Probability models for the Philippine Tennis Association tournaments

Five probability models namely, the Bradley and Terry Type model, Uniform-Strength Assumption model, Normal-Strength Assumption model and two logistic regression models (using a linear and non-linear function of the player's strength) were compared to determine the best model that will predict...

Full description

Saved in:
Bibliographic Details
Main Authors: Sianghio, Christina Sheryl L., Vicencio, Aileen R.
Format: text
Language:English
Published: Animo Repository 2000
Online Access:https://animorepository.dlsu.edu.ph/etd_bachelors/17005
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: De La Salle University
Language: English
Description
Summary:Five probability models namely, the Bradley and Terry Type model, Uniform-Strength Assumption model, Normal-Strength Assumption model and two logistic regression models (using a linear and non-linear function of the player's strength) were compared to determine the best model that will predict the champion in a tournament or the winner in a match of the Philippine Tennis Association (PHILTA) seeded single-elimination tournaments having a balanced structure with a (1, 4, 2, 3) draw. If the players are strictly the ones in the top four seeds, then the logistic regression model using a linear function of the player's strength was proven to be the best model for predicting the champion in the tournament whereas no model was seen appropriate for predicting a winner in a match. On the other hand, if the players are not restricted to the top four seeded players but are paired in such a way that the strongest player competes the weakest, and the second strongest competes with the second weakest, then the logistic regression model using a linear function of the player's strength is still the best one for predicting the champion in the tournament while the Uniform-Strength Assumption model is a good model for the prediction of the winner in a match.