Evaluating ODI bowling spells

The author uses what he calls SVI (Spell Value Index) to rank bowling performances in One-Day Internationals.

Junaid Khan... a superb bowling performance of 9.0-1-43-4 in the ODI between India and Pakistan in Chennai on December 30, 2012.   -  AP

Let us take three ODI bowling spells from matches that were all won by the bowlers’ teams comfortably.

ODI #3739: Sunil Narine 9.5-0-27-6.

ODI #3244: Ben Hilfenhaus 9.3-1-33-5.

ODI #3314: Junaid Khan 9.0-1-43-4.

Without the benefit of additional information, it would be easy to rank these spells as Narine, Hilfenhaus and Junaid. After all, more wickets have been captured in more economic spells. Surely, the Narine spell should have contributed more to the team’s cause than that of Hilfenhaus, and much more than the one essayed by Junaid.

However, a very different story emerges when we delve into the scorecards and look at the batsmen dismissed, and at what score. The batsman’s score is in parentheses.

Narine: Hashim Amla (20), Rilee Rossouw (61), Farhaan Behardien (0), Chris Morris (0), Aaron Phangiso (3) and Imran Tahir (0.)

Hilfenhaus: Sachin Tendulkar (3), Virat Kohli (12), M.S. Dhoni (56), Irfan Pathan (19) and Zaheer Khan (9).

Junaid: Virender Sehwag (4), Virat Kohli (0), Yuvraj Singh (2) and Rohit Sharma (4).

It is clear that Narine’s collection of wickets is nothing great: The better batsmen were dismissed after scoring runs, and the lesser batsmen at low scores.

Hilfenhaus dismissed three top wickets, but one after many runs were scored. And we suddenly have a lot of respect for Junaid Khan’s spell from both points of view: batsmen dismissed and when they were dismissed.

Now if we apply our cricketing knowledge, common sense and instinct, we get the feeling that the spells should be ranked, Junaid Khan, Ben Hilfenhaus and Sunil Narine, in that order. How do we formalise this gut feel and instinct? How do we establish a foolproof methodology? How do we validate our own postulate?

I intend to do that in this article. To do that I have coined a new term called SVI (Spell Value Index). This will be determined by using the following four parameters. By ‘Spell’, I mean the complete bowling effort in the innings.

The RpAI (Runs per Adjusted Innings) of the dismissed batsman.

Notionally, how many runs were saved by the subject dismissal? Applicable only if the batsman was dismissed below his RpAI.

The absolute bowling run-rate (RpO).

The bowling run-rate, relative to the team bowling run-rate.

Even though this is an article on ODI bowlers, the cornerstone of this piece is a new measure that I have developed for batsmen: The RpAI, which is described in a separate box. It is a simple but very important derivation and readers should understand it well.

If career-to date figures are to be used, then I would have to use either the batting average or the RpI, with their inherent deficiencies. Hence, the career RpAI is used to determine the quality of the batsman. The number of wickets captured does not come in directly. However, it is in the picture indirectly through the first two parameters. This method is recognition of the following facts.

The wicket of a top-order batsman is more valuable than that of a lower-order batsman.

The dismissal of a top-order batsman at, say, 5 is far more valuable than when he is dismissed at, say, 116.

Per se, it is essential to reward bowling accuracy, through the absolute value.

However, the importance of relative accuracy has to be recognised. A bowling spell in which the bowler has conceded 5 runs per over in an innings, where the other bowlers have conceded, say, 7 runs per over, is more valuable than a bowling spell in which the bowler has conceded 4 runs per over in an innings, where the other bowlers have conceded, say, 3 runs per over.

A few examples:

Junaid Khan dismissing Kohli for 0 — he gets a credit of 47.3 for dismissing Kohli and 47.3 for dismissing Kohli for 0 (Total: 94.6).

Hilfenhaus dismissing Dhoni for 56 — he gets a credit of 37.3 for dismissing Dhoni and 0 for dismissing Dhoni past his RpI value (Total: 37.3).

Narine dismissing Amla for 20 — he is credited with Amla’s RpI value of 47.0 and 27.0 (Total: 74.0).

