October 5, 2013

How do openers affect No. 3 batsmen?

The success or failure of openers shifts the average of No. 3 batsmen by about ten runs
12

Rahul Dravid: had to follow failing openers in almost a fourth of his innings at No. 3
Rahul Dravid: had to follow failing openers in almost a fourth of his innings at No. 3 © AFP

No. 3: the best batsman in a Test team bats there. A list of most prolific No. 3 batsmen in Test history will confirm this. Donald Bradman, Wally Hammond, Viv Richards, Brian Lara, Ricky Ponting, Jacques Kallis, Kumar Sangakkara, Hashim Amla, Rohan Kanhai, Neil Harvey, Richie Richardson, Rahul Dravid. The list goes on. Both VVS Laxman and Dravid wanted to bat there. Dravid won that argument and VVS had to concede the point after a lean spell in the early 2000s.

Ian Chappell has argued that No. 3 batsmen set up Test matches. At No. 3, the contest between bat and ball has not yet been defined for the innings. No. 3 is a peculiar position - sandwiched between the middle order and the openers - where a batsman has to be prepared to walk in to face the second ball of the innings.

The history of Test cricket reveals that the success or failure of openers affect the average of the No. 3 batsman by about ten runs. The table below provides some figures. I looked at a number of conditions, and finally chose six to present here.

  All Test matches Non-draws*
No.3's Avg Instance No. 3's Avg Instance
Openers score less than 10 33.60 768 30.60 567
All Test innings 39.90 7450 35.90 4850
When at least one opener reaches 35 37.20 4108 40.10 2432
When 1st wicket falls with score under 10 35.90 2171 33.00 2432
Openers score less than 20 33.50 1819 30.30 1351
At least one openers scores 50 or more 45.30 2973 41.00 977
* Excluding all Tests against Bangladesh and Zimbabwe. However, Bangladesh and Zimbabwe batsmen have been included.

The No. 3 has averaged 40 runs per wicket in Test cricket. If we consider Tests which ended in a win, a loss, or a tie, and Test innings in which batsmen faced Bangladesh or Zimbabwe, the average runs made by a No. 3 Test batsman is 36. When both openers are dismissed for single figures, this drops down to 31. In fact, even if both openers are dismissed for less than 20, it drops down to 30 to 31. If at least one opener reaches 35 in non-draws, the No. 3 batsman averages 40. If at least one opener reaches 50, the No. 3 reaches 41.

In the population of No. 3 batting innings, there is some correlation between an opener's score and the No. 3 batsman's score. Part of this has to do with the conditions. Conditions that assist new-ball bowling generally make it less likely for the openers to survive.

Occasionally, the match situation also matters. If a team loses an opener early in its response to a large total, it allows the fielding captain to set attacking fields for much longer periods, as he has more runs to play with.

What of individual No. 3 batsmen? In what follows, I'll work with two definitions. Openers will be said to have failed when both are dismissed for a score under 20. They will be said to have succeeded when both score 20 or more. The former condition means two things: 1) The No. 3 batsman has to come in early in the innings. 2) The second wicket (i.e. the other opener) falls soon after the No. 3 batsman comes to the wicket. The latter condition means that the No. 3 can come into bat when the ball is no longer "new". It is worth pointing out that the median number of runs contributed by openers in innings where both cross 20 is 109.

No. 3 batsmen when both openers are dismissed under 20
Player Team Aggregate Dismissals Innings Average
R Dravid India 1861 48 51 38.80
KC Sangakkara Sri Lanka 2068 43 46 48.10
RR Sarwan West Indies 1213 28 28 43.30
RB Richardson West Indies 1006 24 26 41.90
Habibul Bashar Bangladesh 639 25 26 25.60
DC Boon Australia 804 21 23 38.30
SP Fleming New Zealand 607 22 23 27.60
JH Kallis South Africa 779 19 22 41.00
Younis Khan Pakistan 803 20 21 40.20
BC Lara West Indies 961 20 20 48.10
IM Chappell Australia 849 17 19 49.90
RT Ponting Australia 665 18 19 36.90
AH Jones New Zealand 574 19 19 30.20
Saeed Ahmed Pakistan 506 19 19 26.60
MW Goodwin Zimbabwe 489 18 19 27.20
C Hill Australia 552 17 18 32.50
Zaheer Abbas Pakistan 537 15 18 35.80
AL Wadekar India 406 18 18 22.60
RB Kanhai West Indies 912 16 17 57.00
DI Gower England 509 17 17 29.90
DB Vengsarkar India 283 15 16 18.90
BE Congdon New Zealand 793 13 15 61.00
HM Amla South Africa 767 13 15 59.00
JL Langer Australia 756 15 15 50.40
RN Harvey Australia 579 14 15 41.40

