Save the Singularity is an upcoming cooperative, push your luck dice game that I am working on.


The Technological Singularity, the moment when the first smarter-than-human AI starts self-improving and promptly becomes super powerful and remakes the world in the image of whatever goals it was coded to optimize for, is coming. No one can stop it. What the players are trying to do is ensure that the first such AI is Friendly, a technical term which means that the end-state it turns the world into is one that humans can live in and enjoy, instead of a Terminator-esque hellscape. The default for an AI is to be unFriendly: you have to try really hard, and have a working, fully fleshed out, mathematical theory of Friendliness to even have a chance of making a Friendly AI. Fortunately, you and your friends are a team of computer scientists, programmers, mathematicians, and philosophers who just finished Friendliness Theory. Now you need to actually code the darn thing, before someone else makes an unFriendly AI and dooms us all.

Turns are taken in two parts: coding, and what the rest of the world does. When you code, you roll the coding dice, which each come 3 possible results: Good Code (1 point), a Promising Lead (if you press on, roll an extra die), and a Bug. If you have one bug, that’s okay, you can fix it, but if you have two or more, your code is fatally flawed and must be thrown out: your turn ends and you score no points. After each roll, assuming you didn’t get two or more bugs, you can choose to keep working on that bit of code (roll again) or save and compile your work (score the points you’ve rolled and end your turn). At the end of your turn, you roll the single 12 sided Progress die, to represent what the rest of the world is doing. The three possible results are UnFriendly Code (the rest of the world scores a point), Useful Research (the next player starts with an extra Coding die), and Stop (you keep rolling the Progress die until you roll a Stop).

You’re trying to get to 25 points per player before the rest of the world gets to 3 points per player.

When designing Save the Singularity, I had these ideas in mind very strongly throughout. Save the Singularity is supposed to be fun, but it’s not supposed to be only fun, and that’s not even at the forefront of what it’s supposed to be. At its core, it’s a horror game. Almost a thriller. It’s meant to evoke the heightened alertness, the positive anxiety, the tension, of being on the precipice of something great and terrible. There’s this monstrous, overwhelming force rolling your way, and you and the other players have to band together to have even the slightest chance of getting through it alive. Everything is riding on your efforts, and even if you do your best, the odds are you’ll probably lose.

The game is cooperative. It’s not the players warring against each other, it’s all the players together struggling against the great, indifferent might of the rest of the universe. A tiny, doomed few against the incredible, uncaring ability of the rest of the world. What makes it worse is that the rest of the world isn’t malicious, they’re not trying to kill you-it’s just that they don’t know what powers they’re tampering with, and they will end up releasing forces beyond their wildest nightmares should they succeed before you do.

The game is difficult. It’s not impossible, but you have to play really smartly, and not give in to the biases all gamblers face, to succeed. In addition to all the in-person playtesting, I wrote a computer program to simulate simple agents playing the game a few thousand times in order to iron out some kinks in the probabilities. A basic, naïve algorithm-if you’re about to roll more than 6 dice, stop-wins 35% of the time. A perfect super-agent, able to predict but not affect dice rolls, wins 85% of the time. That implies that there exists a strategy such that if you play really smartly, you have just better than 50% odds of winning. (and in real-person playtesting, this seems to hold true) Which is what I want: if you do everything right, you can barely tip the odds in your favor. It’s sacrificing a little bit of realism (the odds of a negative Singularity vs a positive one in the real world are much worse) for a lot better gameplay. I want to evoke terror and a feeling of nigh hopelessness, but not make it so clearly actually hopeless that people don’t bother to play.

And according to the feedback I’ve been getting in the actual human playtests, I have succeeded. I’m really satisfied with this game, and despite being so simple, I think it brings two important new elements to the casual, push your luck style party game scene: being cooperative, and being primarily about an experience other than “fun”. This is an area of gamespace I think should be explored more.

What games have you played and enjoyed where the focus is on an experience other than “fun”?

Leave a comment

Leave a comment

Your email address will not be published. Required fields are marked *