What I learned about startups by beating a roguelike

May 05, 2020

A couple of years ago, I spent several weeks playing a turn-based RPG called Dungeon Crawl Stone Soup.

It’s one of the most classic examples of a roguelike, a genre of turn-based games that feature randomly generated levels, permanent character death, and extremely complicated strategy. They’re also notoriously difficult.

cover image Dungeon Crawl Stone Soup

Even by those standards, Dungeon Crawl Stone Soup was HARD.

Most players take years (no, really) to learn and beat DCSS, if they ever manage it, because of how many possible unique situations and puzzles there are.

Many situations require prior meta knowledge to survive. Others, the ability to predict ramifications of a choice far into the future. At a guess, 80% of the wins in the game are made by less than 20% of its players.

Starting to sound familiar?

At a high level, the experience of playing a roguelike maps well to launching a startup. They both share an extremely high learning curve, fatal mistakes which can cost you everything, deep knowledge of tactics and strategies, and an obsession with the grind. Stay with me.

I love roguelikes. I’d played many of them, beaten a couple, and even made several of my own. But Dungeon Crawl Stone Soup was something else. I routinely lost my characters ten, fifteen, or twenty minutes into the game (games often last up to 12 hours of play – and that’s for experienced players) to situations that seemed completely unpredictable and/or unavoidable.

It wasn’t long before I was up to my ears in wikis and tutorials, drinking from a firehose of information publicly available about the game online. I upped my iteration speed, killing off characters by the dozens in a flow-state of exhilaration and frustration.

After days of losing marked only by tiny iterative increases character life expectency, something happened that got me nervous: a character didn’t die right away. I’d made it maybe 20% through the game when I stopped playing for a little while and just thought.

This was the farthest I’d ever gotten by far, and I didn’t want to lose this character to a stupid mistake.

So I jumped on a Discord channel for the game and shared what I was doing.

Somebody told me I could live-stream my play through an easy in-game system. I set up the stream. Within minutes, several experienced players were watching my screen, giving me advice as I played. Hours later, they’d left, but other folks had jumped on to replace them.

Two days later, I beat the game for the first time.

I was stunned. I hadn’t expected to beat the game for months. And when I later tried to replicate the feat on my own, I failed miserably. After all, even though I learned a lot, I was still missing out on a lot of that deep knowledge and strategy that took my character over the metaphysical finish line.

It was the help of experts who took time out of their lives to assist me in what I was doing – for no return whatsoever – that enabled me to achieve that win.


How does this apply to startups? Well, there are people are out there that can and want to help you succeed, with no exterior motivation of their own. More, if you’re a first-time founder, you NEED their help.

And to get it, all you have to do is let them know you exist, while being transparent about where you’re at right now. The best way to do that? Work in public and document your efforts as close to real-time as possible. Be humble and authentic. Listen. Ask for advice when you’re stuck. And don’t do anything stupid without talking to an expert first!

P.S. I’m going to be applying this lesson directly as I head into validating my next SaaS idea, plus hopefully turning it into a proof-of-concept app. You can follow my journey on Twitter if that interests you.