Cricket Case Study - Assessing Gujarat Lions Squad, 2017

Skype: Sports Analytics Advantage

THE SEASON SO FAR


The 2017 IPL season has not started well for the Gujarat Lions, with crushing defeats occurring in both their opening matches.

Against Kolkata Knight Riders in the opener, after being inserted, Lions managed a solid-looking 184/8 in their 20 overs, mainly thanks to wicket-keeper Dinesh Karthik's superb 47 off 25 balls late in the innings.  

Arguably, this total could have been bigger, given captain Suresh Raina's 68 off 51 balls.  On the surface, a batsman scoring 68 not out would look to have performed creditably, but Raina used up almost half of the team's resources (balls faced) in this innings, and hit only seven boundaries in 51 balls (13.73%) - a figure below the average for T20 batting.

One of the main issues teams face when batting first is that it is so tough to quantify what a 'par' score is.  Certainly, using historical ground data is fundamentally flawed, given that the match-ups between historical matches and the upcoming one are frequently far from the same.

For example, in this particular match, Gujarat's 'par' score should be significantly higher than the mean ground first innings score, given the lack of quality bowling options available in their XI for the match - a statement which I will quantify later in this article.  This lack of bowling quality was also demonstrated by Kolkata being able to chase down 185 in fewer than 15 overs.

In the second match, against Sunrisers Hyderabad, Gujarat were restricted to 135 from 20 overs, following an early collapse that saw them reach 57-4 from the first nine overs.  Again, this total was chased down in around 15 overs by their opponents.

Essentially, the question that needs to be answered is why are Gujarat failing in 2017, when they topped the table in 2016?


2017 IPL AUCTION STRATEGY


Gujarat Lions acquired 11 players at auction (they were the most active team in the 2017 IPL auction), and these are listed below (includes expected IPL batting and bowling data - for an explanation of this, click here).

Note: 1 crore = £122,447 approx.  10 lakh = £12,248 approx.
Note: N/A is used when data sample sizes are insufficient.
Note: (Red = Overseas Players)



Batting



Bowling


Player

Approx Fee (£)

IPL Exp Ave

IPL Exp SR

Boundary Hit %

IPL Exp Ave

IPL Exp Econ








Jason Roy

122477

27.75

150.85

22.36

N/A

N/A

Basil Thampi

104105

N/A

N/A

N/A

33.49

8.19

Manpreet Gony

73486

14.74

174.60

24.44

21.10

7.69

Nathu Singh

61239

N/A

N/A

N/A

19.95

6.61

Munaf Patel

36743

N/A

N/A

N/A

123.61

7.37

Chirag Suri

12248

N/A

N/A

N/A

N/A

N/A

Shelley Shaurya

12248

N/A

N/A

N/A

N/A

N/A

Shubham Agarwal

12248

N/A

N/A

8.79

14.44

5.43

Pratham Singh

12248

N/A

N/A

N/A

N/A

N/A

Akshdeep Nath

12248

19.79

117.87

13.86

49.15

10.65

Tejas Baroka

12248

N/A

N/A

N/A

N/A

N/A


On the surface, there is a fair bit to like about the Gujarat auction purchases. 

Jason Roy is one of the best T20 opening batsmen in the world, while at a lower cost, Manpreet Gony, Nathu Singh and Shubham Agarwal (from a small sample) look to have solid data, particularly in lower-level Indian T20 events.  

As with any tournament around the world, I can use the expectation average concept to assess how a player would perform in a different league by assessing opposition/league quality, and despite relatively small bowling sample sizes, the likes of Manpreet Gony, Nathu Singh and Shubham Agarwal had such impressive numbers in lower-level cricket that they are likely to translate well to future IPL success.  Singh and Agarwal were also relatively inexpensive, and in their early 20s - the sort of signing which is likely to have positive future expectation.

Munaf Patel seems a strange choice given that he's been unsuccessful in lower-level cricket in recent years and at 33 years of age, is unlikely to improve.  All-rounder Akshdeep Nath is a decade younger, but doesn't have impressive data either.  

10 Lakh purchases Chirag Suri, Shelley Shaurya, Pratham Singh and Tejas Baroka do not have enough data to be able to assess their abilities.

One major area where they did not improve their squad was by recruiting high-quality Indian batsmen, or ones with rich potential, despite the obvious need to do so - demonstrated using analysis later in this article.