A couple of wickets with the same bowler-batsman combination, in the recently concluded India-Australia series, will go a long way in explaining this concept. In Chennai, Coulter-Nile dismissed Kohli for 5. This double blow secured Coulter-Nile 86.8 points (45.9+40.9) and goes a long way towards getting an excellent SVI value. In Kolkata, he dismissed Kohli for 92 and is credited with 45.9 points (45.9+0). He deserves this since the dismissal of Kohli, even after he had scored quite a few runs, was instrumental in keeping the Indian total down by about 30 runs.

The BSI is an amalgam of runs and suitably scaled points for accuracy calculations. As such, the number has been scaled to reflect the relative value in an easy-to-understand manner. A BSI of 100 represents the ultimate: A spell of possibly 9/10 wickets for very few runs against a strong Indian/Australian team. The current highest is around 80. It is just an index value.

It should be remembered that the BSI is non-contextual. The scores, result, importance and location of the match are not considered. For instance, the wonderful bowling effort by Gary Gilmour in the semifinals of the 1975 World Cup is enhanced because of the importance of the match. Similarly, Mohammad Amir’s incisive spell against India in the recent Champions’ Trophy final acquires significant value. As such, the CSI work should not to be treated as a ratings exercise. When we add these four contextual elements, we will have a ratings table. That is for a later date.

Let us look at some tables. I will present the tables containing the top spells, based on the BSI, for the complete selection of 3919 ODI matches, India, World Cups, Champions Trophy and by Innings. The article is current up to the fifth ODI between India and Australia, played in Nagpur.

The first table lists the top-20 bowling spells of all time, based on SVI.

Sno

ODI#

Year

BowlerTeamVs

Analysis

SVI

1

1724

2001

Waqar YounisPakistanEng

10.0-0-36-7

80.2

2

1776

2001

WPUJC VaasSri LankaZim

8.0-3-19-8

79.9

3

321

1985

Imran KhanPakistanInd

10.0-2-14-6

78.6

4

31

1975

GJ GilmourAustraliaEng

12.0-6-14-6

78.2

5

685

1991

Aaqib JavedPakistanInd

10.0-1-37-7

78.0

6

2345

2006

M NtiniSouth AfricaAus

9.3-4-22-6

76.6

7

1652

2000

WPUJC VaasSri LankaInd

9.3-1-14-5

76.0

8

1504

1999

SB JoshiIndiaSaf

10.0-6-6-5

75.8

9

1976

2003

AJ BichelAustraliaEng

10.0-0-20-7

74.5

10

3387

2013

B KumarIndiaSlk

6.0-1-8-4

74.1

11

3321

2013

KMDN KulasekaraSri LankaAus

10.0-2-22-5

73.3

12

782

1992

PV SimmonsWest IndiesPak

10.0-8-3-4

73.3

13

1986

2003

SE BondNew ZealandAus

10.0-2-23-6

73.0

14

2237

2005

Naved-ul-HasanPakistanInd

8.4-1-27-6

72.8

15

2212

2005

PT CollinsWest IndiesAus

10.0-1-43-5

72.8

16

2523

2007

SM PollockSouth AfricaPak

10.0-3-23-5

72.7

17

2735

2008

BAW MendisSri LankaInd

8.0-1-13-6

72.6

18

2754

2008

SCJ BroadEnglandSaf

10.0-3-23-5

72.4

19

1540

2000

GD McGrathAustraliaInd

10.0-4-8-4

72.3

20

2158

2004

DNT ZoysaSri LankaSaf

8.0-0-26-5

72.1

 

What can one tell of Waqar Younis’ spell? The top six English wickets, including 5 single-figure dismissals, have propelled him to the top. It is tough to keep an 7-wicket haul off the top of the table.

Chaminda Vaas captured the first eight wickets, seven at single-digit scores. It was only the fact that this was against Zimbabwe, prevented him from occupying the top position. Then come two identical performances: Imran Khan’s 6 for 14 against a strong Indian batting line-up, albeit in a losing cause, which is in third position, and Gary Gilmour’s 6 for 14 in the semifinals of the 1975 World Cup, which occupies the fourth position. Both spells included five top-order dismissals in single figures.

Aaqib Javed’s devastating spell, the first one in the second innings, completes the top-5. A clear difference is that four of Javed’s dismissals came after the batsmen reached reasonable scores. It can be seen that the positions 6-10 include two five-wicket and one four-wicket hauls. Only Bangladesh and Zimbabwe are not represented in the top-20. Utseya’s 5 for 36 is in 23rd position and Aftab Ahmed’s 5 for 31 is in 83rd position. No bowler has a second entry in the top-30 of the table. Makhaya Ntini has a second entry in the 33rd position.

