Features Icon 1 FEATURES

Zero to Sixty – Important Runs in Test Cricket

Ricky Ponting is the most clutch Test batsman of the modern era, and it isn’t close. In fact, Ponting’s level of clutch batting compared to Don Bradman is the equivalent of Punter’s Test batting average being 92.07! While there is no doubt that others such as Brian Lara, Virender Sehwag and Adam Gilchrist were all highly impactful Test batsmen, none comes close to Ponting when it comes to scoring runs when it mattered most. Here we’re going to discover who were the most clutch Test batsmen of all time.

What is all this ‘clutch’ stuff, anyway
For those traditionalists among you who may not be familiar with the term, ‘clutch’ is coined largely by American sports journalists to describe that trait of ‘intestinal fortitude’ which is not possessed by the average player, and which allegedly separates the merely good from the truly great – those who can step up their game and deliver the goods when the match is on the line. How this relates to cricket generally, and Test cricket in particular, is further discussed below.

Cognitive what?
The issue of sportsmen raising their game when the chips are down has fascinated people in general, and sports fans in particular, for years. As far back as 1908 Yerkes and Dodson posited that too much (and, interestingly, too little) stress causes a decline in performance. More recently, Hanin (1997) claimed that each player had a Zone of Optimal Performance based on his or her individual preferred level of anxiety. Humara (1999) noted that for sports where there are pauses as part of regular play this can lead to increased visualization of upcoming success with a related positive impact on performance, which is of course the case with cricket.

Most of the discussion on clutch performance has largely been related to sports for which a specific function has limited outcomes but instantaneous results. This is not the case for Test cricket – so does clutch performance, or an equivalent of it, even exist in the international long game? The fact that we are dealing with top-class athletes (at least in the modern era), the best in the world at their chosen sport, who in pressure situations have faced down the most dangerous bowlers to have ever played, means that it is likely that the effects noted by the eminent people above are indeed at work in Test cricket. They must be.

What we don’t know is, at what point would the state of a Test match begin to have an impact on what Humara called the ‘cognitive anxiety’ of a particular batsman? The key to this study, as with all of my previous studies for that matter, is that we can’t accept any findings unless they’re based on objective evidence – as the late Christopher Hitchens noted, “What can be asserted without evidence can be dismissed without evidence”. We can’t just say, “Hey, innings A was great but innings B was not as good because of X, Y and Z”; it has to be an objective assessment applied consistently and without any subjective bias. Otherwise, it’s simply opinion.

Other similar studies have always keyed on showing whether or not clutch performance is a repeatable skill, which invariably means looking at smaller performance samples. Regarding Test cricket, while it is possible that batsmen can perform well under pressure for a whole series, for example, we must allow that with smaller and smaller sample sizes this could be in large part due to luck. It could also be that those batsmen who show up as being better under pressure are just those who are or were better anyway, such as Hobbs and Bradman. In any case, if we look at a complete career, then the luck component must be reduced, such that any conclusions we draw can at least be considered to be less impacted by the luck factor.

The impact of…impact
I’ve been banging on for ages now about match impact, i.e. assessing the impact of a cricketer’s performance on the win probability of the match, such that any changes in the match status, whether positive or negative, are assigned to the players involved based on their contributions with bat, ball and in the field. However, it occurred to me that it could be that a 20% increase in win probability may not have the same significance if it takes the win probability from 75% to 95% (or from well ahead to virtual certainty) than if it takes the chance of a win from 30% to 50%, or from noticeably behind to the match being in the balance. The insight, such as it is, which I perceived was that I had information available to me which would enable an objective investigation of those clutch performances, such as they are in Test cricket. This database would inform us as to the match status, in terms of win probability, which existed when a player came in to bat at any time in the match, and irrespective as to the outcome. Armed with such information for every innings for all Test players, it would be possible to build up a picture of a player’s clutch ability based on his situational average achieved with the varying win probabilities present at the initiation of each of his innings.

It’s just sad, Wilson, what a colossal geek you are
As mentioned above, I have compiled a large database of Test match FOWs which allows us to find a team’s Win Probability (WP) at any time in a match based on the results of previous matches which enjoyed the same status in terms of runs scored and wickets remaining. After reviewing the various match states and related WPs, it was noticeable that there are identifiable cut-off points at which a game’s status (in terms of being ahead in the match, the match being in the balance, or the team being behind) would be apparent to the incoming batsman. I found these points to be betwen 60% and 70% WP for the batting team as being evidently ahead, and between 30% and 40% WP for the batting team being evidently behind. In between these thresholds, the match is considered to be in the balance.

