Application of Individual (I) and Moving Range (MR) Control

charts in analyzing the performance of cricket scores

Muhammad Daniyal

Department of Statistics, the Islamia University of Bahawalpur.

Email: [email protected]

Tahir Nawaz

Department of Statistics, the Islamia University of Bahawalpur.

Email: [email protected].

Iqra mubeen

Department of statistics, the University of Sargodha

[email protected]

M. Aleem

Department of Statistics, the Islamia University of Bahawalpur.

Email: [email protected].
Abstract.

Today’s in cricket, the performance of the teams is being analyzed by using very simple

statistical tools and graphics. Most of the times, the average scores, strike rates, coefficient of

variation and graphical measures are being utilized to measure the performance and to make

comparisons between different teams. The main focus of this research paper is to provide the

application of statistical quality control charts in analyzing the performance of the team by using

its batting scores. For this purpose, we analyzed the scores of Pakistan cricket team. We used

individual and moving range control charts because it may be applied in normal conditions

otherwise false special cause signals can be generated. We then used the basic sensitizing rules to

check whether the un-natural or non random variations are present in the batting scores or not.

We will provide the specialists of this game an opportunity to check for the reasons for the un-

natural and non random behavior in the batting scores so that in future the performance can be

improved.

Keywords: scores, performance, moving range control charts, individual control charts,

graphical displays, assignable cause, and normality.

1 Introduction

Cricket has become one of the world class famous and popular games. It is the game of

bat and ball which includes eleven players. Most of analysis is being conducted to measure the

performance of the teams and individual players from all perspectives. Different sites are

available which include the complete information and time to time record of every event of the

cricket. The experts of this game give their opinions in the light of this information. As far as the

ODI career of Pakistan team is concerned, it started its one day career in 1972/1973. It is ranked

at 6th position according to ICC criteria in one day with 2620 points and 105

rating(http:/ / w w w .espncricinfo.com/ rankings/ content/ page/ 211271.html).

very simple statistical tools. During the few years back, a huge study and work has been

conducted to measure the performance of the teams and the prediction of their performance in

the next series. The utilization of other sources are given in Duckworth/Lewis approach

(Duckworth and Lewis, 2002) and its references, (Johnston et al. 1993), (Beaudoin and Swartz,

2003) and (de Silva et al., 2001). The batting orders with optimality are mentioned in (Swartz et

al., 2006) and in (Normanand Clarke, 2010), different batting strategies using the dynamic

programming (Preston and Thomas, 2000) and (Johnston et al. 1993), what kind of impact will

have winning the toss first are discussed in the study of (de Silva and Swartz, 1997). The

methods of prediction and forecasting can be found in the study of (Cohen, 2002), (Gilfillan and

Nobandla, 2000) and (Swartz et al., 2009). Graphical representation of the results are mentioned

in the research study of (Kimber, 1993), (Barr et al., 2008), (Bracewell and Ruggiero, 2009) and

(van Staden, 2009). The performance of batting is keenly measured by the method of batting

average scores and various scholars and researchers have used strike rate and the average score

of batsman as a measure of performance e.g (Croucher, 2000) and (Barr and Kantor, 2004).Bar

and Kantor used a new idea of graphical representation of the results with the strike rate on one

axis and the probability of getting out on the other axis. In this two dimensional structure, they

created the criteria for the selection of the batsman which combines together the average and

strike rate (Basevi and Binoy, 2007) and (Barr et al., 2008). (Barrand van den Honert, 1998)

gave a measure which is based on both averages and consistencies. (Lemmer, 2008a) concluded

in his research study that the batting average can not be reasonable or satisfactory in the case of

a small amount of scores if the player had a large percentage of not out scores.

The present research gives the application of statistical quality control charts as compared

to the simple statistical measures for measuring the batting performance of Pakistan team. It will

help us to see the non random pattern and un-natural variation in the scores of Pakistan team

which will help the experts to make some decisions. These decisions will help us in future to
improve the batting performance of Pakistan team and avoid un-natural situations due to

assignable causes.

2. Methodology

The main theme of this research study is to see the application of statistical control charts

in evaluating the batting performance of Pakistan cricket team by using its ODI scores from 11

Feb 1973 to 22 Mar 2012. Pakistan is ranked at 6th position according to ICC criteria in one day

with 2620 points and 105 rating. We are interested in evaluating the performance of Pakistan

cricket team by applying quality control charts on its batting scores. The individual and moving

average control charts were used for this purpose. The individual and moving range control

charts can be utilized when the data follows the normal distribution otherwise the false special

cause signals will be generated.

(http:/ / w w w .smartersolutions.com/ blog/ forrestbreyfogle/ 2009/ 12/ 08/ non-normal-data-and-

the-creation-of-their-control-charts-and-predictive-statements/ ).

