Prediction of ICC T20 WCC 2022 Winner
Despite Hockey being the national sport of the country, it is cricket which rules over the hearts of the citizens. It creates a lot of excitement and frenzy amongst the fans of the game. Cricket is like a religion in India and the players are considered to be demi-gods. It is the most-watched sport in India and people even miss their schools and offices when any major international match is happening. Cricket unites Indians like nothing else and from kids to adults; everyone keeps track of the cricket score whenever the Indian team is playing.
ICC T20 World Cup 2022
The 2022 ICC Men’s T20 World Cup is scheduled to be the eighth ICC Men’s T20 World Cup tournament, scheduled to be played in October and November 2022 in Australia.
The twelve teams that reached the Super 12 phase of the 2021 ICC Men’s T20 World Cup automatically qualified for the 2022 tournament. Afghanistan, Australia, Bangladesh, England, India, Pakistan, New Zealand and South Africa all qualified directly for the Super 12 phase of this tournament, based on their performances in the 2021 tournament and their rankings as of 15 November 2021. Namibia, Scotland, Sri Lanka and the West Indies were all placed in the group stage of the competition.
How we are Predicting ?
We are living in the world Artificial Intelligence and Machine Learning. So, everything can be predicted using AI & ML but till certain accuracy. We can’t predict will 100% accuracy . In this prediction , we are using SVM (Support Vector Machine) algorithm for predicting the winner.
Support Vector Machine(SVM)
It is machine learning based supervised learning model with associated learning algorithms that analyze data for classification and regression analysis. SVMs are one of the most robust prediction methods, being based on statistical learning frameworks. SVM maps training examples to points in space so as to maximize the width of the gap between the two categories. New examples are then mapped into that same space and predicted to belong to a category based on which side of the gap they fall.
Data
For this project, we collected the T20 international matches records from 2010 to February 2022. Since , that record will be sufficient for testing and preparing the model. Apart from it , we considered the current T20 ranking of the team in February 2022 and for we have taken the match fixtures for ICC T20 WCC 2022 from their official website.
Previous record : https://cricsheet.org/
Current ranking : https://www.cricbuzz.com/
Match fixtures : https://www.t20worldcup.com/
Tools and Environment
- Jupyter Notebook
- Numpy
- Pandas
- Seaborn
- Scikit-learn
Factors we are considering for prediction
We are considering two important factors for predicting. First, Number of wins in last six matches played before current matches. Obviously, the team which is doing well from last six matches is having more probability of winning. But , there may be a case in which both teams have won equal matches in their last six matches. So, in that case, we are taking the current ranking into the consideration that will help to predict in that case. So, the two factors are :-
- Number of Wins in last six matches from current match.
- Current ranking
Assumptions we made
Since, 12 teams will be participating in this World Cup, 4 among them need to first win the qualifier matches to get eligible for World Cup. So, by analyzing the current performance of teams in qualifier matches ,we have assumed the results of qualifier matches as follows:-
Group A Winner : Sri Lanka
Group A Runner-Up : Namibia
Group B Winner : West Indies
Group B Runner-Up : Scotland
We are predicting the results of qualifier matches because there are many teams which haven’t played enough matches so that we can predict i.e. less data is available for it. That’s why we are assuming their results.
Code
Data cleaning code : https://github.com/HarshDeswal/ICC-T20-world-cup-2022-Prediction/blob/main/Notebooks/Data_Cleaning.ipynb
Modelling code : https://github.com/HarshDeswal/ICC-T20-world-cup-2022-Prediction/blob/main/Model/modelling.ipynb
Accuracy
Since, cricket is totally unpredictable game , game changes even in a single ball . But , through the analysis of previous records and accuracy of 62.14% we have done the prediction using SVM.
#Accuracy score
accuracy = model.score(X_test, y_test)
accuracy0.6214285714285714
Prediction Results
Results till League matches
1.Australia Vs New Zealand : New Zealand
2.England Vs Afghanistan : England
3.Group A Winner Vs Group B Runner Up : Group A Winner
4.India Vs Pakistan : India
5.Bangladesh Vs Group A Runner up : Group A Runner up
6.South Africa Vs Group B Winner : South Africa
7.Australia Vs Group A Winner : Australia
8.England Vs Group B Runner UP : Group B Runner UP
9.New Zealand Vs Afghanistan : New Zealand
10.South Africa Vs Bangladesh : South Africa
11.India Vs Group A Runner Up : India
12.Pakistan Vs Group B Runner Up : Pakistan
13.Afghanistan Vs Group B Runner UP : Group B Runner UP
14.England Vs Australia : England
15.New Zealand Vs Group A Winner : New Zealand
16.Bangladesh Vs Group B Winner : Group B Winner
17.Pakistan Vs Group A Runner Up : Pakistan
18.India Vs South Africa : India
19.Australia Vs Group B Runner Up : Australia
20.Afghanistan Vs Group A winner : Group A winner
21.England Vs New Zealand : England
22.Group B winner Vs Group A Runner Up : Group B winner
23.India Vs Bangladesh : India
24.Pakistan Vs South Africa : Pakistan
25.New Zealand Vs Group B Runner Up : New Zealand
26.Australia Vs Afghanistan : Australia
27.England Vs Group A Runner Up : England
28.South Africa Vs Group A Runner Up : South Africa
29.Pakistan Vs Bangladesh : Pakistan
30.India Vs Group B Winner : IndiaFrom the Modelling the Point table as follows:
Group 1
Team Win
Afghanistan 0
Australia 3
England 4
New Zealand 4
Group A Winner 2
Group B Runner Up 2
Group 2
Team Win
Bangladesh 0
India 5
Pakistan 4
South Africa 3
Group B Winner 2
Group A Runner Up 1Semi Finals
India Vs England : India
New Zealand Vs Pakistan : New ZealandFinal
India Vs New Zealand : India
Probable Winner of World Cup : India
Conclusion
According to this model India is going to Win the World.
This article is written to show how Machine Learning can be used to calculate probabilities in a simulation and does not attempt to actually get the results right since the data used is not enough for it .
#Disclaimer : This article is for educational purpose. We are not providing any recommendation for betting and other fantasy games. So, we are not liable for any financial or economical losses of anyone.
Further Scope
- More features can be used for getting better precision and accuracy.
- More advanced machine learning algorithm can be used.
Team Member
Harsh Deswal : www.linkedin.com/in/harsh-deswal-87400a223
Hari Krishna
Suhash Reddy T
References