
The highlight of the match for Australia, statistical or otherwise, was Johnson’s sequence of five wickets for just two runs in 21 balls at the end of the second day. How does this compare to other extreme sequences? This sort of stat is rarely published, so it is worth listing the best known Australian sequences for five wickets:
5/1 (17 balls) Gerry Hazlitt 7/25 Aus v Eng The Oval 1912 (his last 17 balls in Tests)
5/2 (19 balls) Ernie Toshack 5/2 Aus v Ind Brisbane 1947/48
5/2 (21 balls) Mitchell Johnson 8/61 v SAf Perth 2008/09
5/2 (28 balls) Glenn McGrath 6/17 Aus v WI Brisbane 2000/01
5/2 (31 balls) Glenn McGrath 5/53 Aus v Eng Lord’s 2005
Four bowlers from other countries have achieved sequences of five wickets for no runs (excluding Bangladesh Tests): Hugh Tayfield (1953), Fred Trueman (1961), Lance Gibbs (1961/62), and Venkatesh Prasad (1998/99).
The shortest spell for five wickets was 13 balls (approximately) by Monty Noble for Australia at the MCG in 1901/02 (conceding five runs). Bert Ironmonger took five wickets in 17 balls (three runs) against South Africa at the MCG in 1931/32.
Would You Believe?
Australia never lost an Ashes Test when both Glenn McGrath and Shane Warne were in the side, and the team went unbeaten for ten years in all home Tests (from 1996/97) when both champion bowlers were playing. Bowlers win Test matches, they say, but it often takes more than one.
1 January 2009
After loss-free years in 2006 and 2007, the sudden derailment in 2008 of the Australian Test juggernaut has come as quite a shock. Australia has now lost four Test matches in 70 days; the last time this happened was during the Packer upheaval in 1979. Australia’s last series loss at home, against the West Indies in 1992/93, was by a margin of one run. To find a team that defeated Australia at the level of Graeme Smith’s men, we must go back even further, to the West Indies in 1988/89 (3-1 winners).
The nature of the losses is also striking. All of Australia’s previous losses in this century had been by narrow margins, or in dead rubbers, often both. Until this year, Australia had lost only four Tests in eight years where the series was still up for grabs: all of the defeats came narrowly or in extraordinary circumstances. But in 2008 all five of our losses have been in live rubbers, mostly by significant margins.
Fans once grumbled when the Australians tended to lose close matches, or sometimes did not whitewash opponents already beaten. All of a sudden they seem like the good old days.
The age of the team has come under some scrutiny. The link between age and performance is a complex issue. Some teams grow old because the players are so good (and well-paid), and this is why the oldest Australian team ever, apart from those affected by wartime, was the 2006/07 Ashes team that whitewashed England. The team’s average age has dropped by two years since then. Other teams, like the West Indies in the early 1990s, grew old because the supply of good new players dried up. There is no strong relationship between the age of the team and its performance. The oldest Australian teams have been
36 years, in England 1926 (lost)
33.2 years, Ashes 2006/07 (won)
32 years, in England 1905 (lost)
31.6 years, the “Invincibles”, in England 1948 (won)
31.1 years in 1980 (return of Packer’s players) (won)
31.0 years, current (lost)
The youngest teams were aged around 25 years in isolated Tests in 1884, when Test cricket was new, and again just before the Packer rapprochement. The only sustained period where the average stayed below 29 years was from about 1967-1974, before players’ incomes rose, and senior players often chose to retire young, to seek alternative careers.
Would You Believe?
The 180 runs added by JP Duminy and Dayle Steyn was not quite the highest ever partnership for the ninth wicket in Tests – the record is 195. However, it was the longest, at 238 minutes or 382 balls. Asif Iqbal and Intikhab Alam batted for 378 balls for the ninth wicket (190 runs) against England in 1967.
5 January 2009
Various record books include lists of the slowest batsmen to get off the mark, or who spent the longest periods without scoring. It invites the question, who has faced the most balls without scoring in a Test match? A stat rather more difficult to find. I decided to put together a list after I found two new extreme cases while inching my way through old Test scorebooks.
Firstly, there was a classic “Barnacling” by Trevor Bailey, who tried, unsuccessfully, to save the Leeds Test of 1955 v South Africa. Coming in immediately after lunch on the last day, Bailey hustled to one off 23 balls, but then went back into his shell. His second scoring stroke came off his 107th ball faced (give or take, there is one unmarked leg bye somewhere along the way). I believe this is the only innings in Tests to feature just one run off the first 100 balls. Bailey made only one scoring stroke off Trevor Goddard all afternoon, and that came after 62 balls bowled by Goddard. Goddard was the most economical bowler of his day; Bailey vs Goddard was a case of immovable object meeting immovable object.
Bailey was last out for 8 off 135 balls in a single session of play, having gone scoreless at one point for between 78-82 balls.
By coincidence, the next Test I analysed was from Johannesburg in 1938/39. It included an innings of passing strangeness by another king of the block, Bruce Mitchell. When South Africa batted on the second day, Mitchell appeared to have forgotten himself in steaming to 36 in less than half an hour, off 39 balls. Suddenly, when the spinners were brought on, the real Mitchell re-emerged, and it took him 52 balls to go from 36 to 40. Returning from the tea break on 56, Mitchell became becalmed, taking an hour to score again, and he remained on 56 for a total of 85 balls. He was eventually out for 73 off 271 balls. The first half of his innings took 39 balls, the second half 232 balls. Was there ever an innings like it?
Mitchell had not quite surpassed himself, however, since he had once faced 95 balls without scoring (see entry for 20 April 2005). For the moment he holds the top two places in the immobility stakes. The list as it stands follows
Notes: Geoff Allott faced 90 consecutive balls in all without scoring, spanning three innings. The figure for Snedden is quite uncertain. In 1881, George Giffen took about 63 balls to get off the mark in his Test debut.
I will add this list to the “Unusual Records” section. It should be regarded as a work in progress. Other extremes may well be found.
Thanks to Shahzad Khan for some helpful data, and to Peter Lyons for a copy of the Leeds Test of 1955.
21 December 2008
Another Couple of “Sundries” Columns
4 December 2008
After four years in the Australian team, Michael Clarke is a proven Test player. He was named “Player of the Series” against New Zealand. Yet if you go through his resume since his sensational debut in Bangalore in 2004, it looks a little bland. What is missing is a few of those stellar one-off performances that mark the careers of the greats (ironically, his most two eye-catching performance was a spell of 6 for 9 as a part-time spinner). While he won that series award, Clarke has not won a Player of the Match award in his last 36 Tests since 2004. In that time, eighteen Australian players have won at least one, including every one of Clarke’s teammates in the Adelaide Test.[Correction: every one except Nathan Hauritz.]