What stands out immediately is that some No. 3 batsmen have had to follow failing openers far more frequently than others. In the table above, I have only shown those No. 3 batsmen who have had to follow failing openers on 15 or more occasions. Dravid has had to follow failing openers in almost a fourth of his innings at No. 3. Sangakkara has a similar record. In contrast, Ponting has batted at No. 3 when both openers failed only 19 times in his 196 innings in that position. Kallis has had to do it 22 times in 78 innings. Two celebrated No. 3s are not in this list. Richards (59 innings at No. 3) and Greg Chappell (38 innings at No. 3) watched openers fail ten times each. Zaheer Abbas batted at No. 3 60 times. When the openers failed, his averaged dropped to 36.

What happens to No. 3 batsmen in innings where both openers cross 20?

No. 3 batsmen when both openers cross 20
Player Team Aggregate Dismissals Innings Average
RT Ponting Australia 3755 75 85 50.10
R Dravid India 3887 64 69 60.70
KC Sangakkara Sri Lanka 2623 35 40 74.90
HM Amla South Africa 2251 34 38 66.20
DC Boon Australia 1480 28 34 52.90
RB Richardson West Indies 1571 28 31 56.10
IM Chappell Australia 1390 26 28 53.50
MA Butcher England 1009 26 27 38.80
IVA Richards West Indies 1682 25 26 67.30
RB Kanhai West Indies 1352 26 26 52.00
JH Kallis South Africa 1199 20 24 60.00
RN Harvey Australia 1492 22 24 67.80
Younis Khan Pakistan 1279 22 24 58.10
M Amarnath India 1198 19 22 63.10
DB Vengsarkar India 892 19 21 46.90
RR Sarwan West Indies 756 21 21 36.00
BC Lara West Indies 1247 18 20 69.30
IJL Trott England 848 16 20 53.00
AH Jones New Zealand 537 16 19 33.60
BE Congdon New Zealand 750 15 19 50.00
N Hussain England 674 16 19 42.10
DG Bradman Australia 1577 15 18 105.10
ER Dexter England 1240 17 18 72.90
WR Hammond England 1198 13 18 92.20
Zaheer Abbas Pakistan 848 16 17 53.00
C Hill Australia 770 15 16 51.30
JA Rudolph South Africa 522 12 15 43.50
MP Vaughan England 678 15 15 45.20
WJ Edrich England 875 14 15 62.50

As you can see, the top No. 3 batsmen do far better when the openers succeed than when they do not. The phenomenal Bradman scored 5078 runs at No. 3 in only 56 innings. He averaged 105 when both openers succeeded. In seven innings when both openers failed, Bradman made 214 runs at 35. As you will see in the next table, when the first wicket fell for less than 20, and Bradman batted at No. 3, he averaged only 86!

The early wicket doesn't seem to matter very much. I looked at No. 3 batsmen's record in innings where the first wicket fell for a score of under 20. As you can see from the size of the table, this happens quite frequently. There are 65 batsmen who have walked in to bat after the first wicket fell for less than 20. The best batsmen have records under these conditions that are not significantly worse than their career records.

The interesting thing in this table from my point of view is that the batsmen who are normally considered to be technically excellent do worse than batsmen who are considered to be risk takers and stroke players. For example, Ponting, Sangakkara and Richards have done substantially better than Dravid, Kallis and Trott.

