Scribd
Upload
100 views4 pages

Advanced Business Analytics

The document summarizes a soccer result prediction model that was built using four key variables from past Premier League match data: home and away team's goal differentials and points from the previous season. The model achieved around 56% accuracy in predicting matches correctly based on these variables when tested on new data, with away goal differential being the most important predictor. However, the model has limitations as it does not account for situational or subjective factors that could impact matches. Adding more advanced metrics could help improve the model's accuracy.

Uploaded by

Rishabh Shukla
Copyright
© Public Domain
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
100 views4 pages

Advanced Business Analytics

The document summarizes a soccer result prediction model that was built using four key variables from past Premier League match data: home and away team's goal differentials and points from the previous season. The model achieved around 56% accuracy in predicting matches correctly based on these variables when tested on new data, with away goal differential being the most important predictor. However, the model has limitations as it does not account for situational or subjective factors that could impact matches. Adding more advanced metrics could help improve the model's accuracy.

Uploaded by

Rishabh Shukla
Copyright
© Public Domain
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

Soccer Result Prediction Model

ABA Project MyM


Group - 20
FT173071 Rishabh Shukla
FT173104 Yash Misra
FT174034 Jayshree Bhatt
FT174052 Piyush Bhise
FT174104 Vishesh Agrawal
Soccer Result Prediction Model An Overview

What exactly is this model?

It is a system that can identify unbiased reference points with which we


can determine the probability for all the outcomes in a particular game.

This model is supposed to be data driven and objective in nature, it


would overlook the sentiments of the market and try to beat the
bookies.

"The current estimations, which include both the illegal markets and the
legal markets, suggest the sports match-betting industry is worth
anywhere between $700bn and $1tn (435bn to 625bn) a year," says
Darren Small, director of integrity at betting and sports data analysts
Sportradar.

What are we doing to predict the result?

The model that we are building will showcase basically four variables.
Though the model can be made more accurate by adding more
variables:

Home teams goal differential up to that match in a particular


season.
Away teams goal differential up to that match in a particular
season.
Home teams total points from the previous season.
Away teams total points from the previous season.

We have used match results data from past 15 seasons of the well-
known English Premier League to calculate the above variables.
Using these variables we will try and predict the outcome of a given
match.
Results

Train Data
Actual Classification
Prediction A D H Accuracy
A 602 217 323 52.71%
D 282 236 537 22.37%
H 215 223 1475 77.10%
56.28% Total
Test Data 4 Variable
Actual Classification
Prediction A D H Accuracy
A 277 108 92 58.07%
D 88 109 104 36.21%
H 143 235 604 61.51%
56.25% Total

MeanDecreaseGini
Away Goal Differential 632.3599
Home Goal Differential 579.6603
Away Team Points Last Season 547.2345
Home Team Points Last Season 533.9064

According to the results, we found out that the model showed approx.
60% accuracy in predicting away and home wins according on our test
data, it didnt do well in predicting a draw with just about 35% accuracy.
The overall accuracy of the model was about 56% which is pretty good
as we had very few variables. We could also see that the variables were
significant in predicting the outcomes, but Away Team Goal Differential
was the most important among them.

We also benchmarked this model against the odds given by major


betting websites like Bet356, Interwetten, Ladbrokes, and William Hill,
etc. We found out that our model was nearly as accurate as them.

The Limitations of Our Model

Situational factors such as club circumstances, game status etc. and


subjective evaluation of the change of each team during the transfer
window are completely ignored.

In this case, it means the huge x-factor of Manchester Uniteds first


Premier League game with new manager Jose Mourinho is entirely
ignored.

Correlations are also ignored; such as the widely recognized pitch affect
that shows certain matches have a tendency to be either high or low
scoring.

We can later try and incorporate important metrics such as possession,


transfer spends in the last transfer window, chance creation and chance
conversion to make the model more accurate and beat the bookmakers
to make money.

You might also like