Clarke has now made nine Test centuries, without surpassing that 151 on debut. More strikingly, the lowest team total to include a Clarke century is 474, and the average is 534, higher than for any other Test batsman (minimum eight centuries). Now Clarke cannot be blamed if his colleagues score so many runs, but even so, there have been 42 centuries by Australians in totals of less than 450 since Clarke’s debut; Simon Katich alone has made four.
It’s not that Clarke fails completely under pressure. His average when Australia loses is a respectable 38, but the pattern is one of moderate scores, often reaching 20 but with a top score of 81.
There is a way of measuring how well a batsman scores runs “against the flow”. Compare his average in those innings where his teammates score poorly, against his average where teammates score well. If the ratio of these averages is high, this means a batsman keeps his head when all about are losing theirs, while a low ratio means he tends to make his runs in big totals. Clarke’s ratio of 0.71 is lowest of current Australians – Ponting is 0.93 and Hayden 0.86 – although some earlier players, including Steve Waugh (0.58) have still lower ratios. (Waugh’s ratio is affected by his early career, when both he and his team tended to fail regularly.)
At the other extreme, there is Andrew Symonds, who made a duck in Australia’s 535 at Adelaide, and whose ratio of 1.96 is almost unparalleled. Symonds has never shone in big innings: his best score when Australia exceeds 500 is only 53 not out, while Michael Clarke has made seven centuries in such totals.
Would You Believe?
Way back in 1926 at the Oval, when Harold Larwood (a 150kph bowler) was bowling, two intrepid Australians, Herbie Collins and Jack Gregory, decided to bat without gloves. Why they did so is not recorded, but the correspondent for The Age noted drily that “Collins was struck on the fingers by Larwood, and obviously felt it for some time.” Indeed.
12 December 2008
The spotlight is once again on our cricketers’ workloads, as a very busy program looms in 2009. Next year sees the Australian team scheduled to play 62 days’ worth of Internationals on tour, which will pass the peak of 58 set in 1999, if all matches take place and last the distance.
Behind all the concern, there seems to be an assumption that players have never worked harder, and that workloads have increased inexorably. Do the numbers support this? Let’s take a look.
The table compares the number of days cricket played annually by the most active players of different generations. The cricket includes all international, first-class, and senior one-day games, including Twenty20, but not minor cricket. The figures are annual averages over the busiest five years of each player’s career.
So while there has been a long-term increase in workload, it was the players of the 1980s (Allan Border) and the 1990s (Mark Waugh) who spent most time on the field. Waugh, in particular, was one of those who played whenever he could, including the county circuit in England. In 1995 he played on over 170 days, possibly a world record.
Mike Hussey’s workload, while considerable, is only a little more than Greg Chappell’s was, and less than 20% greater than Neil Harvey in the 1950s; and remember that Harvey held down a full-time job. Some might argue that the game has changed, and is somehow more intense than in the past. Perhaps, but the other side to this is that the old-timers often put in more than 120 overs in a day, an inconceivable total nowadays.
They also played more minor cricket then. In the 1928/29 Australian season, the young Don Bradman played at least 68 days at all levels, including district cricket. No current player could find the time for that; there would be no time left for buying Aston Martins for girlfriends, and so forth.
A couple of suggestions for our cricketers, from a non-combatant:
- Get through your overs faster, and matches become shorter! More time to put your feet up.
- Go easy on the practice. It seems more players are injured in practice than in the field, Jason Krejza being just the latest.
Would You Believe?
Philip Hughes scored 93 out of 172 and 108 out of 173 for NSW against Tasmania last week, becoming the first player in Australian first-class cricket to score more than 50% of his team’s total in both innings. It did happen once in an Australian tour match, by JT Tyldesley for Lancashire vs Australians in 1899.
For comments, or to contact Z-score (Charles Davis) email
statz334 at iprimus dot
com dot au
(The address is like this to avoid SPAM. Type the address in the usual format)
11 December 2008
Another Nugget for the Vault
Some of the work on this website is from data extracted over the years from my collection of copies of original Test match scorebooks. Collecting this material is an ongoing project, advancing slowly. Occasionally there is a small breakthrough, and I have Peter Lyons (from the Isle of Lewis!) to thank for a recent one. Peter alerted me to a rather rare book from 1926, “The Greatest Test Match”, by John Marchant, which describes, in loquacious detail, the Oval Test of that year when England won back the Ashes.
The original edition had an unusual feature, a reproduction of a scorebook of that Test (not included in a later edition). As it happens, this is the only significant Ashes Test since 1920 for which a scorebook cannot be found in the major archives, so the book filled the gap nicely. I set about re-scoring the match to get balls faced.
Not so easy. For one thing, the reproduced pages are very small, and the writing practically illegible in places. I sought the help of the very detailed newspaper reports of the day (Australian papers much better than English for this purpose), and still it was difficult. The score seems to have a lot of errors, more than any other I have come across (except perhaps Sydney 1903/04). A few dozen changes seemed to be necessary to make it add up, and even then a complete reconciliation of batting and bowling was not achieved.
Curiously, the problems were alluded to by Marchant in the book. He hadn’t analysed it deeply, but he did find anomalies, some of which arise as early as the third over of an innings.
Still, it is a find, and balls faced for this historic match have been calculated, even if the numbers are not precise.
The match was played on a pitch that varied in quality from session to session. The nexus of the match was the Hobbs/Sutcliffe partnership of 172 on the third day which lasted for 86 overs under the most difficult conditions of the match. Hobbs’ batting that morning baffled his opponents and amazed observers. Collins’ captaincy was criticised: apparently Hobbs foxed Collins into thinking that second-string bowler Richardson was doing very well, and Richardson was kept on for a 19-over unsuccessful spell at a critical time. Sutcliffe, for his part, did not score for 56 balls at one point, but finished with 161 off about 470 balls. Hobbs score 100 off about 255 balls.
(Collins, incidentally, faced about 200 balls in his first innings 61 before he hit his first boundary)
Collins was even accused, more than 50 years later, of throwing the match to accommodate some bookmaker contacts of his. There is no hard evidence for this, although Collins was known to be a heavy gambler.
It would seem that the scorebook in the Marchant volume was an amateur effort; at least, one hopes it was not the official record. In any case, the balls faced record for all Ashes Tests since World War I is now virtually complete. I say virtually because the Nottingham Test of the same year is still missing; however, this match lasted only 46 minutes before being rained out, and no wickets fell. Balls faced are also complete for 1903/04, 1905, 1909, 1911/12, and much of 1912; 1907/08 is missing, as is 1912, which was last seen a few years ago.
If any reader knows the whereabouts of any Test match scorebooks, amateur or professional, I would love to hear about it. A case in point: the only known copy of the 1909 scorebook is a photocopy at Cricket NSW; it must have been copied in recent decades, but where is the original?