No. 3 batsmen when first wicket falls for less than 20
Player Team Aggregate Dismissals Innings Average
R Dravid India 4576 97 106 47.20
KC Sangakkara Sri Lanka 4807 83 87 57.90
RT Ponting Australia 3293 62 67 53.10
DC Boon Australia 1868 42 47 44.50
Habibul Bashar Bangladesh 1692 46 47 36.80
RB Richardson West Indies 2066 43 46 48.00
RR Sarwan West Indies 2024 42 43 48.20
HM Amla South Africa 2012 37 41 56.80
IM Chappell Australia 2014 38 41 55.40
DB Vengsarkar India 1155 37 40 31.20
SP Fleming New Zealand 1357 38 40 35.70
JH Kallis South Africa 1503 34 39 44.20
RB Kanhai West Indies 2178 38 39 57.30
Younis Khan Pakistan 1691 34 35 49.70
MA Butcher England 1172 31 34 37.80
IJL Trott England 1117 32 33 34.90
ALWadekar India 1007 32 32 31.50
JL Langer Australia 1356 31 32 43.70
BC Lara West Indies 1429 30 31 47.60
RN Harvey Australia 1179 31 31 38.00
SaeedAhmed Pakistan 1029 31 31 33.20
DI Gower England 1077 30 30 35.90
AHJones New Zealand 1394 28 29 49.80
AP Gurusinha Sri Lanka 1049 25 29 42.00
Azhar Ali Pakistan 1037 27 29 38.40
C Hill Australia 914 27 28 33.90
ADR Campbell Zimbabwe 925 27 27 34.30
M Amarnath India 1192 27 27 44.10
BE Congdonl New Zealand 1006 24 26 41.90
N Hussain England 1229 22 26 55.90
IVA Richards West Indies 1388 24 25 57.80
Zaheer Abbas Pakistan 649 23 25 28.20

Generally, the figures suggest that the ability of opening batsmen to keep the new ball out is crucial. Given that Dravid and Ponting have played in the same era, it is difficult to argue that their records were built on significantly different wickets. Much of the difference between their records must be explained by the quality of openers they batted after. This matters when one compares their respective career records.

But I want to leave you with a thought about India's highly successful Test line-up in first 12 years of the 21st century. Rahul Dravid averaged 70 as an opener (made three centuries) in Tests. There is much truth, going by the figures, to suggest that the best, most attacking batsman in a team should bat at No. 3. What might have been had India's team management flipped the batting order to have Dravid open the batting, and Virender Sehwag bat at No. 3? Dravid's stated first ambition in each Test innings was to stay at the wicket for 30 overs. He was dismissed once every 123 balls in Tests. He was, by his own admission, a reluctant opening batsman. But if No. 3 batsmen flourish when the openers keep the bowling at bay at the start, and if stroke players thrive at No. 3, then did India make a colossal tactical error by having Sehwag open and Dravid bat at 3?

Kartikeya Date writes at A Cricketing View and tweets here

Comments have now been closed for this article

  • on October 5, 2013, 23:44 GMT

    hi, did you take into account CONFOUNDERS like if the pitch is difficult or opposition bowling is good, openers and number 3 are both likely to score fewer runs; the lower averages are not due to openers failing but rather difficult conditions??

  • gandabhai on October 5, 2013, 19:33 GMT

    It is also true that it was far more easier for a Ricky Ponting to score runs whilst playing in a TOP team than for a Brian Lara playing in a weak team .As it goes SRT also played in a weak team for most of his carear .

  • straightdrive on October 5, 2013, 18:34 GMT

    There's a pretty obvious explanation for your findings that you haven't brought up. When openers fail it could be because they're facing a better than normal attack and that would obviously be true for the #3 as well thereby explaining his lower than normal average.

    If there is a psychological factor that prevents #3's from excelling when openers fail this data doesn't demonstrate that. I respect the effort it took to do the research but you went looking for data to support your hypothesis and not the other way around.

  • on October 5, 2013, 7:26 GMT

    I wonder if Brendon McCullum and Martin Guptill are reading this, and thinking of poor Kane Williamson...

  • kartikeya on October 5, 2013, 6:00 GMT

    A word of caution about the stats, especially when comparing current players and players like Ponting, Richards, Dravid etc. At various points in their careers, these players had numbers comparable to those produced by Amla and Sangakkara at the moment, when these two are near or at their peak.

  • on October 5, 2013, 5:42 GMT

    Gather all the stats you want,analyze how ever you want.those numbers will show that Kumar Sangakkara & Hashim Amla are the best No.3 batsmen world has ever seen after Bradman.Admit it !