For example, data illustrates that any of Aditya Tare, Ankit Bawne (both signed by Delhi Daredevils for 30 and 10 lakh respectively), Ishant Jaggi (10 lakh, Kolkata Knight Riders), Saurabh Tiwary (30 lakh, Mumbai Indians), Tanmay Agarwal (10 lakh, Sunrisers Hyderabad) or the late RCB pick-up, Vishnu Vinod, would all be able to boost Gujarat's squad batting quality and depth despite having very low cost.

To summarise, while I disagree with several Gujarat signings and the lack of signing Indian batsmen, their actual signings were far from horrific and looks to be better than many of their rivals.


SQUAD COMPOSITION - BATTING


Below is our pre-IPL 2017 batting data for Gujarat Lions 2017 IPL Squad (Red = Overseas Players), sorted by their IPL Expectation Average based on the SAA model:-


Batting



Player

IPL Exp Ave

IPL Exp SR

Boundary Hit %





Aaron Finch

33.76

138.29

18.23

Brendon McCullum

27.92

152.57

23.72

Jason Roy

27.75

150.85

22.36

Suresh Raina

27.20

127.92

15.66

Dwayne Smith

27.20

125.39

17.54

Dinesh Karthik

22.06

126.05

15.35

Dwayne Bravo

21.40

130.23

13.62

James Faulkner

20.75

127.52

10.34

Akshdeep Nath

19.79

117.87

13.86

Ravindra Jadeja

17.91

103.63

9.26

Ishan Kishan

15.72

130.96

14.29

Pradeep Sangwan

15.62

99.91

13.00

Manpreet Gony

14.74

174.60

24.44

Jaydev Shah

12.48

83.14

10.71

Andrew Tye

12.34

99.62

11.11

Shivil Kaushik

11.78

97.00

7.69

Dhawal Kulkarni

8.50

113.33

10.00

Praveen Kumar

8.03

91.92

9.38

Shadab Jakati

2.28

42.18

8.00


Immediately we can see that four of the top five players in this list are overseas, and this is a large part of the problem facing Gujarat so far this season, with the Lions shoe-horning all four into their side for both matches.

The added problem with this is that all four players would probably prefer to open the batting, and only two slots are available to do so.   So far, Jason Roy and Brendon McCullum have opened the batting for Gujarat, and these are also the main batsmen with the highest IPL expected strike rate, according to my model (all-rounder Manpreet Gony has higher).

Having openers with a great strike rate is a huge asset to a team.  However, when the rest of the team cannot nearly replicate those strike rates, the inevitable result is a failure to continue the momentum given to a team by openers.  Dwayne Smith, with an expected strike rate of around 125, has been pushed down to six in the order, which seems strange, given that players in that sort of role should be expected to score at strike rates well in excess of 140 - batting in the later stages of innings is a time where teams need to accelerate, not stagnate.  He looks much better suited to the opener role, but faces three statistically better batsmen in Finch, McCullum and Roy as competition.

The second obvious judgement that we can take from the above list of players is the lack of quality Indian batsmen that Gujarat Lions possess.  Suresh Raina and Dinesh Karthik are the only two Lions players with IPL expected averages over 20, and the effect of this dearth of Indian batting talent in the Lions squad is the probable reason why coach Brad Hodge has selected all four overseas batsmen.  18 year old Ishan Kishan could eventually be one of the answers to this problem, and given his numbers at such a young age, his age curve would give him a strong chance of a good future career.  However, his current data suggests that this is unlikely to occur immediately.

Having said this, given the huge value available at the IPL auction due to the highly inefficient recruitment strategies of franchises, there really is no logical reason why Gujarat, or any IPL franchise for that matter, could not have acquired several better Indian batsmen, for inexpensive sums, as mentioned previously.

The final area worth assessing regarding Gujarat's batting is that there is a lack of boundary hitters in the squad.  Only five batsmen have boundary hitting percentages in excess of the 16% I deem to be a solid figure, and again this lack of batting depth will put pressure on the main Gujarat batsmen, as well as give the Lions problems when trying to accelerate an innings in the latter stages.