3. Objective

The basic purpose of this research study is to evaluate the performance of Pakistan cricket team

using the statistical control charts. The appropriate control charts which will be used for this

analysis is the individual and moving range control charts. The sensitizing rules will be used to

check which score is indicating an out of control signal or un-natural variation. In the end, we

will achieve following objectives.

1. We will develop the trial control limits, and analyze if there is any out of control point.

We will look for the assignable cause for that out of control signal. If all the scores are

found in control, the trial limits will be utilized as tolerance limits for the future analysis.

2. If any out of control signal is observed, we will remove that score and analyze the scores

again using control structure.

As it has already been mentioned that individual and moving range control charts can be utilized

when there is normality of data otherwise false special cause signals will be generated. Different

techniques exist to check the normality of the scores. In statistics the Shapiro-Wilk (1965) test

checks the null hypothesis that x1 … xn follows the normal distribution. The test statistic used is:

(1)

If P-value is less than significance level which is normally taken 5%, then we will reject the

hypothesis that the scores follow normal distribution. The SPSS version 17 was utilized for this

purpose and the table 1 was obtained

Since the p value is greater than the significance level 0.05, therefore we can conclude that the

scores follow the normal distribution. The graphical technique which is being utilized to measure

the normality of the scores is the normal probability plot. The scores are plotted against a

theoretical normal distribution in such a manner that the points plotted show the straight line

approximately. If the points deviate from the straight line, this is the indication that the scores are

non-normal. The normal probability plot is a special case of the probability plot in the case of a

normal distribution. The normal P-P plot can be utilized to check the normality of the scores. The

following figure is showing the P-P plot of the batting scores. It can be clearly observed from

figure 1 that the scores are following the normal distribution.

The histogram is the most common and the most useful graphical measure used to explore the shape

of the distribution of the scores. If distribution is bell shaped then it clearly means that they follow

the normal distribution. The figure 2 is showing the histogram for the scores of Pakistan cricket

team made against different teams in its whole ODI career.

The individual and moving range control charts were first introduced by the William Shewhart in

1920. These control charts are utilized as the graphical measure to evaluate the performance of

the manufacturing units and in services (Montgomery, 2005). The main use of these charts is to

evaluate the performance and look for any non random and un-natural variations present due to

any assignable cause.

For the purpose of checking assignable causes and variations in the batting performance of the

batsman, individual (I) and moving average (MR) control charts are utilized because these

control charts are applied when the data is in the form of single observation in each time period.

These types of charts are utilized when each unit of measurement is analyzed. The initial step in

the creation of control charts is to set up the trial control limits. Moving range control chart is

used to control the spread and location parameter. The process will be considered under control

once these parameters are controlled. Moving range control limits are made with the help of two

observations and its statistic can be computed as:

MR  xi  xi 1 (2)

The first value in the moving range can not be calculated because no previous record is available

of the scores which have not been made. Therefore the first period value will be left blank and

the MR will be calculated for the second period and so on.

For the individual control chart, the upper, lower control limits and the central line are calculated

by the following statistic

UCL  X  3
MR
(3)
d2

CL  X (4)
 LCL  X  3
MR
(5)
d2

Where X denotes the mean of the observations and MR denotes the mean of the moving ranges.

The term d2 shows the sigma conversation factor for the specific sample size which is taken 2

because we are utilizing two terms xi and xi 1 to construct the 3 sigma limits. So the value of

d2=1.128 (Hines et al., 2004).

The control limits for the MR (moving range) charts can be calculated by the following

UCL  D4 MR (6)

CL  MR (7)

LCL  D3 MR (8)

D3 and D4 are considered as the parameters of the control limits. They depend upon the sample

size which has been mentioned that it will be taken two. The values of D3 and D4 are chosen 0

and 3.267 respectively (Hines et al., 2004). The charts in figure 3 and figure 4 were constructed

using MINITAB 15.

6. Analysis and Conclusions

The next step after the construction of the I and MR charts is to analyze each score or

case. For this purpose, we selected two basic sensitizing rules to check which score is out of

control and which is showing non random pattern. These rules can be found in Western Electric

Company (1956) and Nelson (1984; 1985).

1. The control chart will be out of control if any point lies above or below the 3 sigma

limits.

2. The control chart will be out of control if 2 points out of 3 above or below the 2 sigma

limits.
By considering these rules, we can see that the individual control chart of the scores in figure 3 is

showing seven out of control signals. On the other hand if we look at the figure 4, the moving

range control chart is showing nine points out of control. These out of control points can be

generated due to any assignable cause. We leave this to the specialists and the experts of the

cricket to look for these assignable causes which have made the scores out of control. The

sensitizing rules applied were formulated in the table using SPSS 17 and are mentioned in table

2. The quality control charts indicate that the process which is being monitored is not in control.

If we go into the deep analysis of the chart, we will find different the sources of variation. This

deep analysis is the work of the specialists and the experts of the game who deeply analyze the

game and its techniques. They would be in the better position to look for the assignable causes

which are making these control structures out of control. They would look for the non random

and un-natural pattern which is showing the out of control signal and disturbing the batting

performance of the team. The limits which are constructed are the trial control limits. If we

remove the out of control points again calculate the control structure, it may be possible that

charts are then under control. These limits will be then tolerance limits and can be used for future

analysis.

