SELECT BLOG
October 26, 2009

## Analysing bowlers in Test wins

Anantha Narayanan
Enlarge

A few days back I posted an article on the runs scored by batsmen in winning cause. A number of comments were received which indicated that the batting averages in winning Tests is a very important indicator. I have done the work but will post the tables in a later article since I want to do justice to the bowlers. In fact the bowlers' analysis is as different from the batsmen analysis as chalk and cheese.

The reason is very simple and fundamental. Look at the following two Tests.

In 1932, Australia scored 153 runs in the match and WON.

```South Africa:36 & 45.
Australia: 153.
```

In 1921, England scored 817 runs in the match and LOST.

```Australia: 354 & 582.
England: 447 & 370.
```

The common thread running through these two extreme matches is that the winning team captured 20 wickets. This is the mandatory requirement of all wins, barring a few matches in which there might have been declarations or retired-hurt situations.

So I am going to take a somewhat different look at the bowlers' analysis. I have also been influenced by Unnikrishnan's excellent suggestion that the % runs should be calculated for each match, summed and averaged. I applied that to the bowler analysis. However let me inform Unni that there is almost no difference at all in the two ways of calculations since the team wickets is 20 for over 99% of the matches. There would obviously be a difference in batting because the total team runs in won matches vary a lot. I have also compared the bowling averages of bowlers, in winning causes, to the bowling averages of the other bowlers.

This time I have done a table of the top 25 for each of these analysis and a single team-based table, listing only the top-10 for each team. The full table is available through a link.

The criteria is simple. The bowler should have been involved in a minimum of 10 wins and captured over 100 wickets in their career.

1. Top 25 bowlers based on % of team wickets in wins

```No Cty  Bowler            Mat Wins  Wkts  Wkts %-of-Wkts
Own  Team

1.Eng  Barnes S.F         27   13   115   260   44.23
2.Slk  Muralitharan M    129   53   430  1060   40.57
3.Nzl  Hadlee R.J         86   22   173   440   39.32
4.Aus  Grimmett C.V       37   20   143   400   35.75
5.Ind  Chandrasekhar B.S  58   14    98   276   35.71
6.Saf  Steyn D.W          33   18   124   360   34.44
7.Saf  Tayfield H.J       37   11    74   220   33.64
8.Ind  Kumble A          132   43   284   860   33.02
9.Aus  Lillee D.K         70   31   203   618   32.80
10.Aus  O'Reilly W.J       27   14    91   279   32.61
11.Eng  Fraser A.R.C       46   12    78   240   32.50
12.Eng  Peel R             20   12    78   240   32.50
13.Eng  Lohmann G.A        18   15    94   300   31.33
14.Aus  McKenzie G.D       60   18   112   360   31.11
15.Eng  Gough D            58   18   105   342   30.83
16.Pak  Imran Khan         88   26   155   520   29.81
17.Win  Marshall M.D       81   43   254   857   29.62
18.Win  Ramadhin S         43   13    76   260   29.23
19.Ind  Bedi B.S           67   17    97   336   28.90
20.Win  Croft C.E.H        27   10    57   200   28.50
21.Pak  Waqar Younis       87   39   222   780   28.46
22.Saf  Donald A.A         72   33   187   660   28.33
23.Eng  Caddick A.R        62   21   114   402   28.27
24.Aus  Davidson A.K       44   16    89   320   27.81
25.Aus  Trumble H          32   14    77   280   27.50
```
Let us give Barnes his place at the top. That is to be expected, considering that he captured 7 wickets per Test which became nearly 9 per Test in won matches. Muralitharan and Hadlee's high +-40% is to be expected considering that they were the leading bowlers for their respectiove teams, by a wide margin. Grimmett is also to be expected. This single position is also enough to show the contribution that Chandrasekhar has made for Indian cricket. Steyn is fast emerging as one of the great bowlers. Then come the two great spinners, Tayfield and Kumble. Lillee's 6.5 wickets per Test for a strong Australia is a revelation. The top-10 is rounded off by O'Reilly, the other great leg spinner of the 1920s.