  • vaidyar on October 5, 2013, 5:35 GMT

    Ideally, VVS, if he had any consistency under normal non-crisis circumstances would've batted no. 3 and RD at no. 5 followed by Ganguly. But then SG was captain and wouldn't bat at 6; even VSS, a better batsman was pushed down to accommodate him. RD at 6 is practically useless as there is no point playing sheet anchor for the tail. Either way, for India's lineup, RD at no. 3 worked because there was Sehwag with a musical chair on the other side, and then SRT, SG and VVS to follow after RD. He was the foil. I personally don't subscribe to the theory of this kind of batsman has to play at only this position. It varies from team to team based on composition.

  • CricFan24 on October 5, 2013, 5:23 GMT

    Kartikeya - you state "But it also suggests, that irrespective of the quality of the pitch, he was exposed to challenges that Ponting wasn't exposed to."

    The point is that it is actually better to come in early especially on good pitches. A No.3/4 is in any case expected to be technically proficient enough to tackle the new ball.

    Dravid may have been exposed to different challenges but this does not mean that the challenges were tougher than those faced by Ponting. Coming in at 20/1 may place more "pressure" on a batsman but this is perhaps preferable to coming in a a bigger score where the batsman is more "complacent". Also as has been shown very few big scores are scored by batsmen who come in with a big score already on the board.

  • kartikeya on October 5, 2013, 4:52 GMT

    Yes, there are clearly a number of reasons. Which is why the aggregate figures are significant in my view. All things considered, in all Test Cricket, if openers are dismissed for less than 20, it hurts the number 3 batsman's average by about 10 runs, as compared to when the openers reach at least 20.

    If we compare Dravid and Ponting, for Ponting, the openers failed 19 times in 196 innings. For Dravid, they failed 51 times in 219 innings - more than twice as frequently. Yet, it is difficult to say that Dravid and Ponting played on dramatically different wickets - they played in the same era, on many of the same grounds, and against many of the same bowlers.

    Its reasonable to assume that Dravid followed weaker openers than Ponting. The numbers suggest this. But it also suggests, that irrespective of the quality of the pitch, he was exposed to challenges that Ponting wasn't exposed to. This is important when their records (which are similar) are to be compared.

  • gsingh7 on October 5, 2013, 4:52 GMT

    in case first 2 batsman find it easy to score it means either pitch is flat or bowling is below par.naturally number 3 will score more in such conditions. hardly a valid compilation of stats .

  • on October 5, 2013, 23:44 GMT

    hi, did you take into account CONFOUNDERS like if the pitch is difficult or opposition bowling is good, openers and number 3 are both likely to score fewer runs; the lower averages are not due to openers failing but rather difficult conditions??

  • gandabhai on October 5, 2013, 19:33 GMT

    It is also true that it was far more easier for a Ricky Ponting to score runs whilst playing in a TOP team than for a Brian Lara playing in a weak team .As it goes SRT also played in a weak team for most of his carear .

  • straightdrive on October 5, 2013, 18:34 GMT

    There's a pretty obvious explanation for your findings that you haven't brought up. When openers fail it could be because they're facing a better than normal attack and that would obviously be true for the #3 as well thereby explaining his lower than normal average.

    If there is a psychological factor that prevents #3's from excelling when openers fail this data doesn't demonstrate that. I respect the effort it took to do the research but you went looking for data to support your hypothesis and not the other way around.

  • on October 5, 2013, 7:26 GMT

    I wonder if Brendon McCullum and Martin Guptill are reading this, and thinking of poor Kane Williamson...

  • kartikeya on October 5, 2013, 6:00 GMT

    A word of caution about the stats, especially when comparing current players and players like Ponting, Richards, Dravid etc. At various points in their careers, these players had numbers comparable to those produced by Amla and Sangakkara at the moment, when these two are near or at their peak.

  • on October 5, 2013, 5:42 GMT

    Gather all the stats you want,analyze how ever you want.those numbers will show that Kumar Sangakkara & Hashim Amla are the best No.3 batsmen world has ever seen after Bradman.Admit it !

