In the study carried out, it has been verified that, based on “the 20 patterns most repeated by each player, we calculate the probability that these patterns end up being winners (EPV). By comparing the EPVs of each player with respect to their opponent or the most played patterns of the top 20 players, we can know the advantage of using some patterns over others in a match”. Image: Master1305 on Freepik.
(17-2-2023). A group of researchers from the UPM and the URJC have analyzed game data from 93 badminton tournaments to increase the chances of players earning points in matches based on their patterns of hitting the shuttlecock.
Knowing the hitting patterns in a rally (players exchanging blows until the end of a point) makes a badminton player more likely to win the next point. This is the conclusion of the study carried out by researchers Javier Galeano (ETSI Agronomy, Food and Biosystems) and Miguel Ángel Gómez (Faculty of Physical Activity and Sports Sciences) of the Polytechnic University of Madrid; Javier M. Buldú from the Rey Juan Carlos University, together with the coach of the French Badminton Federation, Fernando Rivas (also coach of the Spanish player Carolina Marín).
The researchers have analyzed the game data of 93 badminton player tournaments between the years 2015 and 2020, implementing the mathematical methodology known as “Markov chains”. In probability theory, a special type of stochastic process (subjected to chance) is known as a chain or Markov model in which the probability of an event occurring depends only on the immediately preceding event.
Hitting Patterns
According to this analysis, the hitting patterns that provide more chances of winning a point in a badminton match for each player have been detected.
To carry out the study, first the hitting sequences of 3 consecutive shots of each player have been calculated. To calculate the sequences, the badminton court is divided into 12 zones and the sequences are recorded. For example, if a player hits from position 12 sending the shuttle to the opponent’s position 5 and returns it to position 9, the sequence 12-5-9 is recorded. With these sequences for each player, it is possible to obtain the transition matrix that measures the probability of passing from one sequence to any other, also taking into account whether these transitions ended in points won or lost. Using this Markov matrix, the researchers have been able to calculate the probability that this pattern is a winning pattern, obtaining a parameter that they have called the Expected Pattern Value (EPV).
These EPV values can be compared between players to find out their chances of winning.
The key to winning game patterns
The study, which has led to an article published in the journal Chaos, Solitons & Fractalsthe number one magazine in the world ranking of the Journal Citation Reports in mathematical applications, has revealed “that players have preferences for performing certain hitting patterns, these being different between players”, explains Javier Galeano.
“On the 20 most repeated patterns for each player, we calculate the probability that these patterns end up being winners (EPV). By comparing the EPVs of each player with respect to her opponent or the most played patterns of the top 20 players, we can find out the advantage of using some patterns over others in a match”, concludes the researcher.
In addition, “it is possible to analyze each of the rallies separately and as the blows are made, being able to identify rallies that should have been won or lost, providing the players and their technical teams with a very useful tool to prepare and analyze matches”, adds Javier Buldú, from the Rey Juan Carlos University.
Although this methodology has a direct application in badminton, it could also be transferred to other racket sports, such as paddle tennis, tennis or table tennis.
The type of analysis proposed in this article could help coaches and professional badminton players to prepare their training sessions and be more efficient in their hitting patterns depending on the specific context of the game and their opponent. Study Information: Using Markov chains to identify player’s performance in badminton, J. Galeano, MA Gómez, F. Rivas and JM Buldú. Chaos, Solitons and Fractals, 165 (2022) 112828.
More information in www.upm.es
