# Classification of fielders in nippon professional baseball using a Gaussian mixture clustering model

## TL;DR

This paper uses advanced statistical clustering to classify Japanese professional baseball fielders based on their hitting performance data.

## Contribution

A Gaussian mixture clustering model is applied to identify meaningful player groupings for strategic team decisions.

## Key findings

- 115 hitting performance indices were categorized into eight meaningful groups.
- 72 high-performing players were analyzed using a WAR threshold of 1.0.
- The framework aims to support substitution and trade strategies in NPB.

## Abstract

Summary: This study proposes a novel analytical framework for categorizing Japanese professional baseball players based on comprehensive hitting performance data. Our primary goal is to identify player groupings that may inform decision-making related to substitution and trade strategies within teams. The dataset used in this analysis was provided by DELTA Corporation, a Japanese firm specializing in advanced baseball analytics. It includes 115 distinct hitting-related performance indices for 327 fielders who participated in official Nippon Professional Baseball (NPB) games during the 2020 season. To make the analysis more structured, we first organized these 115 indices into eight meaningful categories, following the classification methodology defined by DELTA. These categories represent various aspects of hitting performance, such as plate discipline, power, contact ability, and situational hitting, among others. To focus on players with a significant level of contribution, we filtered the original sample and selected 72 players who recorded a “Wins Above Replacement” (WAR) value of 1.0 or higher during the season.