The top-10 has 6 spinners. Also 6 modern bowlers appear in these positions.

2. Top 5 bowlers for each country based on % of team wickets in wins

```Cty  Bowler            Mat Wins  Wkts  Wkts %-of-Wkts
Own  Team

Aus  Grimmett C.V       37   20   143   400   35.75
Aus  Lillee D.K         70   31   203   618   32.80
Aus  O'Reilly W.J       27   14    91   279   32.61
Aus  McKenzie G.D       60   18   112   360   31.11
Aus  Davidson A.K       44   16    89   320   27.81
...
Eng  Barnes S.F         27   13   115   260   44.23
Eng  Fraser A.R.C       46   12    78   240   32.50
Eng  Peel R             20   12    78   240   32.50
Eng  Lohmann G.A        18   15    94   300   31.33
Eng  Gough D            58   18   105   342   30.83
...
Ind  Chandrasekhar B.S  58   14    98   276   35.71
Ind  Kumble A          132   43   284   860   33.02
Ind  Bedi B.S           67   17    97   336   28.90
Ind  Harbhajan Singh    77   31   168   619   27.13
Ind  Prasanna E.A.S     49   15    81   300   27.00
...
Nzl  Hadlee R.J         86   22   173   440   39.32
Nzl  Martin C.S         50   12    59   240   24.58
Nzl  Cairns C.L         62   16    76   320   23.75
Nzl  Chatfield E.J      43   12    52   240   21.67
Nzl  Cairns B.L         43   12    48   240   20.00
...
Pak  Imran Khan         88   26   155   520  29.81
Pak  Waqar Younis       87   39   222   780  28.46
Pak  Wasim Akram       104   41   211   820  25.73
Pak  Danish Kaneria     54   21   108   420  25.71
Pak  Shoaib Akhtar      46   20    99   400  24.75
...
Saf  Steyn D.W          33   18   124   360  34.44
Saf  Tayfield H.J       37   11    74   220  33.64
Saf  Donald A.A         72   33   187   660  28.33
Saf  Ntini M            99   50   233  1000  23.30
Saf  Pollock P.M        28   10    46   200  23.00
...
Slk  Muralitharan M    129   53   430  1060  40.57
Slk  Vaas WPUJC        111   43   166   860  19.30
...
Win  Marshall M.D       81   43   254   857  29.62
Win  Ramadhin S         43   13    76   260  29.23
Win  Croft C.E.H        27   10    57   200  28.50
Win  Roberts A.M.E      47   21   110   420  26.19
Win  Ambrose C.E.L      98   44   229   878  26.12
```
The list is elf-explanatory. The Indian top-5 are all spinners. Quite surprising is the presence of Ramadhin amongst great West Indian fast bowlers and the very high placing of Fraser, McKenzie and Kaneria.

