[Paper Review] Predicting Outcome of Indian Premier League (IPL) Matches Using Machine Learning

Rabindra Lamsal, Ayesha Choudhary|arXiv (Cornell University)|Sep 26, 2018

Sports Analytics and Performance19 citations

TL;DR

This paper proposes a machine learning framework to predict outcomes of Indian Premier League (IPL) Twenty20 matches using player performance metrics and team strength scores. It models seven key factors influencing match outcomes, trains six models on historical data, and achieves 71.66% accuracy using a Multilayer Perceptron, outperforming other models in predicting 2018 IPL match results 15 minutes before gameplay.

ABSTRACT

Cricket, especially the Twenty20 format, has maximum uncertainty, where a single over can completely change the momentum of the game. With millions of people following the Indian Premier League (IPL), developing a model for predicting the outcome of its matches is a real-world problem. A cricket match depends upon various factors, and in this work, the factors which significantly influence the outcome of a Twenty20 cricket match are identified. Each player's performance in the field is considered to find out the overall weight (relative strength) of the teams. A multivariate regression based solution is proposed to calculate points for each player in the league and the overall weight of a team is computed based on the past performance of the players who have appeared most for the team. Finally, a dataset is modeled based on the identified seven factors which influence the outcome of an IPL match. Six machine learning models were trained and used for predicting the outcome of each 2018 IPL match, 15 minutes before the gameplay, immediately after the toss. Three of the trained models were seen to be correctly predicting more than 40 matches, with Multilayer Perceptron outperforming all other models with an impressive accuracy of 71.66%.

Motivation & Objective

• To develop a predictive model for IPL Twenty20 match outcomes using machine learning.
• To identify and quantify the seven most influential factors affecting match results in IPL.
• To compute team strength based on player performance and historical participation.
• To evaluate multiple machine learning models for predictive accuracy in real-time match prediction.
• To achieve reliable pre-match predictions 15 minutes after the toss using historical data.

Proposed method

• A multivariate regression model computes individual player points based on past performance in the IPL.
• Team strength is calculated as the sum of weighted player points for players with the highest appearances for each team.
• Seven key match-influencing factors are extracted and used to construct a predictive dataset.
• Six machine learning models—including Multilayer Perceptron, Random Forest, and SVM—are trained on 2018 IPL match data.
• Predictions are made 15 minutes after the toss, using real-time team composition and player statistics.
• Model performance is evaluated using accuracy on test data from the 2018 IPL season.

Experimental results

Research questions

• RQ1 Which factors most significantly influence the outcome of an IPL Twenty20 match?
• RQ2 Can player performance metrics be aggregated to form a reliable team strength score?
• RQ3 How accurately can machine learning models predict IPL match outcomes using pre-match data?
• RQ4 Which machine learning model performs best in predicting IPL match results among the evaluated models?
• RQ5 Can predictions made 15 minutes after the toss achieve high accuracy using only historical player and team data?

Key findings

• The Multilayer Perceptron model achieved the highest prediction accuracy of 71.66% among all models tested.
• Three of the six models correctly predicted more than 40 matches in the 2018 IPL season.
• The model was trained and evaluated on data from 15 matches, with predictions made 15 minutes after the toss.
• Player performance and team composition were found to be critical in determining match outcomes.
• The use of multivariate regression to assign player points improved the reliability of team strength estimation.
• The final dataset incorporated seven key factors influencing IPL match outcomes, forming the basis for model training and evaluation.