Now let us look at the top performances in the World Cups.

Sno

ODI#

WC

BowlerTeamVs

Analysis

SVI

1

31

WC-1975

GJ GilmourAustraliaEng

12.0-6-14-6

78.2

2

1976

WC-2003

AJ BichelAustraliaEng

10.0-0-20-7

74.5

3

1986

WC-2003

SE BondNew ZealandAus

10.0-2-23-6

73.0

4

748

WC-1992

EA BrandesZimbabweEng

10.0-4-21-4

71.8

5

3617

WC-2015

MA StarcAustraliaNzl

9.0-0-28-6

70.7

6

3600

WC-2015

MR MarshAustraliaEng

9.0-0-33-5

70.6

7

730

WC-1992

MW PringleSouth AfricaWin

8.0-4-11-4

70.4

8

3607

WC-2015

TG SoutheeNew ZealandEng

9.0-0-33-7

66.8

9

1976

WC-2003

AR CaddickEnglandAus

9.0-2-35-4

65.3

10

3602

WC-2015

Sohail KhanPakistanInd

10.0-0-55-5

64.3

11

3147

WC-2011

Wahab RiazPakistanInd

10.0-0-46-5

63.6

12

3617

WC-2015

TA BoultNew ZealandAus

10.0-3-27-5

63.2

13

1969

WC-2003

A NehraIndiaEng

10.0-2-23-6

62.9

14

3608

WC-2015

JE TaylorWest IndiesPak

7.0-1-15-3

62.8

15

1965

WC-2003

CO ObuyaKenyaSlk

10.0-0-24-5

62.3

 

When we add context to Gary Gilmour’s figures, they become the best ever bowling spell across nearly 4000 ODI games. Look at the context: World Cup semifinals, playing away and a win. Add to this, the collection of wickets, at low scores, there is no doubt this is the greatest bowling spell in ODI history.

Andy Bichel’s collection of seven English wickets in the 2003 World Cup is in second place. Five of these batsmen reached decent scores. Shane Bond’s dismissal list is fascinating: Adam Gilchrist, Matthew Hayden, Ricky Ponting, Damien Martyn, Brad Hogg and Ian Harvey.

Eddo Brandes’ spell is only a four-wicket one, but look at the batsmen he dismissed: Graham Gooch, Allan Lamb, Robin Smith and Graeme Hick.

Mitchell Starc’s spell was in the recently concluded World Cup. However, only three of the six wickets were of top-order batsmen. Ashish Nehra’s match-winning spell against England is the best performance by an Indian bowler in the World Cup.

It is time to look at the top eight SVI-based spells in Champions’ Trophy games.

Sno

ODI#

ICC

BowlerTeamVs

Analysis

SVI

1

2438

IC-2006

M NtiniSouth AfricaPak

6.0-2-21-5

70.3

2

2428

IC-2006

MF MaharoofSri LankaWin

9.0-2-14-6

69.2

3

3894

IC-2017

Mohammad AmirPakistanInd

6.0-2-16-3

67.5

4

2441

IC-2006

KD MillsNew ZealandAus

10.0-1-38-4

64.5

5

3881

IC-2017

M MorkelSouth AfricaPak

7.0-1-18-3

62.3

6

2430

IC-2006

KD MillsNew ZealandSaf

6.0-0-18-3

61.4

7

2894

IC-2009

GC TongeWest IndiesPak

10.0-3-25-4

59.2

8

2428

IC-2006

WPUJC VaasSri LankaWin

6.0-2-6-2

58.6

 

The Champions Trophy and its predecessors do not boast of really great bowling performances. The best performance is by Makhaya Ntini in a group game against Pakistan in Mohali in 2006. Even though South Africa scored only 213 in their 50 overs, Ntini was unplayable. He dismissed Mohammad Hafeez, Imran Farhat, Younis Khan, Shoaib Malik and Kamran Akmal, all for single-digit scores. This was in his first spell and he had no chances to add to his tally. In the same tournament, Farveez Maharoof had a terrific spell of 6 for 14 against West Indies in Mumbai. However, two of the batsmen had reached decent scores and two others were late-order batsmen.