3. Top 25 bowlers based on Ratio of bowling average in wins

```No Cty  Bowler           Wkts  <-Wins Bow Avge-> Ratio
Team   Own Others

1.Eng  Fraser A.R.C       78  24.20 16.53 27.90  1.69
2.Nzl  Hadlee R.J        173  18.38 13.07 21.82  1.67
3.Pak  Imran Khan        155  20.16 14.50 22.56  1.56
4.Eng  Barnes S.F        115  17.71 13.58 20.98  1.54
5.Slk  Muralitharan M    430  20.57 16.04 23.66  1.47
6.Saf  Steyn D.W         124  21.33 16.68 23.77  1.43
7.Pak  Shoaib Akhtar      99  21.78 17.52 23.19  1.32
8.Eng  Briggs J           84  16.01 13.01 16.86  1.30
9.Aus  Davidson A.K       89  19.52 16.04 20.86  1.30
10.Aus  McKenzie G.D      112  23.47 19.49 25.27  1.30
11.Eng  Underwood D.L     123  18.65 15.19 19.67  1.30
12.Aus  O'Reilly W.J       91  17.84 14.96 19.23  1.29
13.Aus  Lillee D.K        203  21.56 18.27 23.18  1.27
14.Win  Gibbs L.R         154  22.93 19.17 24.23  1.26
15.Saf  Goddard T.L        47  23.03 19.09 24.10  1.26
16.Eng  Verity H           71  20.01 16.65 20.97  1.26
17.Eng  Lohmann G.A        94  11.21  9.67 11.91  1.23
18.Ind  Pathan I.K         66  23.70 20.26 24.88  1.23
19.Eng  Peel R             78  16.97 14.67 18.07  1.23
20.Aus  Grimmett C.V      143  19.99 17.60 21.32  1.21
21.Aus  Trumble H          77  20.79 18.00 21.85  1.21
22.Eng  Bedser A.V         74  20.09 17.54 21.04  1.20
23.Ind  Kumble A          284  21.18 18.71 22.40  1.20
24.Saf  Pollock P.M        46  22.86 19.83 23.77  1.20
25.Win  Croft C.E.H        57  19.39 17.12 20.29  1.18
```
I have ordered this table on the ratio of own wickets average to other bowlers wicket average in won matches. Fraser is on top having outr=performed his peers in won matches by 69%. I am not able to expplain this other than possibly the relatively weaker English attacks. Hadlee is next. However note the stunning contributions made by Imran Khan in their wins, over 55% better. Muralitharan, is next. Shoaib Akhtar comes into the top-10 as also the great left arm fast bowler, davidson.

Note the low averages by the concerned bowlers in wins. No doubt these figures would be influenced, partly, by the outstanding analysis against weaker teams. But neither Fraser nor Hadlee had one easy match in their careers.

4. Top 5 bowlers for each country based on Ratio of bowling average in wins

```Cty  Bowler           Wkts  <-Wins Bow Avge-> Ratio
Wins  Team   Own Others

Aus  Davidson A.K       89  19.52 16.04 20.86  1.30
Aus  McKenzie G.D      112  23.47 19.49 25.27  1.30
Aus  O'Reilly W.J       91  17.84 14.96 19.23  1.29
Aus  Lillee D.K        203  21.56 18.27 23.18  1.27
Aus  Grimmett C.V      143  19.99 17.60 21.32  1.21
...
Eng  Fraser A.R.C       78  24.20 16.53 27.90  1.69
Eng  Barnes S.F        115  17.71 13.58 20.98  1.54
Eng  Briggs J           84  16.01 13.01 16.86  1.30
Eng  Underwood D.L     123  18.65 15.19 19.67  1.30
Eng  Verity H           71  20.01 16.65 20.97  1.26
...
Ind  Pathan I.K         66  23.70 20.26 24.88  1.23
Ind  Kumble A          284  21.18 18.71 22.40  1.20
Ind  Bedi B.S           97  19.43 17.66 20.14  1.14
Ind  Chandrasekhar B.S  98  20.83 19.28 21.69  1.13
Ind  Prasanna E.A.S     81  19.04 17.62 19.57  1.11
...
Nzl  Hadlee R.J        173  18.38 13.07 21.82  1.67
Nzl  Cairns C.L         76  21.35 20.20 21.70  1.07
Nzl  Bracewell J.G      35  19.54 19.29 19.59  1.02
Nzl  Chatfield E.J      52  18.39 19.00 18.22  0.96
Nzl  Vettori D.L       109  19.07 21.40 18.52  0.87
...
Pak  Imran Khan        155  20.16 14.50 22.56  1.56
Pak  Shoaib Akhtar      99  21.78 17.52 23.19  1.32
Pak  Waqar Younis      222  19.84 18.21 20.49  1.13
Pak  Sarfraz Nawaz      75  21.47 20.52 21.76  1.06
Pak  Wasim Akram       211  18.63 18.49 18.68  1.01
...
Saf  Steyn D.W         124  21.33 16.68 23.77  1.43
Saf  Goddard T.L        47  23.03 19.09 24.10  1.26
Saf  Pollock P.M        46  22.86 19.83 23.77  1.20
Saf  Tayfield H.J       74  20.98 18.85 22.05  1.17
Saf  Donald A.A        187  18.77 16.80 19.56  1.16
...
Slk  Muralitharan M    430  20.57 16.04 23.66  1.47
Slk  Vaas WPUJC        166  20.20 22.64 19.62  0.87
...
Win  Gibbs L.R         154  22.93 19.17 24.23  1.26
Win  Croft C.E.H        57  19.39 17.12 20.29  1.18
Win  Marshall M.D      254  18.70 16.79 19.50  1.16
Win  Ambrose C.E.L     229  18.66 16.86 19.29  1.14
Win  Ramadhin S         76  19.08 17.80 19.61  1.10
```
The table is self-explanatory. Note the vast difference between Muralitharan, Hadlee and their support bowlers. Also Gibbs leads the West Indian list.

