Estimating the Handicap Effect in the Go Game: A Regression Discontinuity Design Approach

TL;DR

This paper estimates the impact of handicaps in the game of Go using a regression discontinuity design, revealing that additional handicaps significantly alter game odds, with effects varying by handicap level.

Contribution

It introduces a novel application of regression discontinuity to quantify the handicap effect in Go, leveraging a unique assignment rule and archival data.

Findings

An extra handicap changes game odds by about 30 percentage points.

The effect of handicaps varies across different levels.

The study provides empirical evidence on handicap effectiveness in Go.

Abstract

This paper provides an estimate for the handicap effect in the go game, a board game widely played in Asia and other parts of the world. The estimation utilizes a unique handicap assignment rule of the game, where the amount of handicaps changes discontinuously with the players' strengths. A dataset suitable for this estimation strategy is collected from game archives of an online platform. The result implies that an additional handicap typically changes the game odds by about 30 percent points, while the impact varies across the handicap level.