10 December 2008
More ‘SUNDRIES’ Columns Written for The Melbourne Age November 2008
20 November 2008
Catches win matches, we are told. Certainly they help, as New Zealand showed on the first day at the ‘Gabba, but this is one area where statisticians, lacking data, traditionally fear to tread. Perhaps it is time some of these boundaries were pushed back. I have collated, from ball-by-ball and other accounts, almost 2,000 missed chances in Tests since 2002, allowing some basic questions to be addressed, like just how many chances are dropped? And who drops them the most?
About 6.5 catches and stumpings are missed per Test match, on average. The most recorded was 19 for an India/England Test in Mumbai in 2006, including twelve by India. A few months earlier in Faisalabad, Pakistan set a benchmark by dropping nine in one innings against England. At the other extreme, there appeared to be no missed chances at all in the Johannesburg Test in 2006, South Africa vs Australia. Australia, though, has twice missed six in an innings against India, at Chennai in 2004 and at Adelaide this year.
Overall, about 25% of chances offered to fieldsmen are put down, if you include very difficult chances. The figure varies from team to team, but perhaps not as much as you might expect. Australia, not surprisingly, has a strong record, but not as strong as South Africa, as shown.
Surprisngly, these rankings were just about identical in the first half of the study (2002-04, and the second half (2005-2008).
Of course, some misses have far more dramatic effects than others. Just ask Kumar Sangakkara, who was dropped on 0 against Zimbabwe in 2004 and went on to make 270. Missed chances off recognised batsmen average out at a cost 36.0 runs. Oddly enough, the average score at which catches were dropped is also 36.0 runs. This exact match is coincidental to a degree, but it is linked to the fact that the runs lost in a missed chance tends to be, on average, similar to the batsman’s batting average.
Much more can be said on this topic. I hope to return to it, and discuss how dropped catches have affected individual players, in a later column.
Would You Believe?
Jason Krejza is the first player to be dropped from a Test team after taking 12 or more wickets on debut. Few have been dropped at any stage of a career after such a haul. The only Australian precedent was a sensational, and strange, case. Clarrie Grimmett took 13 wickets at Durban in 1936 (and an Australian record of 44 wickets in the series), and yet never played another Test. Apparently he was dropped at the behest of Don Bradman, captaining Australia for the first time in the 1936/37 series.
27 November 2008
Byes play a small part in most matches, never smaller than in the ’Gabba Test, where there were no byes at all. Kudos to wicketkeepers Brad Haddin and Brendon McCullum for this unusual feat, very rare when all 40 wickets fall (only two precedents, Johannesburg 1970 and Adelaide 1991/92). Strangely, only six weeks ago Haddin set a new Australian record by conceding 39 byes at Bangalore. Pitch conditions must certainly be important: in the previous Bangalore Test in 2007, Dinesh Karthik conceded a world record 47 byes.
Byes can tell us a few surprising things about wicketkeeping and how the game works. Another reason Brisbane was so different to Bangalore is the bowling mix: there was far more spin bowling at Bangalore, and byes are more prevalent from spinners. Medium and fast-medium bowlers attract the fewest byes, about 0.5 per 100 balls. True fast bowlers attract more (0.85); the most bye-prone modern bowlers are Shoaib Akhtar (1.8) and Lasith Malinga (2.2), which will not surprise anyone who has seen them bowl. However, as a group, spin bowling results in more byes, with the rate for wrist spin 0.99, and finger spin 1.42.
In fact, bye rates from spinners are much higher again if you consider only those balls that pass through to the keeper. How many balls go through to keepers? Simple question, difficult answer, which seems to be “about 20%”, based on a dozen recent Tests. Again there is a big difference between pace (28%) and spin (8-10%); the rate at Brisbane was 28%, at Bangalore 18%. The bye rates run counter to this, showing that keepers have a much harder time with spin than pace.
*Note: based only on those balls that go through to the keeper.
It has been said that spin bowling is the true test of a keeper. From these figures, there can be little doubt that spin represents a greater challenge to keepers than pace.
Andrew Symonds’ hit for eight runs off the bat at the ’Gabba has only two precedents, both in Melbourne. Patsy Hendren hit an eight off Percy Hornibrook in 1928/29 (not in Brisbane, as reported elsewhere), and John Wright did the same off Len Pascoe in 1980/81.
Would You Believe?
The highest batting average recorded by any Australian after five Test matches is 99.0, not by any of the usual suspects, but by Mitchell Johnson! Johnson was out only once. Johnson’s average peaked, briefly, at 112 just before his dismissal in his next Test.
For comments, or to contact Z-score (Charles Davis) email
statz334 at iprimus dot
com dot au
(The address is like this to avoid SPAM. Type the address in the usual format)
23 November 2008
‘SUNDRIES’ Columns Written for The Melbourne Age November 2008
Bowling in India must sometimes feel like a Himalayan challenge for touring cricketers. Visiting fast-medium bowlers have averaged 41.0 in India over the last decade, higher than in any other country. There is nothing new about this, but Australia’s current bowling ranks are really under the hammer. Australia have yet to bowl India out twice in this series: we last went through a long series without a win in 1984 in the West Indies.
Is it the conditions or the opposition? More the latter than the former, perhaps. VVS Laxman, in particular seems to specialise in harassing Australian bowlers. He has just tallied 323 runs between dismissals (200*, 59*, 64), the most by any batsman against Australia since Len Hutton made 364 in one dig in 1938 (Shivnarine Chanderpaul notched 313 earlier this year). Laxman works best in partnership: four of the five largest partnerships against Australia in the last decade have involved Laxman (376 and 303 with Dravid, 353 with Tendulkar, and 278 with Gambhir).
Selectors must have been relieved that Jason Krejza had a reasonable first day with the ball at Nagpur. Never before has a bowler debuted for Australia after such an inauspicious lead-up (0 for 199 in his last match). There are few parallels: one fast bowler named Keith Slater debuted for Australia in 1959 after taking 0 for 118 in a Shield game; it was his only Test appearance.
One other is Gavin Robertson, a spinner who struggled for a place in the NSW side, but was chosen in similar circumstances to Krejza for the 1997/98 India tour (although he did have ODI experience). Robertson did well enough first up, with four wickets and a half-century, but his career at the top was short – 4 Tests.
WOULD YOU BELIEVE?
The “Devil’s Number” for Australian batsmen, 87, remains a popular, and persistent, myth. It was inevitably mentioned when Ricky Ponting was dismissed for 87 in Delhi, but in fact, Ponting is only the second Australian in more than 200 Tests to fall for this score.
Ken Mackay claimed to have originated the superstition, as did Keith Miller, though neither were ever out for 87 in first-class cricket. It is easily debunked, as the table shows, but it could be said that there is an inverse effect, which could be psychological.
As it happens, 87 is the luckiest score in the 80s for Australian batsmen. Australians are out for 87 only half as often as batsmen from other countries.
