Skip to content

John Lees-Miller personal blog

The Mathematics of 2048: Optimal Play with Markov Decision Processes

jdlm.info published 2018-03-1824 min read

Finding provably optimal strategies for 2048 using Markov Decision Processes

Read on jdlm.info → Discuss on HN (43) Archived copy

Points: 312
HN comments: 43
Submissions: 1
Cited by: 0 commenters

Math & Algorithms Games & Graphics 2018

On Hacker News

Submitted	Title on HN	Points	Comments
Apr 9, 2018	The Mathematics of 2048: Optimal Play with Markov Decision Processes by jonbaer	312	43

More from John Lees-Miller

All 7 posts →

How a Technical Co-Founder Spends His Time

2016245 points60 comments Career & Management details

The Mathematics of 2048: Optimal Play with Markov Decision Processes (2018)

2021209 points50 comments Math & Algorithms details

The Mathematics of 2048: Counting States with Combinatorics

2017159 points46 comments Math & Algorithms details

The Mathematics of 2048: Counting States by Exhaustive Enumeration

2017150 points22 comments Math & Algorithms details

Lessons from Building a Node App in Docker

2016134 points28 comments Infrastructure & DevOps details

Lessons from Building Node Apps in Docker

2019107 points30 comments Infrastructure & DevOps details

Related posts from the same era

Announcing the first SHA-1 collision

Google Security Blog20173,030 points485 commentscited by 13 details

Going Critical

Kevin Simler2019989 points81 commentscited by 8 details

A Programmer's Introduction to Mathematics

Jeremy Kun2018937 points214 comments details

Calculus for Beginners and Artists (2003)

Daniel Kleitman2019836 points104 comments details

I found two identical packs of Skittles among 468 packs

Possibly Wrong2019826 points207 comments details

Deep image prior 'learns' on just one image

Dmitry Ulyanov2017782 points223 commentscited by 4 details