Time management in a chess game through machine learning. Issue 1 (2nd January 2023)

• Record Type: Journal Article
• Authors: Burduli, Guga; Wu, Jie
• Is Part Of: International journal of parallel, emergent and distributed systems. Volume 38:Issue 1(2023)
• Page Start: 14
• Page End: 34
• Publication Date: 2023-01-02
• Subjects: Time management -- neural network -- segmented least square approximation -- chess -- collecting the data; Parallel computers -- Periodicals; Electronic data processing -- Distributed processing -- Periodicals; Computer algorithms -- Periodicals; 004.35
• DOI: 10.1080/17445760.2022.2088746
• Languages: English
• ISSNs: 1744-5760

Abstract

ABSTRACT: Chess includes two significant factors: playing good moves and managing your time optimally. Time, especially in blitz games, is just as essential to the game as making good moves. Nowadays, several incredible engines are already developed, more than enough to defeat all the best human chess players. For studying how to make good moves, these engines are crucially useful. Professional chess players are using them in addition to coaches to prepare for the matches or to examine the mistakes in their played games. However, managing time still is a huge challenge. There are no basic rules for managing time. A lot of factors influence the decision about how much time should be spent in a particular position. For computers, it is easier because they calculate much faster and they have all the theoretical knowledge. However, even grandmaster chess human players are struggling with time trouble. In this article, we describe how the data was collected from an online chess platform and show methods of how time can be managed based on different features. In this regard, we will use two different models: using a customised neural network and using a proposed segmented least square approximation method. In both of the models, we will use our collected data. GRAPHICAL ABSTRACT: UF0001