No Indian cricket follower would have forgotten Mohammad Amir’s burst in the recent Champions Trophy final at Lord’s. He dismissed Rohit Sharma for 0, Shikhar Dhawan for 21 and Kohli for 5. There was no way India would have chased the huge target after these blows. Incidentally, this is the only 3-wicket spell in the top-50 of the SVI table. Kyle Mills has two entries in the top-8.

Finally, let us have a look at the best ODI spells for India.

Sno

ODI#

Year

BowlerTeamVs

Analysis

SVI

1

1504

1999

SB JoshiIndiaSaf

10.0-6-6-5

75.8

2

3387

2013

B KumarIndiaSlk

6.0-1-8-4

74.1

3

1219

1997

BKV PrasadIndiaPak

5.0-1-17-4

66.1

4

1540

2000

J SrinathIndiaAus

10.0-2-30-4

65.6

5

594

1989

M PrabhakarIndiaPak

5.0-2-5-3

65.1

6

2745

2008

Zaheer KhanIndiaSlk

9.5-3-21-4

64.6

7

3498

2014

STR BinnyIndiaBng

4.4-2-4-6

64.2

8

1969

2003

A NehraIndiaEng

10.0-2-23-6

62.9

9

2274

2005

IK PathanIndiaZim

10.0-3-27-5

62.6

10

1300

1998

SR TendulkarIndiaAus

10.0-1-32-5

62.4

 

Sunil Joshi’s spell against South Africa in Nairobi in 1999 leads the entries. Joshi dismissed Boeta Dippenaar for 17, Herschelle Gibbs for 18, Hansie Cronje for 2, Jonty Rhodes for 1 and Shaun Pollock for 0. This spell helped dismiss South Africa for a paltry 117 and India won comfortably.

Bhuvneshwar Kumar’s spell is of recent vintage and was in some ways more effective than Joshi’s. In a rain-affected match, Sri Lanka was set a target of 177 runs in 26 overs. Kumar sent back Upul Tharanga for 6, Mahela Jayawardene for 11, Kumar Sangakkara for 0 and Lahiru Thirimanne for 0. Sri Lanka was reduced to 31 for 4 and could not recover.

Then come the two gentlemen pace bowlers from Karnataka. Venkatesh Prasad dismissed Saeed Anwar for 0, Aamer Sohail for 1, Inzamam-ul-Haq for 0 and Saleem Malik for 10. The match, however, was abandoned. Javagal Srinath’s dismissal of four top Australian batsmen was in a lost cause. Stuart Binny’s is of recent vintage. Nehra’s performance was in an important World Cup match. And finally, Tendulkar’s performance wiped out the Australian middle order in Kochi. Readers might wonder at the omission of Anil Kumble’s 6 for 12 in the Hero Cup final. The problem was that, barring Carl Hooper and Jimmy Adams, the other four dismissals were of late-order batsmen. However, once context is considered, Kumble’s tour de force gets into the Indian top-10 comfortably.

The top first innings spells, based on SVI, are given without any comments since most of these have been covered already.

Sno

ODI#

Year

BowlerTeamVs

Analysis

SVI

1

1724

2001

Waqar YounisPakistanEng

10.0-0-36-7

80.2

2

1776

2001

WPUJC VaasSri LankaZim

8.0-3-19-8

79.9

3

321

1985

Imran KhanPakistanInd

10.0-2-14-6

78.6

4

31

1975

GJ GilmourAustraliaEng

12.0-6-14-6

78.2

5

1504

1999

SB JoshiIndiaSaf

10.0-6-6-5

75.8

6

1976

2003

AJ BichelAustraliaEng

10.0-0-20-7

74.5

7

3321

2013

KMDN KulasekaraSri LankaAus

10.0-2-22-5

73.3

8

1986

2003

SE BondNew ZealandAus

10.0-2-23-6

73.0

9

2212

2005

PT CollinsWest IndiesAus

10.0-1-43-5

72.8

10

2523

2007

SM PollockSouth AfricaPak

10.0-3-23-5

72.7

 

Similarly the top second innings spells are presented.

