A Bradley-Terry type model for forecasting tennis match results

McHale, Ian G.; Morton, Alex S.

doi:10.1016/j.ijforecast.2010.04.004

Cited by 91 publications

(50 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The pointbased models extend the earliest approach to implementing the IID model, where only a player's average ability on serve was utilized (Newton and Keller 2005), using more sophisticated models for the probabilities of winning a point on serve and return. Two models -an application of the Bradley-Terry model (1952) proposed by McHale and Morton (2011) and the FiveThirtyEight model -are examples of paired comparison approaches. Finally, as a standard of reference for predictive performance, the study includes predictions from the bookmaker consensus model of Leitner, Zeileis, and Hornik (2009).…”

Section: Modelsmentioning

confidence: 99%

See 1 more Smart Citation

Searching for the GOAT of tennis win prediction

Kovalchik

2016

Journal of Quantitative Analysis in Sports

View full text Add to dashboard Cite

Sports forecasting models -beyond their interest to bettors -are important resources for sports analysts and coaches. Like the best athletes, the best forecasting models should be rigorously tested and judged by how well their performance holds up against top competitors. Although a number of models have been proposed for predicting match outcomes in professional tennis, their comparative performance is largely unknown. The present paper tests the predictive performance of 11 published forecasting models for predicting the outcomes of 2395 singles matches during the 2014 season of the Association of Tennis Professionals Tour. The evaluated models fall into three categories: regressionbased, point-based, and paired comparison models. Bookmaker predictions were used as a performance benchmark. Using only 1 year of prior performance data, regression models based on player ranking and an Elo approach developed by FiveThirtyEight were the most accurate approaches. The FiveThirtyEight model predictions had an accuracy of 75% for matches of the most highly-ranked players, which was competitive with the bookmakers. The inclusion of careerto-date improved the FiveThirtyEight model predictions for lower-ranked players (from 59% to 64%) but did not change the performance for higher-ranked players. All models were 10-20 percentage points less accurate at predicting match outcomes among lower-ranked players than matches with the top players in the sport. The gap in performance according to player ranking and the simplicity of the information used in Elo ratings highlight directions for further model development that could improve the practical utility and generalizability of forecasting in tennis.

show abstract

Section: Modelsmentioning

confidence: 99%

“…A popular type of paired comparison model is the Bradley-Terry model, which was applied to forecasting outcomes in tennis by McHale and Morton (2011). With this approach, each player has a latent match win ability, α i .…”

Section: Paired Comparisonmentioning

confidence: 99%

Searching for the GOAT of tennis win prediction

Kovalchik

2016

Journal of Quantitative Analysis in Sports

View full text Add to dashboard Cite

show abstract

“…We estimate team strength via paired-comparison (Bradley and Terry, 1952) and margin-of-victory models which have been applied to ranking teams in a variety of sports (McHale and Morton, 2011;Koehler and Ridpath, 1982;Sire and Redner, 2009;Martin, 1999). The growth in information about the relative strength of teams is quantified via a predictive accuracy curve which describes how paired-comparison models fitted from X% of the games in a season predict the outcomes of the remaining (100 − X)% of games, across multiple values of X (games are partitioned into training and test sets at random to reduce the impact of longitudinal trends over the course of a season).…”

Section: Introductionmentioning

confidence: 99%

Who’s good this year? Comparing the information content of games in the four major US sports

Wolfson

Koopmeiners

DiLernia

2018

JSA

View full text Add to dashboard Cite

Abstract. In all four major North American professional sports (baseball, basketball, football, and hockey), the primary aim of the regular season is to identify the best teams. Yet the number of regular season games used to distinguish between the teams differs dramatically between the sports, ranging from 16 (football) to 82 (basketball and hockey) to 162 (baseball). While length of season is partially determined by factors including travel logistics, rest requirements, playoff structure and television contracts, from a statistical perspective the 10-fold difference in the number of games between leagues suggests that some league's games are more "informative" than others. In this paper, we propose a method to quantify the amount of information games yield about the relative strengths of the teams involved. Our strategy is to estimate a predictive accuracy curve which assesses how well simple paired comparison models fitted from X% of games within a season predict the outcomes of the remaining (100 − X)% of games, across multiple values of X. We compare predictive accuracy curves between seasons within each sport and across all sports, and find dramatic differences in the amount of information yielded by game results in the four major North American sports leagues.

show abstract

“…Several studies in the field of sport have reported more accurate forecasts by statistical models compared to the subjective judgment of experts (4,6). Sport researchers used of several methods for forecasting the athletes' performance in various sports including football, tennis, basketball and professional league competitions, collegiate tournament and national, continental and world championships (3,7,8,11,14,15).…”

Section: Introductionmentioning

confidence: 99%

The relationship between the countries performances at the major sport events with their performance at the Olympic Games

BAHRAMI¹

2014

TJSE

View full text Add to dashboard Cite

The purpose of this study was study of Relationship between countries performance at multi-sport events and world championships with their performance at Olympic Game. The population of all the countries in the Olympic Games medals have been won that their number is equal 1166 (excluding repeat). Based on Morgan table, samples were selected equal 321. Information was collected from multiple sites such as international federations, Olympic Continental Councils and international Olympic committee. Also, for analyzing used descriptive and inferential statistics (Kolmogorov-Smirnov test and Spearman correlation coefficient) by SPSS software and significant level p≤0.05. The results showed that there is positive and significantly relationship between the number of medals of gold, silver, bronze and total medals won by countries at former Olympic Games, world championships, Asian and Pan-American Games with the number of Medals of gold, silver, bronze and total medals won at upcoming Olympic Games, However, there is not significantly relationship between the number of medals of gold, silver, bronze and total medals won at all-African Games with the number of medals of gold, silver, bronze and total medals won at Olympic Games.

show abstract

A Bradley-Terry type model for forecasting tennis match results

Cited by 91 publications

References 18 publications

Searching for the GOAT of tennis win prediction

Searching for the GOAT of tennis win prediction

Who’s good this year? Comparing the information content of games in the four major US sports

The relationship between the countries performances at the major sport events with their performance at the Olympic Games

Contact Info

Product

Resources

About