I had intended at first to group innings into three distinct categories, based on the scenarios outlined above, however further consideration suggested this wasn’t the best solution. For example, if we made the ‘team behind’ situation as being represented by a WP of between 0% and 30%, this would include the ‘nothing to lose’ slug-fests typified by Ian Botham at Headingley in 1981 and Nathan Astle in 2002, who both stated they felt no pressure coming into bat as there was nothing to lose by going for it. One solution to that scenario would be to raise the low threshold to 10%, but then we would also lose classic performances such as Brian Lara’s virtually single-handed match-winning 153* in 1999 and Ranji’s 154* in the 1896 Ashes at Old Trafford, the latter which was rated by Wisden at that time as ‘the finest display ever seen on a great occasion’.

I had also originally decided to include only performances made during the team’s second innings, however this would exlude gems like Frank Worrell carrying his bat for 191* at Trent Bridge in 1957. After discussions with informed and interested CW forum members, I decided that the critical performances should include all innings when the batting team had a WP below 60% when the batsman came to the crease.

Going back to Botham – if Bob Willis had taken 5-60 and Australia had won by three wickets, would Botham’s batting heroics be viewed in the same light now? Either way, his state of mind on coming in to bat would have been the same, as he couldn’t have known at that time that Willis would scuttle out the Aussies for just 111. So that’s why I decided to use the WP at the start of each innings, rather than on the match result.

The Importance of Runs
If we assume that runs made while the team is evidently ahead have less significance than when the team is behind or the match is in the balance, we can separate a batman’s important career runs from those which are less critical. So what we’ll do is group together batting performances made when the batting team’s WP is less than 60% when the batsman started his innings. While we will compare all batsmen on those performances, we will also compare the batsman to himself by contrasting his critical performances to those made when his team was ahead when he walked in.

Here are some aggregate numbers to set the scene:-

Important innings: 51914
Important runs: 1,398,611

Regular innings: 20026
Regular runs: 536612

So that’s about 72% of all Test innings played which meet our classification of important (i.e. WP <60%).

OK now, brace yourselves; below is the first of many tables to come. This first one shows those batsmen who have compiled the highest number of important runs over their career. For comparison I’ve also shown their total number of important innings, the number of important innings as a percentage of total innings played, the number of important runs as a percentage of the total runs scored, the player’s batting average in important innings, his regular innings batting average and the difference between them, and finally the difference between the important runs percentage and the important innings percentage – this measure indicates if the player turned a lower number of innings into a higher number of runs:-

Imp. Runs Player Imp% ImpInns ImpAv RegAv AvDiff %Diff
11812 Tendulkar (234) 74.1% 71.3% 55.98 47.94 +8.04 +2.8%
9988 Lara (183) 83.8% 80.3% 56.43 42.76 +13.67 +3.5%
9560 Chanderpaul (227) 81.0% 81.4% 49.79 58.97 -9.18 -0.4%
9558 Ponting (165) 72.2% 57.9% 65.02 33.40 +31.62 +14.3%
9307 Dravid (201) 71.2% 71.8% 51.42 53.14 -1.72 -0.6%
8919 Gavaskar (176) 88.1% 82.2% 54.72 36.46 +18.26 +5.9%
8606 Cook (210) 70.4% 72.7% 42.01 48.76 -6.75 -2.3%
9327 Sangakkara (164) 67.2% 70.6% 53.72 66.62 -12.90 -3.4%
7957 Kallis (173) 61.6% 63.4% 50.36 63.51 -13.15 -2.2%
7572 Border (186) 68.1% 70.5% 48.23 56.38 -8.15 -2.4%
7473 Jayawardene (176) 63.9% 71.3% 44.22 64.95 -20.73 -7.4%
7360 Younis Khan (156) 72.9% 74.3% 51.11 58.28 -7.17 -1.4%

I’ve included those additional details as the top “aggregate” batsmen shown above may not feature in the other tables, and I thought it would be interesting to compare them on those parameters. What stands out like a beacon is Ponting’s amazing performance when it matters, his average of 65.02 being way ahead of the rest of these modern all-time greats. Also, look at the number of times Sunil Gavaskar came in to bat when his side was level or behind as compared to Ponting. Nonetheless, Ponting turned 58% of all his Test innings into 72% of his total of Test runs, when it mattered most. That difference of +14.3% is miles ahead of the batsmen mentioned above. Ponting is also significantly ahead in terms of average differential, with +31.62, followed by Gavaskar (+18.26), Lara (+13.67) and Tendulkar (+8.04).

