Show HN: 1v1 coding game that LLMs struggle with

13•levmiseri•22h ago

This is a game I wish I had as a kid learning programming. The concept of it is fairly similar to other coding games like Screeps, but instead of a complex world with intricate mechanics, Yare is a lot more minimal and approachable with quick 1v1 <3 min matches.

It's purely a passion project with no monetization aspirations. And it's open source: https://github.com/riesvile/yare

The first version 'launched' several years ago and I got some good feedback here: https://news.ycombinator.com/item?id=27365961 that I iterated on.

The latest overhaul is a result of simplifying everything while still keeping the skill ceiling high. And at least the LLMs seem to struggle with this challenge for now (I run a small tournament between major models - results and details here: https://yare.io/ai-arena

I'd love to hear your thoughts

Comments

javadhu•18h ago

Cool project, this is my first time seeing such project using LLMs. Took me a while to understand what's happening on the home page.

A question though, why such powerful bots like Gemini 3.1 failed against Clowder bot? Is it because of inefficient code or the LLMs did not handle edge cases? Or they are not as good as humans when it comes to strategy.

levmiseri•18h ago

I’m not sure honestly. It could be some combination of bad spatial reasoning of the LLMs and lack of any training data for this specific challenge.

You can see replays for all of the matches if you hover over the cells in the table.

dang•3h ago

Macroexpanding the previous threads:

Show HN: Yare 2 – Programmable RTS game - https://news.ycombinator.com/item?id=32394902 - Aug 2022 (26 comments)

Show HN: Yare.io – game where you control units with JavaScript - https://news.ycombinator.com/item?id=27365961 - June 2021 (64 comments)

(Btw, reposts are fine after a year or so; links to past threads are just to satisfy extra-curious readers!)

vessenes•2h ago

Cool!

From the prompt it looks like you don’t give the llms a harness to step through games or simulate - is that correct? If so I’d suggest it’s not a level playing field vs human written bots - if the humans are allowed to watch some games that is.

levmiseri•1h ago

That’s true, I’m trying to figure out a better testing environment with a feedback loop.

I did try letting the models iterate on the bot code based on a summary of an end-of-game ‘report’, but that showed only marginal improvements vs. zero-shot

Plasma Bigscreen – 10-foot interface for KDE plasma

UUID package coming to Go standard library

this css proves me human

Can a wealthy family change the course of a deadly brain disease?

Maybe There's a Pattern Here?

LLMs work best when the user defines their acceptance criteria first

C# strings silently kill your SQL Server indexes in Dapper

Galileo's handwritten notes found in ancient astronomy text

Hardening Firefox with Anthropic's Red Team

Querying 3B Vectors

Show HN: Moongate – Ultima Online server emulator in .NET 10 with Lua scripting

Tell HN: I'm 60 years old. Claude Code has ignited a passion again

The Shady World of IP Leasing

Show HN: 1v1 coding game that LLMs struggle with

Launch HN: Palus Finance (YC W26): Better yields on idle cash for startups, SMBs

Tech employment now significantly worse than the 2008 or 2020 recessions

CT Scans of Health Wearables

Show HN: Kula – Lightweight, self-contained Linux server monitoring tool

What canceled my Go context?

Entomologists use a particle accelerator to image ants at scale

Ada 2022

A Modular Robot Dashboard

A tool that removes censorship from open-weight LLMs

Workers who love ‘synergizing paradigms’ might be bad at their jobs

Good Bad ISPs

Analytic Fog Rendering with Volumetric Primitives (2025)

Astra: An open-source observatory control software

Art Bits from HyperCard

Multifactor (YC F25) Is Hiring an Engineering Lead

Game about Data of America

Show HN: 1v1 coding game that LLMs struggle with

Comments

Plasma Bigscreen – 10-foot interface for KDE plasma

UUID package coming to Go standard library

this css proves me human

Can a wealthy family change the course of a deadly brain disease?

Maybe There's a Pattern Here?

LLMs work best when the user defines their acceptance criteria first

C# strings silently kill your SQL Server indexes in Dapper

Galileo's handwritten notes found in ancient astronomy text

Hardening Firefox with Anthropic's Red Team

Querying 3B Vectors

Show HN: Moongate – Ultima Online server emulator in .NET 10 with Lua scripting

Tell HN: I'm 60 years old. Claude Code has ignited a passion again

The Shady World of IP Leasing

Show HN: 1v1 coding game that LLMs struggle with

Launch HN: Palus Finance (YC W26): Better yields on idle cash for startups, SMBs

Tech employment now significantly worse than the 2008 or 2020 recessions

CT Scans of Health Wearables

Show HN: Kula – Lightweight, self-contained Linux server monitoring tool

What canceled my Go context?

Entomologists use a particle accelerator to image ants at scale

Ada 2022

A Modular Robot Dashboard

A tool that removes censorship from open-weight LLMs

Workers who love ‘synergizing paradigms’ might be bad at their jobs

Good Bad ISPs

Analytic Fog Rendering with Volumetric Primitives (2025)

Astra: An open-source observatory control software

Art Bits from HyperCard

Multifactor (YC F25) Is Hiring an Engineering Lead

Game about Data of America