Sno

ODI#

Year

BowlerTeamVs

Analysis

SVI

1

685

1991

Aaqib JavedPakistanInd

10.0-1-37-7

78.0

2

2345

2006

M NtiniSouth AfricaAus

9.3-4-22-6

76.6

3

1652

2000

WPUJC VaasSri LankaInd

9.3-1-14-5

76.0

4

3387

2013

B KumarIndiaSlk

6.0-1-8-4

74.1

5

782

1992

PV SimmonsWest IndiesPak

10.0-8-3-4

73.3

6

2237

2005

Naved-ul-HasanPakistanInd

8.4-1-27-6

72.8

7

2735

2008

BAW MendisSri LankaInd

8.0-1-13-6

72.6

8

2158

2004

DNT ZoysaSri LankaSaf

8.0-0-26-5

72.1

9

748

1992

EA BrandesZimbabweEng

10.0-4-21-4

71.8

10

2284

2005

B LeeAustraliaIcc

9.0-2-30-4

71.4

 

It will be of interest to list the top-3 bowling spells in the recently concluded India-Australia and England-West Indies ODI series. These will be very fresh in the minds of the followers. All three have SVI values exceeding 50. I am particularly happy at the recognition of Bhuvneshwar Kumar’s spell, especially as I felt that he should have been the Man of the Match, ahead of Kohli.

ODI # 3910 (Chennai): Coulter-Nile 10.0-0-44-3 (SVI=60.1).

ODI # 3916 (The Oval): Alzzari Joseph 8.1-0-56-5 (SVI=57.3)

ODI # 3912 (Kolkata): Bhuvneshwar Kumar 6.1-2-9-3 (SVI=50.4).

An interesting bit of information: Coulter-Nile’s bowling performance is placed at 211th position in the all-time SVI table.

 

Runs per adjusted innings

The ODI Batting Average is a lopsided measure. It favours the middle-order batsmen to such an extent that some of the high values we see are grossly unfair to the top-order batsmen. Even amongst the middle-order batsmen, the percentage of not outs varies considerably. I have given below the not out percentage for a few top batsmen.

Kohli (17.2%), Bevan (34.2%), Dhoni (28.6%), Amla (6.5%), Hussey (28.0%), Richards (14.4%), Tendulkar (9.1%), Warner (4.3%), Lara (11.1%), Gilchrist (3.7%), Smith (14.3%), de Kock (4.7%) and so on. The range is mind-boggling: 3.7% to 34.2%. The range across all qualifying batsmen is still wider: 0.7% to 36.5%.

It is no wonder that Bevan, Dhoni and Hussey have high averages and Gilchrist and Warner have middling averages.

In order to correct this anomaly and create an equitable measure, I have created a new measure called Runs per Adjusted Innings (RpAI). It is very simple in concept and very easy to derive. The salient points are explained below. I have accepted that the adjusted number of innings will be a decimal number.

*All dismissals will be treated as one completed innings.

*All not out innings above the RpI value will be treated as one complete innings.

*All not out innings below the RpI value will be treated as part innings and determined pro-rata.

The RpAI is equal to the total runs scored divided by the sum of the above adjusted innings.

A simple example: Suppose a batsman plays 10 innings and the scores are 5, 75*, 16*, 45, 22, 80, 8*, 31, 85 and 33, his aggregate score is 400 and his bare RpI works out to 40.0. However, his batting average is a very flattering 57.14 (400/10-3). This is eminently incorrect. The RpAI working is given below. Each innings is adjusted as explained already.

1.0 (5), 1.0 (75*), 0.4 (16*), 1.0 (45), 1.0 (22), 1.0 (80), 0.2 (8*), 1.0 (31), 1.0 (85), 1.0 (33).

The total adjusted innings value is 8.6. The RpAI works to 46.51 (400/8.6), which is a very fair reflection of this mini-career. The batsman does not get undeserved credit for remaining unbeaten on low scores, nor does he get penalised. It is a fair measure irrespective of the positions the batsman batted in.

A question may be asked — why not the straight RpI (Runs per Innings)? That would be eminently unfair to the middle-order batsmen who do not have chances to complete their innings. The RpAI value will be somewhere in between the Batting Average and the RpI.

Unfortunately, the Batting Average still tends to get used everywhere because most of the followers of the game are familiar with that metric. However, all my analytical work is based on this very sound and almost foolproof measure. RpAI is developed only at the career level.