Anantha Narayanan has written for ESPNcricinfo and CastrolCricket and worked with a number of companies on their cricket performance ratings-related systems

Feeds: Anantha Narayanan

Posted by Vikram Maingi on (November 9, 2009, 7:08 GMT)

I do not know much about Syd Barnes but his stats are really too good. As far as Hadlee and Murali are concerned, they hardly got any support from the rest of the team. This was the main reason of their percentages going up. When Sir Hadlee retired from Test Cricket with 431 wickets in his kitty, Lance Cairns was New Zealand's next highest wicket taker with 130 wickets. Glenn McGrath and Warne do not find a place in tehe list due to exactly opposite reasons. Whole of the team was such a great unit that all the bowlers were amongst wickets, a few at least. Due to similar reasons the percentage of Garner, Holding and Wasim Akram are not so high.

Posted by Arvind on (November 7, 2009, 2:45 GMT)

This analysis is a prominent measure of how pathetic the other bowlers in the team were. It is also the reason why Warne and McGrath are somewhat lower in the table, because they had each other for support, besides thier teammates like Brett Lee, Jason Gillespie were good bowlers in thier own right.

Posted by ted on (November 3, 2009, 11:06 GMT)

good to see grimmett there in the top of aussie lists.a forgotten genius. warne&mcgrath not in the top 5 maybe because they were the go to men when times where tough & wickets hard to come by. [[ Ted They are not in the top 5 because of the reasons you mention. They are not there because they shared the wickets. Still McGrath is in 6th position. Ananth: ]]

Posted by Ashwath on (October 31, 2009, 3:10 GMT)

Hello sir, refreshing article as always. Is it possible for you to do an analysis on captains. To analyse which captains brought out the best in their players for instance dravid averaged about 71 under ganguly which is 19 more than his career mean and kumble averaged 24 with the ball under azhar which is 5 less than his career mean. Any chance of an analysis under these aspects. Am sorry if it sounds tedious. thanks [[ Ashwath Captains' analysis has been done already. Pl look at earlier articles. Ananth: ]]

Posted by Unni on (October 31, 2009, 2:13 GMT)

Ananth: thanks for taking the suggestion. But obvious that it wouldn't make any difference when all denominators are same. No further comments on this article as the idea is same. Waiting for the batting analysis ....

Posted by Youvi on (October 28, 2009, 3:46 GMT)

Ananth- Very stimulating analysis. It started me thinking in different ways. Not surprisingly perhaps, in the top five by %-age of wickets, Hadlee, Barnes and Muralitharan are way ahead of the second bowler for their respective countries. I also notice that among the top five WI bowlers by % of wickets, the spread is the least ~3.5%. Interesting also to find Ramadhin in the top five ! Moreover, other than Ramadhin the careers of the other WI bowlers overlapped quite a bit when WI were the dominant force. Also, the %-spread among the top five Pakistan bowlers is 5% (least after WI). Interestingly, both Ramadhin and Kaneria are spinners while rest of top five in respective teams are fast bowlers. For both WI and Pak these fast bowlers had overlapping careers with their other team-mates in the top five %-age of wickets list. Intriguing stuff !