Jason Krejza’s freakish bowling figures at Nagpur went beyond mere “debut” records. In taking 12 for 358, Krejza conceded far more runs than any bowler in history in a time-limited Test match (previously 290 by ‘Roley’ Jenkins, Lord’s 1950). One bowler, Oscar ‘Tommy’ Scott of the West Indies, conceded more (374), but that was in a 9-day “Timeless” match in 1930. Stranger still, Krejza bowled just 75 overs, a total exceeded over 400 times in Tests: Muralitharan alone has bowled more than this on eighteen occasions.
Krejza conceded over half of India’s runs (off the bat) in the match. This is a rare occurrence, particularly when all twenty wickets fall. The last instance for Australia was at Adelaide in 1894 (George Giffen), which, coincidentally, saw Albert Trott on debut take 8 for 43 in the second innings. Let’s hope that Krejza enjoys a longer career than Trott or Bob Massie (16 for 137 on debut), who between them played fewer than 10 times for Australia.
Over rates are back in the news. Those unfamiliar with the ‘good old days’ might be amazed at how much Test over rates have fallen. Today’s 90-over daily quota almost always requires extended playing hours, but there was a time when 130 or more overs in a day was commonplace. The record for a Test match day is scarcely believable: 162 overs by England and India at Lord’s in 1946. The chart shows how rates have fallen, particularly in the 1950s and 1970s. Who knows how low they might have gone had authorities not mandated minimums in the late 1980s. At that time the West Indies, with their all-pace attack, had pioneered super-low rates below 80 balls per hour.
Nowadays most teams have fallen into this habit. The biggest historical change seems to be with spinners. In the 1950s spinners wheeled through their overs 30 or 40% faster than pace-dominated attacks. Most of that difference has disappeared; no one seems to be able to lift over rates more than 10 or 15% regardless of the types of bowlers used.
Would You Believe?
The Nagpur Test was the first time in almost 20 years that three top order Australian batsmen had been run out in the same match (since Adelaide 1988/89 vs West Indies). Based on (my own) calculations, the cost of each incident is roughly the average of each affected batsman, so by running out Hayden (avge 52.5), Ponting (58) and Hussey (66), India may well have saved 176 runs. Yes, that’s actually less than India’s winning margin. Of course, Australia did mitigate their losses by running out Tendulkar.
29 October 2008
The Ultimate Bradman
When we come to a score of 452 not out, as Don Bradman made for New South Wales v Queensland at the Sydney Cricket Ground in the first few days of the 1930s, it doesn’t take a lot of analysis to know what an extraordinary innings it was. After all, it held the record for highest first-class innings for the best part of 30 years. Even so, closer analysis can still be rewarding, and reveal important, previously unknown features of the innings; for example, by “re-scoring” the surviving original scorebook, we can calculate for the first time how many balls the Don faced.
One of the surprising features is that Bradman made his runs in the second innings, after being out for three in the first. Bradman’s score remains, by a margin of more than 100 runs, the highest in the second innings of a match, ahead of WG Grace’s 344 in 1876; there are no other second innings among the Top 40 first-class scores.
Bradman was fortunate that Sheffield Shield matches were scheduled for 4 and a half days in 1929/30; the 5th day was abolished after that season, and Bradman’s innings, under later circumstances, would have been curtailed by an earlier declaration. Bradman, 205 not out overnight, also may have appreciated the rest day on the 5th of January. Still, scoring so many in a second innings required extremely fast, sustained scoring and intense concentration, which has been praised by many commentators.
Bradman by all accounts was rarely troubled by the Queensland bowlers (though Thurlow bowled well, and there were near chances on 264 and 345), but the secret of the gigantic score seems to lie in discipline and consistency. Bradman seemed to quickly find a comfort zone and he stuck with it. The table below gives the times and balls faced for the innings milestones.
Going on the batting times, Bradman appeared to be accelerating during his last 100 runs, but the balls faced show no such acceleration. The reason is that his last 200 was scored with the lower order, and Bradman farmed the strike heavily in the final stages. Earlier on, when Bradman added 272 with Kippax, and 156 with McCabe, the strike was shared fairly evenly, but he took 70% of the strike in the 8th-wicket stand of 92 off 105 balls with Davidson. This is a very difficult percentage to sustain when scoring rapidly, because strike-farming basically means sacrificing scoring opportunities.
The consistency of scoring is emphasised by his session-by-session scoring. Bradman netted 85 (in 92 minutes), 120 (101 mins), 105 (106 mins) and 142 (116 mins) runs in each session, coming close to four centuries in a session in a row. It is intriguing that Bradman did not hit a six in all that time.
Bradman reached his 300 off 310 balls faced. While extreme, this is not record-breaking. It is not even the fastest triple century for Bradman, that being his 369 v Tasmania in 1935/36. We don’t know the exact balls faced in that innings, but it was probably similar to the 221 balls faced by Charlie Macartney in reaching 300 against Notts in 1921, the fastest triple for which BF figures are known (Dennis Compton once reached 300 in 25 minutes faster than Macartney, but he faced 261 balls).
The 452 has parallels with Brian Lara’s 501 off 474 balls in 1994, when Lara, after reaching 300 off 278 balls and 400 off 367, eased off just a little in the latter stages to maintain concentration.
NSW v Queensland, 3, 4, 6, 7 January 1930.
NSW First Innings: All out 235
Queensland First Innings: All out 227
NSW Second Innings (4th and 6th January 1930)
Bradman hit 13 threes, 46 twos, and 125 singles.
Queensland Second Innings: All out 84
NSW won by 685 runs.
3 October 2008
Added to the Longer Articles section:
Some Statistical Snippets from Don Bradman’s Career Some information supplied to Ken Piesse for use in his book Our Don Bradman. A recommended publication!
2 October 2008
Ten Years of Player vs Player Records
The increasing scope of available Test match data in recent years creates new opportunities for cricket statistics. One area where we have more information than in previous years is specific player vs player data. Just how well does a batsman do against a specific bowler? This sort of question has long been of interest to commentators, but in the past this could only be answered in general terms by statisticians.
I have extracted a few player vs player extremes from Cricinfo’s data (either in specific player v player form or as ball-by-ball text commentary), supplemented by other sources (hat tip to Andrew Samson) so that the record can be extended back to the 1998/99 Ashes series. The data covers over 450 Tests, and is about 99.5% complete, with a majority of the gaps being in some Zimbabwe Tests. This forms a new class of cricket records.
Most of the records below are based on a qualification minimum, with minimum of either 200 balls bowled, or five dismissals, in encounters between specific bowlers and batsmen. A ‘recognised’ batsman is one with an average batting position of less than 7.1.
Some Player vs Player Records 1998 – 2008.
Most balls bowled by one bowler to one batsman:
736 N Boje to DPMD Jayawardene (410 runs). Boje bowled 221 balls to Jayawardene in one innings during Jayawardene’s 374 at Colombo in 2006, a single-innings record.