The next list shows the highest values of important runs (iRuns) as a percentage of total runs, with a 40/40 rule applied, i.e. a minimum of 40 important innings (iInns) played, as well as a minimum batting average of 40.00:-

%iRuns Player (iRuns & iInns)
95.5% B Sutcliffe (2603-93.4%)
93.4% AD Nourse (2765-84.1%)
90.2% Flower (4323-88.4%)
88.3% Faulkner (1549-85.1%)
88.1% Gavaskar (8919-82.2%)
88.1% HW Taylor (2587-82.9%)
87.8% EAB Rowan (1726-84.0%)
87.8% Woodfull (2020-79.6%)
87.4% Sarwan (5106-82.5%)
87.3% H Sutcliffe (3975-81.0%)
86.9% B Mitchell (2953-88.6%)
85.9% Hazare (1882-96.2%)
85.7% Viswanath (5209-79.5%)
85.6% Hunte (2779-82.1%)
84.5% Chandimal (2978-80.5%)
84.3% Sehwag (7163-77.8%)
84.3% Amiss (3042-79.5%)
83.8% Lara (9988-80.3%)
83.7% Gayle (5974-81.0%)
83.6% Hanif (3272-82.5%)

Geoff Boycott is next, with 83.4% on 6769 runs. Largely, the top of the list features good (or great) batsmen on poor teams. Herbert Sutcliffe’s presence can’t be explained in the same way, however many of these players, like Gavaskar, Woodfull and Sutcliffe, were openers – openers will often come in to bat, at least in their first innings, with the match classified as “in the balance”.

Next is a list of those who have crafted the highest number of important 50s, showing the total number of 50s, the name of the player (with total important innings) and the number of important 50s as a percentage – the list below includes all players with at least 40 important 50s:-

i50s Player %i50s
90 Tendulkar (234) 38.46%
76 Chanderpaul (227) 33.48%
74 Ponting (165) 44.85%
70 Gavaskar (176) 39.77%
68 Lara (183) 37.16%
67 Dravid (201) 33.33%
63 Cook (210) 30.00%
62 Sangakkara (164) 37.80%
60 Kallis (173) 34.68%
57 Border (186) 30.65%
56 Boycott (152) 36.84%
55 SR Waugh (155) 35.48%
54 Jayawardene (176) 30.68%
52 Miandad (137) 37.96%
52 Inzamam (144) 36.11%
52 Laxman (156) 33.33%
52 Gooch (168) 30.95%
50 Gower (156) 32.05%
50 Younis Khan (156) 32.05%
48 Atherton (165) 29.09%
47 Stewart (177) 26.55%
45 Sehwag (137) 32.85%
44 Viswanath (124) 35.48%
44 Bell (127) 34.65%
43 V Richards (113) 38.05%
43 SP Fleming (141) 30.50%
42 Root (99) 42.42%
42 Vengsarkar (145) 28.97%
42 Gayle (145) 28.97%
41 Youhana (116) 35.34%
41 MA Taylor (124) 33.06%
40 Hutton (97) 41.24%
40 ME Waugh (114) 35.09%
40 CH Lloyd (118) 33.90%
40 Amla (121) 33.06%
40 NA Hussain (124) 32.26%
40 Cowdrey (125) 29.41%
40 Thorpe (136) 29.41%

Tendulkar is miles ahead with 90 fifties in 234 important innings. Again we can see the massive success of Ricky Ponting, who passed fifty in almost 45% of his important innings. Not far behind percentage-wise is Joe Root – Root has an impressive number of fifties in important innings, though the criticism typically aimed at him is that he doesn’t convert these into hundreds. But his percentage of fifties is nonetheless impressive, and compares favourably for example with Australia’s banned pair of Steve Smith and David Warner; Root notched up 42 important fifties out of 132 total innings (32%), while Smith scored 28 out of 114 (25%) and Warner 22 of 137 (16%). Gooch and Gower are very close to each other, with Gower slightly edging Gooch percentage-wise.

Looking at the number of important fifties as a percentage, and applying the earlier 40/40 rule, the below list shows the percentage, the player and both the number of important fifties and from how many important innings these were crafted:-

%i50s Player (i50s/iInns)
52.54% Bradman (31/59)
48.53% H Sutcliffe (33/68)
44.85% Ponting (74/165)
43.82% Barrington (39/89)
43.40% AD Nourse (23/53)
42.42% Root (42/99)
41.86% Woodfull (18/43)
41.27% May (26/63)
41.24% Hutton (40/97)
40.82% Walcott (20/49)
40.70% DA Warner (35/86)
40.58% SPD Smith (28/69)
40.00% MH Richardson (20/50)
40.00% EJ Barlow (16/40)
39.77% Gavaskar (70/176)
38.81% Katich (26/67)
38.46% Tendulkar (90/234)
38.36% Hobbs (28/73)
38.05% IVA Richards(43/133)
37.97% MEK Hussey (30/79)

As well as having scored by far the highest number of fifties, Tendulkar also has a good percentage, also showing up in the top 20 when limiting to the front-line batsmen. Stanley Jackson enjoyed a percentage of 40.91% from 22 innings, while Shane Watson just missed our threshold, managing a fifty in 15 of his 39 important innings for 38.46%. Ponting has many more important innings than most of the batsmen in this list, certainly in the top 14, yet maintained a very high percentage. Bradman is the only one of the recognised batsmen to manage a fifty in more than half of his important innings.

