Member-only story

I Built An AI Human-Level Game Player

Old-school game trees can be incredibly effective.

Rafe Brena, Ph.D.
TDS Archive
11 min readOct 15, 2024

--

Photo by Jay Bhadreshwara on Unsplash

A while ago, I built the decision-making part of a human-level automatic player for a game.

The automated player for a board game was so strong that nobody I knew (including myself) was able to consistently beat it, so much so that I had to make it dumber in a controlled way to make it enjoyable for casual players.

In this post, I explain how I programmed the “brain” of this little game using standard techniques from the AI playbook before Deep Learning. You’ll learn (in case you are not already familiar with it) how to build Minimax trees and how to develop heuristics for board evaluation. The Minimax technique can apply to any adversarial game regardless of the specifics of the game I describe here.

Enter Kromate

When mobile devices started to become popular (around 2008, after the launch of the iPhone), there were few available board games. And some of the games were not really playable on the tiny screens of the time. That was, of course, the case of chess: you can’t really display a chessboard in a 3.5-inch display with a resolution of 320 x 480 pixels. You can, but it’s going to strain your eyes.

--

--

TDS Archive
TDS Archive

Published in TDS Archive

An archive of data science, data analytics, data engineering, machine learning, and artificial intelligence writing from the former Towards Data Science Medium publication.

Rafe Brena, Ph.D.
Rafe Brena, Ph.D.

Written by Rafe Brena, Ph.D.

AI expert, writepreneur, and futurologist. I was in AI way before it became cool.

Responses (1)