Most runs by one batsman off one bowler:
441 BC Lara off SCG MacGill (4 dismissals, batting average 110.3).
Most runs by one batsman off one bowler (single innings):
130 in 161 balls by BC Lara (400*) off GJ Batty, St John’s 2004. (Note: Garry Sobers scored 133 of his 365* off Khan Mohammad in 1957/58)
Most balls bowled by one bowler to one batsman without dismissing him:
556 Harbhajan Singh to S Chanderpaul (196 runs), in eight matches.
Most runs scored by one batsman off one bowler without dismissal:
223 by RS Dravid off SCG MacGill (354 balls in five matches)
Highest batting average:
238.0 by JH Kallis off DL Vettori (238 runs for once out).
Lowest batting average (recognised batsman):
1.00 by Matthew Bell (NZ) off J Srinath. This is a remarkable case. Srinath dismissed Bell (an opening batsman) five times in Tests and only conceded five runs in 103 balls bowled.
11 by SK Warne bowling to AG Prince (164 runs, batting average 14.5). Greater numbers can be found going further back than 1998. For example, Mike Atherton fell to Glenn McGrath 19 times in Tests, including pre-1998 matches: a full analysis is not yet available. Atherton’s vulnerability to McGrath is well-known; perhaps less well-known is his failure against Chaminda Vaas, against whom he averaged just 6.6 with five dismissals.
Highest batting strike rate
104.3 Runs /100 balls RT Ponting off AR Caddick (batting average 72). This does not include earlier encounters of these two players in 1997, which would take the strike rate down to 91.6. Chris Cairns had a strike rate of 103 against Brett Lee, although his batting average was only 13.4. Shahid Afridi has scored 202 runs at a strike rate of 93.1 against Anil Kumble.
Highest Bowling Strike Rate (recognised batsman)
Makhaya Ntini dismissed Nathan Astle six times in just 92 balls bowled to him, conceding 37 runs. Glenn McGrath took Sanath Jayasuriya’s wicket five times in just 76 balls, twice dismissing him with the first ball of an innings, but these figures don’t include the Adelaide Test of 1995/96, where Jayasuriya got the better of McGrath.
Ajit Agarkar faced only two balls from Mark Waugh, and was dismissed both times. Agarkar was also out to his first two balls from Brett Lee, and has been out three times in the five balls faced he has faced from Lee.
A final curiosity: If it needed any confirmation, take a look at Brian Lara’s head-to-head batting averages against some leading spin bowling since 1998:
vs M Muralitharan 124.0
vs SK Warne 74.0
vs SCG MacGill 110.3
vs Danish Kaneria 86.7
vs N Boje 212.0
Kudos to Anil Kumble, who seems to have a much stronger record against Lara than other spinners(though data from their encounters in 1994-96 is not available).
Over time, it will be possible to extend this data to earlier Tests. However, chances are that earlier data will be more incomplete, as there are quite a few Test matches even in the 1990s for which complete scorebooks have not yet been located. If any readers, especially in India, Pakistan, and the West Indies, know of the existence of detailed Test match scores (not necessarily official ones) from the 1990s or earlier, please get in touch with me through this blog.
31 August 2008
A new article, describing the ‘uncertainty’ of Don Bradman’s batting average of 99.94, and the possibility that four ‘missing’ runs have been found to round out his average to 100, has been included in the longer articles section. This was published in three Australian newspapers and made it to the front page of the Sydney Morning Herald.
A list of most balls bowled before first wicket in Tests has been included in the “Unusual Records” section. It has been updated with the discovery that RGCE Wijesuriya bowled 561 balls before his first wicket, which places him second on the all-time list behind AG Kripal Singh (651 balls). Thanks to Shahzad Khan for the new info. See also the entry for 14 August 2006.
14 August 2008
The Hot 100 list, the fastest- and slowest- scoring batsmen in Tests, has been updated. Somewhat belatedly – it has been almost a year since the last update – but it is interesting to note that there have not been a lot of changes. Interesting, at least, in that while a batsman’s batting average can fluctuate up and down quite a lot, his scoring rate tends to remain at a characteristic level. Batting failures will affect the batting average much more than the average scoring rate.
Adam Gilchrist’s career has ended with him in second place on the all-time list. He is not far ahead of Kapil Dev. There remains a little uncertainty about Kapil’s figure because we have balls faced for only 93% of his career (96% including minutes batted), but it is doubtful if Kapil’s missing innings were unusually fast or slow.
Since the last update a little more information has come to light on Maurice Tate, the hard-hitting allrounder/tailender of the 1920s and 30s. Some innings that were previously only estimates now have exact balls faced. (We now have balls faced for 81% of Tate’s runs, and minutes batted for 98%.) It turns out that in Tate’s case the estimates were spot on and his estimated scoring rate (74.5 runs/100 balls) has not changed.
In the Most Tenacious section, new information on Don Bradman and Herbert Sutcliffe is available. It has not changed the leaderboard, but the two greats are now very close in terms of Balls Between Dismissals. The exact figures are now 163.4 balls for Bradman and 163.1 balls for Sutcliffe. In reality, the underlying data is not precise enough to distinguish between them.
It is unlikely that the list of slowest batsmen will change anytime soon. In the modern game, no batsmen could make a career for himself batting at the speed of the stonewallers of the 1800s or 1950s. Besides, with today’s superbats and shrunken grounds, it is much easier to score quickly.
13 August 2008
The New Unbowlable
Although he is not the most dynamic batsman going round, Shivnarine Chanderpaul has been carving himself a most unusual statistical place in recent years. His career has been studded with extraordinary spells where he becomes near-impossible to dismiss. Since overcoming a foot problem and undergoing surgery in 2000, he has made a habit of stringing unbeaten innings together, so much so that he has batted for more than 1000 minutes between dismissals on four separate occasions, twice in the past year. His last four Test innings, in the series against Australia, spanned 1115 minutes before until he at last fell lbw to Stuart Clark. For comparison, consider that only five other batsmen have ever gone unbeaten for over 1000 minutes, none of them more than once.
Longest Batting Between Dismissals
These are all relatively recent events, thanks partly to the fact that over rates are much slower than in olden days. But if we turn to Balls Faced, Chanderpaul still leads. He is the only batsman to ever face 1000 balls without getting out.
Most Balls Faced Between Dismissals
*Balls Faced Figure for Hanif is an estimate only.
A striking feature is that Chanderpaul’s highest score in these purple patches is only 136 not out. It is a sign of the general weakness of his team’s batting, as well as his defensive nature, that he so often get left unbeaten without making huge scores. It could certainly be argued that the West Indies are losing out on potential runs because of this; perhaps he should bat higher in the order.
Curiously, the 362 runs he scored in that 1000-ball sequence is not even in the Top 20 for most runs between dismissals, which is led by Tendulkar (497) and Sobers (490) in the tables above. Chanderpaul is in the Top 20 thanks to his 371 runs in 2004/05, but well down the list.
