How did a game destined for failure become a cult smash? In an article originally written for Game Developer magazine, Edmund McMillen (Super Meat Boy) discusses how he added religion to The Legend of Zelda, mixed it with a roguelike, and came out with a surprise hit.

On paper, there is simply no reason for a game like The Binding of Isaac to have become as huge as it has. It makes no sense -- and this is coming from the person who believed in it the most. I knew Isaac was special, but if you asked me to bet on whether Isaac would sell over one million copies in less than a year, I would have bet against it.

You see, The Binding of Isaac was made to clash against mainstream games -- it was designed to be a niche hit at best. I had hoped it would gain some minor cult status in small circles, kind of like a midnight movie from the 1970s. From any mainstream marketing perspective, I designed Isaac to fail -- and that was my goal from the start.

When I started working on The Binding of Isaac, I was still haunted by the end of Super Meat Boy's development, and the hoops we had to jump through to get there. I wouldn't say Super Meat Boy was "selling out," but it was the closest I was going to come to it when it came to playing by the rules to make sure that we could sell the game that consumed two years of our lives (and all of our money).

After SMB, I no longer had those worries -- I could afford to take a bigger risk and fail, if I felt like failing. I wanted to make something risky and exciting now that the financial aspects of that risk were gone. And I wanted to really push my limits to get back to where I had come from -- a place where there were no boundaries, where I could create anything without worrying about making a profit.

The Binding of Isaac started in a weeklong game jam. Tommy Refenes (Super Meat Boy co-developer) was taking a vacation, so I decided to do the game jam with Florian Himsl, who programmed a few of my previous Flash games (Triachnid, Coil, and Cunt). Florian is the kind of guy who is up for anything; he wasn't worried about his reputation, and was basically down with whatever I wanted to do in terms of content. This was good, because I had two clear goals when I started designing Isaac: I wanted to make a roguelike game using the Legend of Zelda dungeon structure, and I wanted to make a game about my relationship with religion.

Both goals were challenging but very fun to design, and after seven days we had something that was turning into a game. It seemed too good to pass up, so we continued working on it in Flash (using ActionScript 2). At this point in the process, I wasn't thinking about how we were going to sell this game (or if we were going to be able to sell the game at all!); it was just a challenge we both wanted to finish.

We finished The Binding of Isaac after about three months of part-time development. We released it on Steam, and it was selling okay; for the first few weeks, the game was averaging about 100-200 copies a day, eventually stabilizing at about 150 a day after a few months. By this point, the game had already exceeded my expectations, but five months after release something very odd happened. Our daily average started to climb. 200 copies per day turned into 500 copies, then 1,000 copies, and by the seven-month mark Isaac was averaging sales of more than 1,500 copies a day and climbing. I couldn't explain it -- we hadn't put the game on sale or anything, so I was clueless as to why sales were continuing to grow.

Then I checked out YouTube, and I noticed that fans of the game were uploading Let's Play videos constantly -- over 100 videos every day, each getting tons of traffic. Isaac had found its fanbase, and that base was growing larger and larger. Not bad for a game that was meant to fail!

What Went Right

1. Roguelike Design

The roguelike formula is an amazing design plan that isn't used much, mostly because its traditional designs rely on alienatingly complicated user interfaces. Once you crack the roguelike formula, however, it becomes an increasingly beautiful, deep, and everlasting design that allows you to generate a seemingly dynamic experience for players, so that each time they play your game they're getting a totally new adventure.

I wanted to combine the roguelike formula with some kind of real-time experience, like Spelunky, but I also wanted to experiment more with the traditional role-playing game aspect of roguelike games Crawl and Diablo. Fortunately, using the basic Legend of Zelda dungeon structure as the game's skeleton made it easy to rework almost all the elements of a traditional roguelike formula (procedurally generated dungeons, permadeath, and so on) into a real-time dungeon crawler format. Almost every aspect of the game seemed to fall perfectly into place with little effort.

Let's start by looking at the Legend of Zelda dungeon and resource structure -- it's simple, and really solid. Keys, bombs, coins, and hearts are dropped in various rooms in the dungeon, and the player needs to collect and use these resources to progress through each level. In Isaac, these elements were randomly distributed and not required to progress, but I included them to add structure to the experience.

I also pulled a lot from Zelda's "leveling structure," where each dungeon would yield an item as well as a container heart to level up the character and give the player a sense of growth; in Isaac, each level contains at least one item, and the player can get one stat-raising item by beating the boss. These items are random, but still designed in a way that made it so your character would have some kind of physical growth as you progress through the game.

I approached the roguelike design from many different directions with Isaac, but at its core, what made Isaac different than most roguelike games (well, aside from its visuals) was how I dealt with the difficulty curve. Instead of using traditional difficulty settings, I simply made the game adjust to players as they played, adding increasingly difficult content to the game as they progressed. This made Isaac feel longer, richer, and gave it the appearance of a story that writes itself. Using this design also allowed me to reward the player for playing and playing well, with more items that would help aid in their adventures and keep the gameplay fresh and exciting.

Once the player finally overcomes Mom, they usually assume the game is over, but instead get a new final chapter, six new bosses, a new final boss, and new items that shuffle into the mix. When the player beats the final chapter, they unlock new playable characters and items, and when they beat the chapter with each new character, they'll unlock even more content that makes the game even deeper still.

With Isaac, my goal was to create "magic." I wanted players to feel like the game was endless and alive, that the game had a mind of its own and was writing itself as they played. I remember the original Zelda having this feeling of magic and mystery. You weren't sure what things did until you experimented with them, and you had to brainstorm with your friends and put all your findings together in order to progress. I felt like since I was referencing Zelda so much in Isaac's core design, I should also complement it with the feeling of mystery I felt it had back in the day.