Dear all,

we are glad to announce the following seminar by Dr. Vincenzo Candila (University of Salerno) on Wednesday March 29th 2023 starting from 4:30 p.m. in room 2062 at the Department of Statistics and Quantitative Methods, University of Milano-Bicocca (Milan).

On the tennis winner forecasting methods: from the Weighted Elo rates to the network-based indicators

Vincenzo Candila 

Department of Economics and Statistics

University of Salerno

Abstract:

Statistical methods for predicting the winner in tennis matches have increased in popularity over the last two decades. Currently, several techniques are available, such as the regression-based, point-based, and paired-comparison approaches. Within this latter class, the Elo rating is one of the most popular methods to predict the probability of winning tennis matches. Even with its widely recognized merits in terms of good performance, the Elo method only partially considers each player's current form and recent performances. To overcome this drawback, we have proposed a new version of the Elo rating method, labeled Weighted Elo (WElo), where the standard Elo rates are weighted according to the scoreline of the players’ last match. More recently, we have further enlarged the paired-comparison class of models, introducing the network-based indicators for the first time to allow the ratings of the whole set of players to vary every time there is a new match. This work aims to illustrate the main characteristics of the WElo and network-based rates. In the empirical applications, we show the statistically meaningful benefits of using the proposed approaches with respect to many popular competing specifications.

March 29th, 2023 at 4:30 p.m.

Department of Statistics and Quantitative Methods

Room 2062 - Building U7 (Second floor)

Bicocca degli Arcimboldi street, 8 – 20126 Milan

For further information, please write to rosanna.grassi@unimib.it