  • vaidyar on October 5, 2013, 5:35 GMT

    Ideally, VVS, if he had any consistency under normal non-crisis circumstances would've batted no. 3 and RD at no. 5 followed by Ganguly. But then SG was captain and wouldn't bat at 6; even VSS, a better batsman was pushed down to accommodate him. RD at 6 is practically useless as there is no point playing sheet anchor for the tail. Either way, for India's lineup, RD at no. 3 worked because there was Sehwag with a musical chair on the other side, and then SRT, SG and VVS to follow after RD. He was the foil. I personally don't subscribe to the theory of this kind of batsman has to play at only this position. It varies from team to team based on composition.

  • CricFan24 on October 5, 2013, 5:23 GMT

    Kartikeya - you state "But it also suggests, that irrespective of the quality of the pitch, he was exposed to challenges that Ponting wasn't exposed to."

    The point is that it is actually better to come in early especially on good pitches. A No.3/4 is in any case expected to be technically proficient enough to tackle the new ball.

    Dravid may have been exposed to different challenges but this does not mean that the challenges were tougher than those faced by Ponting. Coming in at 20/1 may place more "pressure" on a batsman but this is perhaps preferable to coming in a a bigger score where the batsman is more "complacent". Also as has been shown very few big scores are scored by batsmen who come in with a big score already on the board.

  • kartikeya on October 5, 2013, 4:52 GMT

    Yes, there are clearly a number of reasons. Which is why the aggregate figures are significant in my view. All things considered, in all Test Cricket, if openers are dismissed for less than 20, it hurts the number 3 batsman's average by about 10 runs, as compared to when the openers reach at least 20.

    If we compare Dravid and Ponting, for Ponting, the openers failed 19 times in 196 innings. For Dravid, they failed 51 times in 219 innings - more than twice as frequently. Yet, it is difficult to say that Dravid and Ponting played on dramatically different wickets - they played in the same era, on many of the same grounds, and against many of the same bowlers.

    Its reasonable to assume that Dravid followed weaker openers than Ponting. The numbers suggest this. But it also suggests, that irrespective of the quality of the pitch, he was exposed to challenges that Ponting wasn't exposed to. This is important when their records (which are similar) are to be compared.

  • gsingh7 on October 5, 2013, 4:52 GMT

    in case first 2 batsman find it easy to score it means either pitch is flat or bowling is below par.naturally number 3 will score more in such conditions. hardly a valid compilation of stats .

  • CricFan24 on October 5, 2013, 4:42 GMT

    So for eg. 1) Poor openers fail - and Batsmen No.3/4 etc play their usual games they will put up par scores in general. 2)Good bowling/poor pitch - Batsmen 3/4 etc would struggle in any case.If good openers can handle the bowling/pitch then so can good No.3/4s.

    It is also worth noting that most batsmen mentioned including Richards/Lara etc avg. much higher at No.3 . In fact, on a good pitch( or perhaps in general) it is perhaps ideal for the best batsman to either open or come in early to make full use of his innings. i.e a 20/2 is almost always preferable to a 100+/2 - conventional wisdom notwithstanding.

    Also note that most truly big scores ( triples and 250+) scores have been scored by openers or No.3s. Otherwise lower order batsmen have come in early due to early wickets. For eg. Bradman's triple lower down the order was because he "reverse ordered" the batting line-up.

  • CricFan24 on October 5, 2013, 4:18 GMT

    If the openers fail there are certain reasons: 1)Poor openers - this will hardly matter to the batsmen to come. 2)Good pitch for bowlers - in which case the No.3 basman would have struggled in any case. 3)Good opening bowling - again same as point no.2.

    What we require to see is "why" openers failed . This is far more revealing than just having failed without going into the reasons for the same.

  • No featured comments at the moment.

  • CricFan24 on October 5, 2013, 4:18 GMT

    If the openers fail there are certain reasons: 1)Poor openers - this will hardly matter to the batsmen to come. 2)Good pitch for bowlers - in which case the No.3 basman would have struggled in any case. 3)Good opening bowling - again same as point no.2.

    What we require to see is "why" openers failed . This is far more revealing than just having failed without going into the reasons for the same.

  • CricFan24 on October 5, 2013, 4:42 GMT

    So for eg. 1) Poor openers fail - and Batsmen No.3/4 etc play their usual games they will put up par scores in general. 2)Good bowling/poor pitch - Batsmen 3/4 etc would struggle in any case.If good openers can handle the bowling/pitch then so can good No.3/4s.