One reason for his success is that Chanderpaul has become the nearest thing to an unbowlable batsman seen in Test cricket. This has developed in recent years as his technique has changed. Even though he has been out bowled in 11% of his dismissals, not an especially low figure, many of these dismissals came earlier in his career. From 2004 to 2007, Chanderpaul played a sequence of 57 innings without being out bowled. He scored 2629 runs, faced 5693 balls, and batted over 138 hours without anyone hitting his wicket! Javed Miandad (2055) is the only other batsmen to score over 2000 runs without being bowled, although Kumar Sangakkara is now right in the hunt for this record, having scored 1983 runs since he last heard the death rattle. Adam Parore of New Zealand batted 77 times before he was out bowled for the first time in a Test match, scoring 1937 runs.
One other curiosity: Chanderpaul is known for his caution, yet has made one of the fastest Test centuries of all time. His normal scoring rate is just over 43 runs/100 balls: among current batsmen, only Rahul Dravid has scored more runs at a slower rate. Yet Chanderpaul has to his credit the fastest century ever scored against Australia, and the fourth-fastest in all Tests, 69 balls at Georgetown (his home ground) in 2003. Has anyone ever batted quite so “out of character”? Perhaps not. The next fastest century by Chanderpaul, 140 balls (in the same series) is less than half as fast as his best, and the average of his other centuries is 212 balls. No other major batsmen has a fastest century so unlike all his others.
22 July 2008
COMPARING THE RECORD TEAM SCORES
By Charles Davis
When comparing the biggest team scores in Tests, the results can be a bit messy. This is because cricket often does not allow teams to carry their innings to completion, and big innings are often truncated by declaration or lack of time. We know for sure that the highest innings in a Test match is Sri Lanka’s 952 for 6 in 1997, but an interesting side question would ask if this is also the most ‘extraordinary’ score in Tests. For example, we know that the West Indies once made a score of 790 for 3. Where might such an innings have gone if it had continued? Can we compare it to Sri Lanka’s record?
While we can never know for sure, it is possible to make a statistical estimate. The approach is to look at the way that innings naturally progress over a wide range of scores. Of course, there is plenty of variation between innings (part of cricket’s appeal), but there are statistical patterns. A team that is, say, five wickets down, will on average add a certain number of runs if the innings is played to completion.
This average number of runs added depends also on the starting point. A team on, say, 50 for 5, can be expected to add fewer runs than a team on 500 for 5, before being bowled out. But there is a surprising result to be found here. Contrary to expectation, the number of runs at the starting point is not very important, with only a limited effect on the future progress of the innings. This is shown in the following table, calculated from the outcomes of all relevant Test innings, which gives the average number of runs added by teams with five wickets down, at different starting points.
What we see here is that above a certain level, in this case about 300 runs, there is very little change in the potential scoring of a team. This is surprising, but it probably comes down to the fact that a batsman coming in at a score of 600 for 5 is likely to bat in a riskier manner, or with less intensity, than one who comes in at 300 for 5. This would appear to balance out any advantage from tired bowling or benign conditions. This pattern is also seen at 6, 7, 8 or 9 wickets down.
It should be stressed that these runs added will often be theoretical in practice. For example, the Projected All Out score for teams that reach 600 for 5 is 710, but in practice most such innings will not reach 700, often because of declarations. What the projected All Out score gives us is an estimate of where the innings was headed if the limits of time and tactics had been removed – its trajectory if you will.
With modern computer power, the result of this process is an “Innings Projector” that can give a projected estimate for any score. (In practice, it only works for innings with two or more wickets down.) Estimates for extreme innings must remain provisional because of the rarity of the situations, but the fact that trends are so stable, as illustrated by the first table, adds confidence to the results.
So what are the most extreme projected scores? Here is a list of the results:
So Sri Lanka retains the #1 position under this calculation. However, the West Indies 790 for 3 moves up to second place, while England’s 849 all out in the Timeless Test of 1930 moves down to seventh.
Another aspect to these scores is that the distribution of the scores around these projections can be calculated, which means that the probability of a specific score can also be calculated. For example, the probability of a score of 790/3 actually exceeding the 1028 assigned to Sri Lanka’s record is about 24%.
One other possible calculation here is a re-appraisal of the most one-sided innings victories in Tests. Using the projected score, the margin of victory can be re-calculated and compared more evenly. The most one-sided Tests in this analysis are:
(Please, no comments that the ‘highest’ does not mean the ‘greatest’. No one is claiming that it does. We are just looking at extremes here.)
[Technical note: the trajectory at large scores must be calculated with care, because teams that continue with great success from a high starting point rarely complete their innings. This must be allowed for in the calculation. The way to do this is through an iterative process, where big innings that are declared closed are themselves calculated through to completion, firstly for innings that are nine wickets down, then eight, seven, and so forth, and these results are then fed back into the calculation for end points starting from fewer wickets down.
For example, take a score of 500 for 3. This has occurred 37 times in Test matches. The projected score in this case is 705 all out. However, only three of the 37 teams have actually reached or exceeded a score of 705, while nine have been bowled out for less than 700. The reason that the projected score is above 700 is that many teams continue to do well but declare before reaching 700. Careful iterative analysis of these declared scores produces the average estimate of 205 runs added, or 705 all out for a projected score.]
17 June 2008
Gibbs Update (see also 30 May 2007 and 7 July 2007)
Back from holidays, where some more detail was found regards the unique 8 for 6 spell against India by Lance Gibbs at Bridgetown in 1962. The Barbados Advocate gives some useful detail not available elsewhere, including the times of each wicket, and in combination with other sources, an over by over sequence of sorts is possible. One complication is that some sources, including Wisden, give the length of the spell as 15.3 overs with 14 maidens. The Advocate, and one or two other sources, say 16.3 overs with 14 maidens, and on close inspection, this appears to be correct.
The following reconstruction distills various sources, but is based on the Advocate account, starting with Gibbs being brought on for Worrell just before 2 o’clock. Prior to this spell, Gibbs had bowled 37 overs without success. The overs in Gibbs’ spell are numbered. The overs at the other end, mostly bowled by Solomon, are given in italics. Note that there was no new ball taken in an innings that lasted 185.3 overs. When Gibbs was brought back on, there had been no wicket for close to 100 overs, and the partnership between Sardesai and Manjrekar had produced just 98 runs. The score, 149/2 off 142 overs at lunch, was now 158/2 off 153 overs.
1. Wicket maiden (Sardesai, 60 in 392 mins) 2:00pm
2. One run to Umrigar.
3. Double wicket maiden (Manjrekar 51 in 259 minutes, Pataudi 0, 2nd ball). 2:11
Solomon, maiden. India now 159 off 159 overs.
4. To 10. Seven overs with six maidens, and one over for 5 runs.
Solomon bowled at the other end, conceding 10 runs. For 25 mins, Borde faced Gibbs and Umrigar faced Solomon.
