By now you know the story. Threes was developed by a small group of indie game developers over more than a year. They’ve posted a treasure trove of their development emails so anyone can see just how much work they went through. It shows: Threes has a very steep learning curve and cannot be beat by following a simple algorithm. Yet, the game is still fun to play. The cards have personality, increasingly so at higher values, and combined with the bright primary colors they create a laid-back, almost joyful atmosphere. The music — somehow — manages to capture the carefree “everyone loses eventually, so don’t take it too hard” aspect. All the subtle animations give the game emotional weight.
Not two weeks after Threes was released, a 19-year-old Italian web developer released 2048, which was actually based on an iOS app 1024 which marketed itself as “no need to pay for Threes”. 2048 is open source on GitHub, the de facto home of all recent, hip open source projects. Therefore it has spawned no end of derivatives and copy-cats. Most of these versions are meant to be funny, parodies, or even satiric. They are transient, gimmicky, and not all that fun. And yet, they are hackable. Tufts’s web development class had its students add a database-backed high score system. You could build a whole ecosystem of services related to 2048, all free and open source, malleable to anyone who knows how to code. And everyone knows how to code, right? And everyone wants nothing more in life that tile games, right?
Threes is like an Apple product: meticulously designed for functionality and emotion. The experience is open, with the bright white background and endlessly fun gameplay, but the implementation is closed, behind a paywall and with private source code. 2048 is like Linux: free, open source, but of lower quality. Sifting through a half dozen 2048 clones to find one you like and that doesn’t crash or spam you is remarkably similar to managing Linux software packages. 2048 is just numbers, and lacks the emotional charm from the music, animations, and voices of Threes.
Indeed, it’s possible to do fairly well in 2048 just by mindlessly shuffling tiles into a corner. This strategy is much less effective in Threes because the 1 and 2 cards get in the way. Threes also has a “next” indicator, telling you the color (but not value!) of the next card. This strikes the perfect balance: knowing the exact value would make the corner strategy workable, but knowing something allows the player to plan their next move. Being mindful of where cards are entering the board makes a tremendous difference in score. Threes is a game that rewards concentration; you have to think. 2048 is so easy a robot could do it.
The Threes developers spent a tremendous amount of time exploring every nook and cranny of the sliding numbered card concept. They eventually found a great game, hidden in the space of all possible games. But that space, as huge as it is, is actually fairly small. There’s only so many different choices you can make regarding core game mechanics, and yet it took months to explore. So when other people attempt to tackle much larger problems, like education or global poverty, without having even laid the groundwork that the Threes developers did, I’m extremely skeptical.
That said, much of what makes Threes stand out is the auxiliary material: art direction, sound, menus, the tutorial. And indeed that’s a lot of what the developers went back and forth on. Here is one except from their massive email archive:
But recently I’ve found myself thinking about the game as an exploration of identity. Like how the phrase “be yourself” is utter bullshit because “yourself” isn’t a thing that exists until you create it. We start out as just the sum of our nature+surroundings (1+2) and eventually we coalesce that into a sense of self that we can define and present to others. (3) And then eventually you get enough perspective to self reflect (3+3) and decide how to change.
And the metaphor actually holds up because your constantly learning about new ideas (1s) while dealing with genetics (2s) and also memories of how you dealt with stuff in the past (3s, 6s, 12s…) and really the trick is about how to line them all up and figure out how to grow.
I’d like to see that put together by a nineteen year-old in a weekend.