SQUAD COMPOSITION - BOWLING


Below is our pre-IPL 2017 bowling data for Gujarat Lions 2017 IPL Squad (Red = Overseas Players), sorted by their IPL Expectation Average based on the SAA model:-


Bowling


Player

IPL Exp Ave

IPL Exp Econ




Shubham Agarwal

14.44

5.43

Nathu Singh

19.95

6.61

Dhawal Kulkarni

20.81

7.64

Manpreet Gony

21.10

7.69

Andrew Tye

24.82

8.23

James Faulkner

27.00

8.38

Dwayne Bravo

27.04

8.73

Pradeep Sangwan

32.08

7.29

Ravindra Jadeja

32.14

7.68

Basil Thampi

33.49

8.19

Dwayne Smith

34.30

9.31

Shadab Jakati

40.31

8.09

Suresh Raina

43.54

7.77

Aaron Finch

43.68

8.48

Praveen Kumar

47.98

8.62

Akshdeep Nath

49.15

10.65

Shivil Kaushik

52.78

8.37

Munaf Patel

123.61

7.37


In both matches, Gujarat Lions bowlers have struggled, conceding 324 runs in 30.2 overs - a rate in excess of 10 runs per over.  

Nine bowlers - Praveen Kumar, Dhawal Kulkarni, Manpreet Gony, Shival Kaushik, Dwayne Smith, Shadab Jakati, Suresh Raina, Basil Thampi and Tejas Baroka - have all been used by captain Raina in an attempt to stem the flow of runs, albeit with little success.

Generally speaking, in T20, there is very little need to chop and change a line-up, although obviously this should be pitch and match-up dependent.  Players thrive on rhythm and the physical demands of T20 are not nearly as heavy as playing 50 over or Test cricket.  Using expectation data helps eliminate the process of subjective, and frequently biased, guesswork from coaches and captains.

Of the nine bowlers used so far in the two matches, only Dhawal Kulkarni and Manpreet Gony have solid IPL bowling expectation data - the rest have expected averages in excess of 33, and of those remaining players, only Suresh Raina has expected economy of below 8 runs per over.

Using these bowlers with poor expectation data has created what was essentially a self-fulfilling prophecy for Gujarat.  Essentially, given their poor expectation data and inconsistent selection, it was always extremely probable that the Lions would struggle with the ball.  Without addressing this, they are likely to continue to do so.

While the top two bowlers in the list, Shubham Agarwal and Nathu Singh, have relatively small data sizes, they are young and have huge upside.  Picking them over bowlers with far worse expectation data, and giving them more than one match to produce performances really makes a lot of sense.

Two overseas bowlers, James Faulkner and Andrew Tye, would add quality to the bowling attack, but have been ignored so far due to Gujarat playing four overseas batsmen. Tye may continue to be unlucky by being left on the sidelines - his numbers are good but with the lack of batsmen in the Gujarat squad, countryman Faulkner's superior batting may give him the nod, despite Tye having better bowling data.


INJURIES


All-rounders Dwayne Bravo (hamstring surgery) and Ravindra Jadeja are both currently unavailable due to injury. 

Given the necessity to play numerous overseas batsmen, there is some doubt that Bravo would make the Gujarat first-choice team, and data indicates he is a marginally better batsman than fellow overseas all-rounder James Faulkner, and a slightly worse bowler.  

Regarding Jadeja, coach Brad Hodge has been quoted here saying “No, he can’t be replaced..he is a three-dimensional player…he bats, bowls and fields like a jet. So its very hard to replace him in any cricketing side.”

It is not easy to quantify a player's fielding ability, but data indicates that Jadeja is likely to be over-rated by Hodge.  His bowling is economical, although not as much of a wicket-taking threat, while his batting strike rate and boundary hitting abilities leave much to be desired in this format, given the probable necessity to accelerate an innings in the scenarios that he is likely to bat in.  There is no doubt that Jadeja is a superb player, but numbers would suggest he is one much better suited to Test cricket, as opposed to T20.

If data was to select the Gujarat best XI, Jadeja would probably make the team, but the numbers make him far from irreplaceable.


TEAM SELECTION FOR NEXT MATCH


Gujarat face Rising Pune Supergiant on Friday, April 14, and this article suggested that the four overseas batsman strategy could change.  The SAA model agreed with the necessity to change this, and selected the following XI as the best team for the Lions in this match (batting order should be viewed as flexible according to game state):-

Brendon McCullum
Jason Roy
Aaron Finch
Suresh Raina (c)
Dinesh Karthik (wk)
James Faulkner
Manpreet Gony
Dhawal Kulkarni
Shubham Agarwal
Pradeep Sangnam
Nathu Singh
Comments