We saw earlier that Ponting had maintained a very high average in important innings, 65.02 – let’s now see how that ranks among the recognised batsmen, again applying the minimum 40 clutch innings threshold, showing batting average in important innings, player and number of important innings:-

iAvg Player (iInns)
111.60 Bradman (59)
65.02 Ponting (165)
64.11 H Sutcliffe (68)
62.57 Hutton (97)
61.14 Barrington (89)
60.54 SPD Smith (69)
59.51 Sobers (107)
58.83 AD Nourse (53)
58.02 Hammond (90)
56.52 May (63)
56.43 Lara (183)
56.02 Walcott (49)
55.98 Tendulkar (234)
55.91 Hobbs (73)
55.34 Miandad (137)
55.09 SR Waugh (155)
54.72 Gavaskar (176)
54.52 AC Gilchrist (66)
54.13 Kohli (83)
53.86 Sehwag (137)
53.72 Sangakkara (164)
53.08 GS Chappell (94)
52.62 IVA Richards (113)
52.35 Compton (79)
52.25 SM Nurse (40)
52.08 Flower (99)

Stan McCabe just missed the threshold, with 38 important innings, but his important average of 63.61 would have seen him in fourth place. Bradman is miles ahead with a ridiculous average of 111.60 in his 59 important innings, even further ahead than he is in traditional batting average. This list includes many of the top batsmen as suggested by the received wisdom, and a number of those batsmen who are renowned as punishers.

Finally, below is the comparison of the players with themselves. Shown here is the differential for each player between their important average and their regular average (based on runs made in less important innings, when the player’s team was ahead), minimum 40 important innings and minimum regular innings average of at least 30 (shown is the differential, player, important average and regular average:-

Diff Player (iAvg-rAvg)
+34.32 Bradman (111.60-77.28)
+31.62 Ponting (65.02-33.40)
+23.39 May (56.52-33.13)
+19.60 Hutton (62.57-42.97)
+19.50 H Sutcliffe (64.11-44.62)
+19.00 Prince (51.18-32.19)
+18.94 SM Nurse (52.25-33.31)
+18.75 Sehwag (53.86-35.11)
+18.26 Gavaskar (54.72-36.45)
+15.63 Leyland (51.28-35.65)
+15.22 Slater (48.76-33.53)
+14.63 CC Hunte (47.91-33.29)
+14.54 Boycott (50.89-36.35)
+14.23 AR Morris (51.91-37.69)
+13.67 Lara (56.43-42.76)
+13.44 Amiss (49.06-35.63)
+13.42 MH Richardson (48.02-34.60)
+13.15 Asif Iqbal (43.19-30.04)
+12.86 Kanhai (50.69-37.84)
+12.76 AC Gilchrist (54.52-41.76)

While I appreciate that by placing a limit on the number of innings I’m introducing some selection bias, this is because otherwise the list is skewed by players on poor teams who didn’t have many regular, or non-important, innings.

We are used to seeing Bradman streets ahead of everyone else, and indeed he tops this list with his crazy averages, but in differential between important innings and regular innings he’s not far above the amazing Ponting, and these two are significantly above the rest. I have often felt that Peter May might have been over-rated and that his legend was perhaps undeserved, however it appears I was wrong. Ashwell Prince’s excellent conversion rate is just as impressive during important innings as with regular innings, with 8 of 17 fifties converted to centuries.

My co-staff member Martin (fredfertang) is a fan of Maurice Leyland, and he shows up well in this study (Leyland, not Martin). I am pleased to see Boycs, Arthur Morris and Dennis Amiss (another fertang fave) represented, and also not surprised to see Brian Lara and Adam Gilchrist highly ranked.


How did I arrive at the extrapolated average of 92.07 for Ponting? This was based on his differential as compared to that of DGB and compared to that famous number, 99.94.


Next time I’ll post a summary of all of the top batsmen from WG Grace to Jonny Bairstow, highlighting their performances in important innings.

Leave a comment

Comments are moderated, and will not appear until they have been approved

More articles by Dave Wilson