11. Wicket maiden (Umrigar, 10 with 1x4) 2:45, Borde 6 not out.
Solomon, conceding two runs from Borde plus a bye
13. Double Wicket maiden (Borde, 8 with 1x4, Engineer 0, 2nd ball) 3:00
Solomon, boundary by Durani, plus two singles (Durani, Nadkarni)
14. Wicket maiden (Durani, 5 with 1x4) 3:07. Gibbs spell at this point is 14-12-6-7
15. Maiden. At this point India had scored 183 runs in 182 overs.
Solomon’s last over, with a single to Nadkarni, single to Desai and 2 leg byes. Solomon’s spell: 21 overs for 24 runs, no wickets.
Sobers on, maiden
17. Desai (1) out off third ball. 3:22pm
All out for 187 off 185.3 overs in 474 minutes.
This sequence confirms that Gibbs took his last 5 wickets without conceding a run. Oddly enough, none of the sources mention this, though they mention his 3 for 1 spell earlier. It may have escaped notice because it was stretched over a period of 12 overs during which runs were scored at the other end.
12 May 2008
The website/blog Cricket Buzz has produced a list of “50 Best Cricket Blogs”, and included this site on the list. So there is a compliment. Thanks, guys. Actually, I didn’t know there were 50 cricket blogs, best or otherwise, but with millions of people blogging, I suppose it is not surprising.
So Near but So Far (Written for Cricinfo)
When Virender Sehwag strode to the Chennai wicket on the fourth day of the recent Test against South Africa, he already had 309 runs to his name. There would have been a great many fans wondering how far he could go: could he top Brian Lara’s 400?
Statistics, however, say that the fans were very likely to be disappointed (as they were). The truth is, that while 309 and 400 sound like reasonably similar scores, they are not. In fact, it is harder for a batsman to add another 100 runs if he has already made 300, than it is at almost any other score.
There have now been 22 Test triple centuries, enough for some statistics. Only one of those triples has gone on to produce the magic 400, while 17 have been dismissed. Only one out of 18: that is only a 5.6% conversion rate. (The other four innings finished not out between 300 and 399; it is better not to include them in this calculation.) It is interesting to compare this to the conversion rates at other scores:
*0-99 data involves only recognised batsmen (#1-#6 in batting order). “Number of successes” refers to the number of innings that have passed through the specified range without dismissal, e.g., for 0-99 it refers to the number of centuries.
While interesting, this data is not very robust for the 300-399 range. If the next batsman to make a triple century happens to go on to 400, the conversion rate will almost double (to a rate similar to the 300-400 conversion rate in first-class cricket of 11%). However, the difficulty that batsmen encounter above 300 can also be seen when we look more closely, at 20-run increments
Note the similarity of the pattern at the 200 mark and the 300 mark. As batsmen approach 200, their conversion rate rises, only to fall suddenly after reaching the milestone; the same thing happens at 300. A dismissal between 280 and 299 is a rare thing.
It is also striking that a batsman’s ability to add runs once he has reached 300 (67% and 58% for 300-319 and 320-339) is, in effect, no better than for those who have just reached 100 (62% and 65%).
Further perspective can be gained by looking at the one batsman who did make it to 400, Brian Lara at St John’s in 2004. In that innings, Lara played with caution and great focus after reaching 300, taking 178 balls to go from 300 to 400 (56 runs per 100 balls). This is probably the slowest progression from 300 to 400 in first-class cricket: in doing this under very benign conditions when quick runs were called for, Lara also sacrificed any chance his team had of winning the match.
Few triple-centurions take this approach. The surprisingly high failure rate after reaching 300, when scoring should be easiest, is probably a combination of mental exhaustion and the need of teams for quick, high-risk runs in those circumstances. The typical scoring rate for triple centurions in their first 300 runs is about 63 runs per 100 balls, but for runs beyond the 300 mark (apart from Lara), the rate is over 80 runs per 100 balls, in time-limited Tests.
5 May 2008
Sreeram has pointed out a bowling spell that belongs in the list of longest-ever spells: Garry Sobers at the MCG in 1960/61. He bowled 41 consecutive eight-ball overs over the second and third days, coming on at 124/0 and finally relieved at 335/9. This is the third-longest known bowling spell after Hirwani in 1990 and Athol Rowan in 1948/49.
George Giffen’s 300-ball spell in 1894/95 is no longer the longest-known in Australia. There could still be others even longer than Sobers’, although this seems unlikely.
I have added these records to the Unusual Records section.
Something Completely Different
In the Longer Articles Section, I have posted a paper I wrote and presented for the “Olympic Legacies” conference at Oxford University in March 2008. The title is fairly self-explanatory:
Hope you find it interesting.
1 May 2008
here is some basic information on Australian batting partners in Tests. The pairs who scored most in partnership are as follows
As you can see, a number of opening pairs at the top, although Australia has had relatively few long-term successful opening pairs. Interesting that Hayden/Ponting and Langer/Ponting have more century stands than Hayden/Langer.
Here is a list of highest partnership averages among batting pairs. Minimum 1000 runs in partnership.
Bradman and Barnes added 992 runs together at 124.0
- Darren Lehmann and Ricky Ponting only had four partnerships together. They were 315, 141, 27, and 121.
- Bradman and Ponsford only batted together 10 times in Tests, even though their careers overlapped by six years. Bradman and Woodfull, by contrast, batted together 22 times over exactly the same period.
- Andrew Symonds and Brett Lee have only batted together once, with a partnership of 114.
- Mike Hussey and Jason Gillespie put on 320 the only time they batted together in Tests. This is a record for Australians, but there is one other (non-Australian) batting pair who scored more runs in their only partnership together. That would be a good trivia question. (I'll keep you guessing.)
23 April 2008
I recently obtained copies of press clippings for the 5th Test in South Africa’s tour of Australia in 1931/32, a match remarkable for the tourists’ collapse, out for 36 and 45, Australia (all out 153) winning by and innings and 72 runs. Oddly enough, there seemed to be a consensus that the pitch on the first day was not too bad, even though 20 wickets fell (SA 36, Aus 153, SA 5/1). Australia’s poor scoring was put down to the absence of Ponsford (injured) and Bradman (in the team but did not bat). South Africa’s batting was derisory.
Since this was the final Test of a 5-0 whitewash, the reports were not as detailed as they might have been, but with four Melbourne dailies reporting independently, there is enough info to reconstruct the two South African innings over by over, and identify every scoring shot. On the first morning, the South African batsmen abandoned all discipline after the veteran champion Taylor was out for a duck and Christy for four off 30 balls. By lunch they were 33/9. The last six wickets fell in less than six overs for 17 runs, the 49-year-old Bert Ironmonger taking 5/3 with his last 17 balls.
After Australia fell for 153, the South Africans started just as badly with Christy out first ball. The next day was rained out, and it was after this that the notorious Melbourne “sticky” wicket emerged. Nothing much happened until the 17th over, when nightwatchman Bell was out for 6 off 50 balls. When Curnow was out for 16 off 70 balls, the floodgates opened again. Once again, Ironmonger captured five wickets for three runs, this time in the space of about 21 balls. It took a number of dropped catches for South Africa to exceed their first innings; Taylor was dropped twice before he scored and was still out for two off five balls.