Posted by alex on (October 27, 2009, 18:03 GMT)

Ananth - congratulations on a top class article. One suggestion: in the long run (cut-off of, say, 40 wins), normal distribution may be expected to figure in somehow. Your columns give the mean values ... could you pls consider adding a column on variance? Thanks.

Posted by Xolile on (October 27, 2009, 14:30 GMT)

In order for a bowler to do well in this measure he needs to meet at least three of the following criteria:

1 The bowler was very good 2 He had the ability to bowl long spells 3 He was utilized to his optimal capacity 4 The conditions generally suited his style of bowling 5 The other bowlers in his team were fairly weak 6 The opposition batsmen were generally below average 7 The statistical depth is insufficient

Barnes ticks boxes 2 to 7; Murali boxes 1 to 6; and Hadlee boxes 1, 3 and 5. That is why they are at the top of the list.

Roberts, Holding, Marshall and Garner only meet the first criteria and are therefore not high on the list. [[ Deon I am not saying that I agree with all your conclusions. However the idea , subjective though it is, is very nice. Kumble is probably 1 to 4. Warne, 1 to 3 and so on. In defence of Barnes, let me suggest that either any 3 or 2-7 should do. Ananth: ]]

Posted by Abhi on (October 27, 2009, 6:05 GMT)

Interesting analysis. As with the similar batting analysis this also indicates the mediocrity of the rest of the bowlers in the team…considering that the likes of Murali, Hadlee, Kumble have high percentages. Also, considering that all the top Indian bowlers are spinners, just goes to show how pathetic a bowling unit India was till a few years back.Steyn is a bit of a surprise, but these are relatively early days for him…and he will require to play at least twice the number of matches more in both formats ( maintaining his standards) before we can fairly compare him to the legends. Lillee, my all time favourite bowler, is the biggest and most pleasant surprise!! Much like Tendulkar, the way he came back from injury (doctors told Lillee he may never walk again, never mind bowl)-strengthening his body, restructuring and remodeling his entire bowling action and technique and STILL being able to perform at the highest level…Talent, brute force of will, bloody minded toughness…the Ultimate fast bowler. Love it when some entirely unexpected stats and analysis reinforce some old notions!

@aditya jha…again goes to show the depth of the Aus team!! Blasphemy though it may seem, I would wager that even without McGrath the Aus team would have done as well. They have always produced quality quicks. Warney would have been much more difficult to replace. [[ Abhi As I had explained to Angus, in many of the analysis I let the readers draw their own conclusions. They are, of course, free to share their views with other readers. McGrath/Warne and the great West Indian fast bowlers would not figure at the top in such an analysis since they necessarily had to share the spoils. In no way does this lower their greatness an iota. In their own diffrerent ways, Warne, Murali and Kumble share the pedestal. An aside. Can you send in a comment with a mailid through which I can reach you. Your current id seems to be a dummy one and bounces mails. The mailid will remain confidential with me. Ananth: ]]

Posted by Sesha on (October 27, 2009, 3:27 GMT)

Hi Anath,

Great job and good thought process. Your introduction to the article has popped up a few possible future analyses...Which team took the least wickets to win a test(minimum should be 10 assuming no retired hurts)2. Overall which team (may be decade wise) have taken the least & most wickets and runs for their victories....consider these analysis so that we can see the relative strength of bowling and batting of each team in each decade.... [[ Sesha Your mail has given me an idea that I should, once in a while, do an article presenting a collection of snippets like the ones you have mentioned, which are not big enough to fill an entire article but deserve a paragraph or two. Thanks Ananth: ]]