7. References

[1] Barr, G.D.I. and Kantor, B.S. (2004) A criterion for comparing and selecting batsmen in

limited overs cricket. Journal of the Operational Research Society 55, 1266-1274.

[2] Barr, G.D.I. and van den Honert, R. (1998) Evaluating batsmen’s scores in test cricket.

South African Statistical Journal 32, 169-183.

[3] Barr, G.D.I., Holdsworth, G.C. and Kantor, B.S. (2008) Evaluating performances at the

2007 cricket world cup. South African Statistical Journal 42, 125-142

[4] Basevi, T. and Binoy, G. (2007) The world’s best Twenty20 players. Available from

URL http:/ / content-rsa.cricinfo.com/ columns/ content/ story/ 311962.html .

South African Statistical Journal 37, 203–222.

[6] Bracewell, P.J. and Ruggiero, K. (2009) A parametric control chart for monitoring

individual batting performances in cricket. Journal of Quantitative Analysis in Sports 5,

1-19.

[7] Cohen, G.L. (2002) Cricketing chances. In: Proceedings of the Sixth Australian

Conference on Mathematics and Computers in Sport. Eds: Cohen, G. and Langtry, T.

Sydney University of Technology,NSW. 1-13.

[8] Croucher, J.S. (2000) Player ratings in one-day cricket. In: Proceedings of the Fifth

Australian Conference on Mathematics and Computers in Sport. Eds: Cohen, G. and

Langtry, T. Sydney University of Technology, NSW. 95-106.

[9] De Silva, M.B. and Swartz, T.B. (1997) Winning the coin toss and the home team

advantage in one-day international cricket matches.The New Zealand Statistician 32, 16-

22.

[10] Duckworth, F. and Lewis, T. (2002) Review of the application of the Duckworth/Lewis

method of target resetting in one-day cricket.In: Proceedings of the Sixth Australian

Conference on Mathematics and Computers in Sport. Eds: Cohen, G. and Langtry,

T.Sydney University of Technology, NSW. 127-140.

[11] Gilfillan, C. and Nobandla, N. (2000) Modelling the performance of the South African

national cricket team. South African Journal for Research in Sport, Physical Education

and Recreation 22, 97-110.

[12] Hines, W., Montgomery, D., Goldsman, D., and Borror, C. (2004). Probability and

Statistics in Engineering, 4th Edition, Wiley & Sons, Inc., Hoboken, NJ.
[13] Johnston, M.I., Clarke, S.R. and Noble, D.H. (1993) Assessing player performance in

one-day cricket using dynamic programming.Asia-Pacific Journal of Operational

Research 10, 45-55.

[14] Kimber, A.C. (1993) A graphical display for comparing bowlers in cricket. Teaching

Statistics 15, 84-86.

[15] Montgomery, D. (2005). Introduction to Statistical Quality Control, 5th Edition, Wiley &

Sons, Inc., Hoboken, NJ.

[16] Nelson, Lloyd S. (1984). "The Shewhart Control Chart—Tests for Special Causes,"

Journal of Quality Technology, 16(4), 237 – 239.

[17] Nelson, Lloyd S. (1985). "Interpreting Shewhart Control Charts," Journal of Quality

Technology, 17(2), 114 – 116.

[18] Norman, J.M. and Clarke, S.R. (2010) Optimal batting orders in cricket. Journal of the

Operational Research Society 61, 980-986.

[19] Preston, I. and Thomas, J. (2000) Batting strategy in limited overs cricket. The

Statistician 49, 95-106

[20] Swartz, T.B., Gill, P.S. and Muthukumarana, S. (2009) Modelling and simulation of one-

day cricket. Canadian Journal of Statistics 37, 143-160

[21] Swartz, T.B., Gill, P.S., Beaudoin, D. and de Silva, B.M. (2006) Optimal batting orders

in one-day cricket. Computers and Operations Research 33, 1939-1950.

[22] Van Staden, P.J. (2009) Comparison of cricketers’ bowling and batting performance

using graphical displays. Current Science 96, 764-766.

[23] Western Electric Company (1956), Statistical Quality Control Handbook. (1st ed.),

Indianapolis, IN: Western Electric Co.

matches. South African Statistical Journal 42, 83-105.

[25] (http:/ / w w w .smartersolutions.com/ blog/ forrestbreyfogle/ 2009/ 12/ 08/ non-normal-data-and-the-

creation-of-their-control-charts-and-predictive-statements/ ).