CRICKET PREDICTOR BY:

NAMAN JAIN(9911103490)

INTRODUCTION

In this work we address the winning factors in the sport of ONE DAY CRICKET.Winning in ODI depends on factors such as scoring as well as physical strength of the teams. But there is scope of further research on these factors and analyzing them. Interesting factors include home game-advantage, day night effect, winning the toss and batting first.

Our main aim is to predict the outcome of the cricket match and assisting the coach in selecting the best possible team.

PROBLEM STAEMENT

Part 1 To predict the result of particular match based on various parameter such as Home advantage, First Bat, Ground Conditions, Toss, Team Combination etc. On the basis of these results, decide the betting rates.Part 2To assist the coach in various areas such as Team Selection, Batting Order, Strategy etc.

PROPOSED SOLUTION

For Part 1We are predicting the outcome of a match by simulating research papers and improving upon them by adding more parameters and also comparing the accuracy in prediction of various techniques. Betting rates will also be decided based on these techniques. Some of the algorithms and techniques:Naive BayesID3For Part 2:We are assisting the coach of the cricket teams in selecting the best possible team using individual player records to determine the best team for a particular match. Also use set matching algorithms such as Gale-Shapley to decide the batting order.

 

ALGORITHMS

We will do so using 2 algorithms :1. Naïve Bayes Algorithm : It can be seen as a way of understanding how

the probability that a theory is true is affected by a new piece of evidence.

2. ID3:- The ID3 algorithm begins with the original set S as the root node. On each iteration of the algorithm, it iterates through every unused attribute of the set S and calculates the entropy H(S) (or information gain IG(A)) of that attribute. It then selects the attribute which has the smallest entropy (or largest information gain) value. The set S is then split by the selected attribute (e.g. age < 50, 50 <= age < 100, age >= 100) to produce subsets of the data.

COACH ASSISTOR

STABLE MARRIAGE PROBLEMStable marriage problem (SMP) is the problem of finding a stable matching between two sets of elements given a set of preferences for each element. A matching is a mapping from the elements of one set to the elements of the other set. A matching is stable whenever it is not the case that both: some given element A of the first matched set prefers some given

element B of the second matched set over the element to which A is already matched, and

B also prefers A over the element to which B is already matched

RESEARCH PAPER 1

Auto-play: A Data Mining Approach to ODI Cricket Simulation and PredictionIn this paper, we build a prediction system that takes in historical match data as well as the instantaneous state of a match, and predicts future match events culminating in a victory or loss. We model the game using a subset of match parameters, using a combination of linear regression and nearest-neighbour clustering algorithms

RESEARCH PAPER 2

Fuzzy Logic based Cricket Player Performance EvaluatorWe propose a fuzzy logic based technique to evaluate the performance of cricket players. Various input parameters are being considered which are scaled using linguistic variables and a very simple yet effective software tool is developed to compute the effect of input parameters on the ranking of the players.

RESEARCH PAPER 3

Gale-Shapley Stable Marriage,Problem Revisited: Strategic Issues and ApplicationsWe study strategic issues in the Gale-Shapley stable marriage model. In the first part of the paper, we derive the optimal cheating strategy and show that it is not always possible for a woman to recover her women-optimal stable partner from the men-optimal stable matching mechanism when she can only cheat by permuting her preferences. In fact, we show, using simulation, that the chances that a woman can benefit from cheating are slim

LITERATURE SUMMARY

From our literature survey, it was found that very limited machine learning work has been done on game of cricket. Though cricket shares some attributes with other sports such as baseball, it still remains unique in certain respects and deserves to be analyzed independently. Most of analyzing studies on cricket so far have been conducted using statistical methods. Furthermore, many of them have addressed the five day long test matches but not the One Day Internationals. We present some relevant studies below.

The statistical research on Cricket has been started very early stage of the cricket. In 1945 Wood used the geometric distribution to model the total score in cricket . This was not a study on the ODI form of the game but has been recognized among the pioneering research in the game of cricket. Bailey and Clarke conducted a study to predict the outcome in one day international cricket while the game is in progress .

SUMMARY(cont)

This study was performed using statistical models. The interesting fact about this article is that the authors have statistically proved how the match resources (number of overs and batsmen left) affect the final result. However, they deal with analysis during the current game. They do not predict in advance the chances of winning a new game based on previous matches.

Chedzoy studied the issue of umpiring errors in cricket matches . An umpire is the term used for a referee in cricket. This article focuses on umpiring decisions and how they affect the outcome of the match. This study was also based on a statistical approach. Moreover, it focused only on one aspect of the game, namely, the effect of umpires. Sparks and Abrahamson developed a mathematical model to predict award winners in a game.

SUMMARY(cont)

This study was conducted using machine learning techniques and focused on baseball matches. They have employed this model in the national league and correctly predicted winners prior to the award announcement. Smith and Lipscomb have done similar study for Predicting CY Young award winners for Baseball Pitchers. Interestingly, they have found that Naïve Bayes performed well in their research as well. Bandulasiri has written an interesting article on predicting the winner in an ODI cricket match. This article addressed similar datasets as we used in our research. In this paper, the author has used statistical methods to find wining factors for an ODI match. We have explored the machine learning path, considering popular classifiers, and developed a software tool based on our results. This AI-based tool would be very helpful in predictive analysis in cricket.

SUMMARY(cont)

 It is also possible to apply the machine learning techniques we used in our research to predict the outcome in other outdoor sports suchas baseball. The specific approach used may depend on the nature of the given datasets and applications

GALE SHAPLEY

In this approach, 2 11*11 matrix are considers1: Preference list of Players2.Preference list of Coach

GALE SHAPLEY(cont)

THANK YOU