    It is also worth noting that most batsmen mentioned including Richards/Lara etc avg. much higher at No.3 . In fact, on a good pitch( or perhaps in general) it is perhaps ideal for the best batsman to either open or come in early to make full use of his innings. i.e a 20/2 is almost always preferable to a 100+/2 - conventional wisdom notwithstanding.

    Also note that most truly big scores ( triples and 250+) scores have been scored by openers or No.3s. Otherwise lower order batsmen have come in early due to early wickets. For eg. Bradman's triple lower down the order was because he "reverse ordered" the batting line-up.

  • gsingh7 on October 5, 2013, 4:52 GMT

    in case first 2 batsman find it easy to score it means either pitch is flat or bowling is below par.naturally number 3 will score more in such conditions. hardly a valid compilation of stats .

  • kartikeya on October 5, 2013, 4:52 GMT

    Yes, there are clearly a number of reasons. Which is why the aggregate figures are significant in my view. All things considered, in all Test Cricket, if openers are dismissed for less than 20, it hurts the number 3 batsman's average by about 10 runs, as compared to when the openers reach at least 20.

    If we compare Dravid and Ponting, for Ponting, the openers failed 19 times in 196 innings. For Dravid, they failed 51 times in 219 innings - more than twice as frequently. Yet, it is difficult to say that Dravid and Ponting played on dramatically different wickets - they played in the same era, on many of the same grounds, and against many of the same bowlers.

    Its reasonable to assume that Dravid followed weaker openers than Ponting. The numbers suggest this. But it also suggests, that irrespective of the quality of the pitch, he was exposed to challenges that Ponting wasn't exposed to. This is important when their records (which are similar) are to be compared.

  • CricFan24 on October 5, 2013, 5:23 GMT

    Kartikeya - you state "But it also suggests, that irrespective of the quality of the pitch, he was exposed to challenges that Ponting wasn't exposed to."

    The point is that it is actually better to come in early especially on good pitches. A No.3/4 is in any case expected to be technically proficient enough to tackle the new ball.

    Dravid may have been exposed to different challenges but this does not mean that the challenges were tougher than those faced by Ponting. Coming in at 20/1 may place more "pressure" on a batsman but this is perhaps preferable to coming in a a bigger score where the batsman is more "complacent". Also as has been shown very few big scores are scored by batsmen who come in with a big score already on the board.

  • vaidyar on October 5, 2013, 5:35 GMT

    Ideally, VVS, if he had any consistency under normal non-crisis circumstances would've batted no. 3 and RD at no. 5 followed by Ganguly. But then SG was captain and wouldn't bat at 6; even VSS, a better batsman was pushed down to accommodate him. RD at 6 is practically useless as there is no point playing sheet anchor for the tail. Either way, for India's lineup, RD at no. 3 worked because there was Sehwag with a musical chair on the other side, and then SRT, SG and VVS to follow after RD. He was the foil. I personally don't subscribe to the theory of this kind of batsman has to play at only this position. It varies from team to team based on composition.

  • on October 5, 2013, 5:42 GMT

    Gather all the stats you want,analyze how ever you want.those numbers will show that Kumar Sangakkara & Hashim Amla are the best No.3 batsmen world has ever seen after Bradman.Admit it !

  • kartikeya on October 5, 2013, 6:00 GMT

    A word of caution about the stats, especially when comparing current players and players like Ponting, Richards, Dravid etc. At various points in their careers, these players had numbers comparable to those produced by Amla and Sangakkara at the moment, when these two are near or at their peak.

  • on October 5, 2013, 7:26 GMT

    I wonder if Brendon McCullum and Martin Guptill are reading this, and thinking of poor Kane Williamson...

  • straightdrive on October 5, 2013, 18:34 GMT

    There's a pretty obvious explanation for your findings that you haven't brought up. When openers fail it could be because they're facing a better than normal attack and that would obviously be true for the #3 as well thereby explaining his lower than normal average.

    If there is a psychological factor that prevents #3's from excelling when openers fail this data doesn't demonstrate that. I respect the effort it took to do the research but you went looking for data to support your hypothesis and not the other way around.