Ironmonger finished with 11/32 for the match. But perhaps the strangest thing was that Clarrie Grimmett, who had taken no fewer than fourteen wicket in the previous Test, did not even get a bowl.
The Ironmonger spells have been entered in to the Unusual Records section.
15 April 2008
Test cricket has changed in many ways over the decades; to the statistician, one of the most striking is the speed at which it is played. By that, I don’t mean the speed of bowling or scoring, though these are important, but simply the sheer amount of cricket that gets played in any given hour or day. Today, it is rare to see even 90 overs bowled in six hours, but in days gone by, 140 or even 150 overs in a day was commonplace. On the second day of the Lord’s Test of 1946, India and England wheeled through no fewer than 161 six-ball overs.
For spectators, it must have been rich entertainment when batsmen were on the attack. One of the most productive innings came at Lord’s in 1924, when England put South Africa’s bowlers to the sword, scoring 503 runs on the second day, for just two wickets, in less than five and a half hours. England scored 200 runs before lunch and another 223 between lunch and tea. While 200 or more in one session is rare enough, keeping it up for two sessions in a row appears to be unique.
While doing a bit of general research, I came across more details of this Test in the original scorebook, thankfully preserved by the archivists at Lord’s. The 200 before lunch was greatly assisted by the bowlers getting through 57 overs (!) in an extended session. Jack Hobbs made his highest Test score, 211 off 300 balls. Hobbs was not given to collecting giant scores, and The Times commented that towards the end he batted as though he “seemed to think someone else might as well have a turn at batting”. One of those others was Frank Woolley, one of the most aggressive batsmen of his generation, who scored 134* off 123 balls, fine hitting in any era.
A tally of 223 runs in one session raises the question of records. Where does it stand? No one seems to have assembled a list before, so here is my attempt. This is one record that favours old-time Tests, but there are a few modern entries (all involving “minnows”). Pre-War Tests in England predominate, mainly because sessions and days in other countries in the days of high over rates tended to be shorter than in England (a pre-War Test day in England was often 6.5 hours, but in Australia only 5 hours). I have examined only those Tests that had specified tea breaks; tea breaks were not always taken in Tests before 1910.
In fact, there are quite a few extreme cases from sessions that were extended beyond the normal two hours, for various reasons. These have been put into a separate list. Note that all of the two-hour cases were the lunch-tea session, whereas all of the extended-session cases are in the opening or closing sessions.
Most Runs in a Two-Hour (maximum) Session
236 (43 overs) Aus v SA, Lunch-Tea, Joburg 1921 (119 off 85 balls by Jack Gregory)
233 (41 overs) Eng v Pak, Lunch-Tea, Nottingham 1954 (Denis Compton 173)
231 (45 overs) Eng v NZ, Lunch-tea 3rd day, Leeds 1949 (both teams batted)
223 (43 overs) Eng v SA, Lunch-Tea, Lord’s 1924
220 (47 overs) Eng v NZ, Lunch-Tea, Wellington 1933 (Wally Hammond 151)
216 (28 overs) Pak v Ind, lunch-tea, Lahore 2006 (two teams)
208 (32 overs, 100 minutes) Aus v SA, lunch-tea, Sydney 1910/11
207 (29 overs) Aus v Zimbabwe Lunch-Tea Perth 2003 (both Matt Hayden and Adam Gilchrist scored centuries in the session)
Most Runs in a Longer Session
249 (33 overs) SA v Zim, post-tea 1st day, Cape Town 2005
244 (58 overs, 165 minutes), Eng v Aus, post-tea, Oval 1921
239 (45 overs, 140 minutes), Eng v NZ, pre-lunch, Lord’s 1937 (two teams)
236 (35 overs, 160 minutes) Eng v Aus, post-tea, Edgbaston 2001 (two teams)
223 (35 overs, 150 minutes) Eng v Ban, post-tea, Chester-le-Street 2005 (Marcus Trescothick 127)
221 (150 minutes) Eng v SA, pre-Lunch, Oval 1935 (Les Ames 123)
219 (35 overs, 150 minutes) NZ v Zimbabwe, post-Tea, Harare 2005 (Daniel Vettori 127)
~210 (150 minutes) Eng v India, pre-Lunch, Oval 1936
208 (47 overs, 154 minutes) Aus v SA, post-tea, Melbourne 1910/11 (Victor Trumper 133)
200 (57 overs, 150 minutes) Eng v SA, pre-Lunch, Lord’s 1924
Readers are invited to submit others that I may have overlooked.
Speaking of remarkable sessions, I was asked if India, all out for 76 against South Africa at Ahmedabad, had become the first team to be bowled out before lunch on the first day of a Test match. Not quite, as it happens, but there appears to be only one precedent, and that was 112 years ago. In the Lord’s Test of 1896, Australia was bowled out for 53 in 85 minutes, allowing England to reach 37/0 before lunch.
India also became the first team to be bowled for less than 100 after scoring over 600 in their previous innings in the same series. Sri Lanka once went from 713/3 to all out 97 in consecutive innings in 2004, but as their opponents were Zimbabwe and Australia respectively, it’s not quite the same thing.
7 April 2008
I don’t study first-class cricket a lot, but some time ago (14 July 2006) I took a look at the fastest first-class triple centuries, and showed that Charlie Macartney’s 345 (300 off 221 balls) was the fastest known in balls faced. Even though Denis Compton had once scored a faster 300 in time batted, he faced 261 balls reaching the milestone.
A couple of contenders for the crown of fastest triple were identified, whose balls faced were unknown. I have now obtained (from the library at Lord’s) a score for one of these, Frank Woolley’s 305 against Tasmania in 1912, and re-scored the innings to get balls faced.
Woolley scored 305* off 235 balls, reaching 300 off 230 balls. This is the second-fastest known, but Macartney’s record stands. A comparison of several of these super innings follows, including a triple by Graeme Smith in 2005, and Don Bradman’s 452*, which I have also re-scored recently. (The Bradman innings is included for reference; this is the first published reference to the balls faced for this historic innings, and it will be described in more detail in a later post)
One other innings, Bradman’s 369 vs Tasmania in 1935/36, remains a contender for fastest triple. No scorebook is known to exist; if Bradman received half the strike, he would have faced 215 to 230 balls reaching 300. However, a characteristic of Bradman in full flight was that he tended to get more than half the strike; this is evident in the 452*, his 334 at Leeds in 1930, and his 244 at the Oval in 1934, which was his fastest Test double century. It seems likely then that Bradman would have taken more than half the strike against Tasmania, and thus received more than 230 balls for his 300. Macartney rules.
For entries May 07 to March 08 click here
For entries November 06 to March 07 click here
For entries April 06 to October 06 click here
For entries January 06 to March 06 click here
For entries Nov 04 to June 05 click here