An interesting idea I had: epistemology sparring. Figure out a way to make model battlefield rationality work, say in some kind of combat-like game—think boffer LARP, paintball, or even lacrosse or dodgeball—just make it immediate and physical. Make success in the game tied directly to the ability to determine the probability of truth of some statement quickly, and more quickly and accurately than your opponents, and make the games short—no more than half an hour each. Do these frequently to allow for averaging out the failures that will definitely result during play, both at the beginning, due to inexperience, and due to bad luck or inconsistent performance. Any suggestions?
I occasionally go airsofting, which is quite similar to paintball. The organizers mainly borrow from First Person Shooters for their rules, but if you adapt the rules to something similar to Quirrell’s games in Methods of Rationality and throw in ways to do easily conceal information, you might be able to make it work.
First of all, have more than two teams. Two teams don’t allow for interesting scenarios such as teaming up, betraying, bluffing… Adding an extra one does.
Don’t have “respawns”. Failure should be meaningful and you can have “wipe out the enemy” as an objective. It also lowers the time needed to complete an objective.
Give incomplete or inaccurate information to your players about the goals or the layout of your arena. This forces them to quickly process new information they encounter on the battlefield.
There are other things you can do, but I’ll have to give it some more thought.
You can still have interesting stuff with more than two teams!
Players compete for points, and get 50 points if their team wins, but get points for other things: negative points for being killed, positive points for sub-objectives that don’t particularly help their team (picking up items hidden throughout the playing field, standing on tall things...).
For extra fun, give each player a random secret objective (including a “you actually play for the other team, and get scored accordingly” card).
Or for a totally different feel, have an “accomplish objectives” game (find hidden items, or bring back water to your jar with a spoon, the one with the most wins), but also give them paintball/airsoft guns, with rules on death/respawning etc.
I used to play airsoft a few years ago. The main problem with that would be the relatively large expense to start playing, and the need for a very large area. I agree that there should be more than two teams, as I’ve said elsewhere, and the lack of respawns would make sense.
Possibly, however, limited-range weaponry would be more fun/useful/easy to deal with—nerf/boffer swords, (which I admit I have no experience with) thrown weaponry, like tennis balls or dodgeballs, and so on.
In a way every game is a rationality game, because in almost every game you have to discover things, predict things, etc. In another way almost no game is one, because domain-specific strategies and skills win out over general ones.
One idea is based on the claim that general rationality skills matter more when it’s a fresh new game that nobody has played yet, since then you have to use your general thinking skills to reason about things in the game and to invent game-speficic strategies. So what if there were “mystery game” competitions where the organizers invented a new set of games for every event and only revealed them some set time before the games started? I don’t know of any that exist, but it would be interesting to see what kinds of skills would lead to consistent winning in these competitions.
There are various other ways you could think of to make it so that the game varies constantly and there’s no way to accumulate game-specific skills, only general ones like quick thinking, teamwork etc. Playing in a different physical place every match like in HPMoR’s battles is one.
Possibly. You’re giving me an idea—have a simple game with a few interacting rules drawn randomly from a larger set of interacting rules, and see who figures out how to take advantage of how the rules interact. Remind me to get back to you on this.
have a simple game with a few interacting rules drawn randomly from a larger set of interacting rules, and see who figures out how to take advantage of how the rules interact.
Hi, this is P. from New Jersey, right? (don’t want to give away your personal info if you don’t wish). I spoke to someone at the MeetUp yesterday who brought up this exact issue.
I think it’s a reference to the post at the bottom of the page, that appeared to receive a few downvotes and ad hominems, and the site blocked my attempt to fix the post or reply to clarify my intent. Not the best foot to start off on or feel welcomed. But your reply is much appreciated.
Whenever you become better at executing successful strategies in-game, you’re improving your instrumental rationality concerning your goal of “beating the game”. Already. As is. Probabilities don’t need to be explicitly stated, and typically aren’t.
Yes, of course. But the OP basically wanted to gamify teaching rationality, in particular by providing immediate feedback to decisions in a game setting. What I am saying is that modifying an FPS game so that specific rationality challenges (which reflect what you want to teach) result in gaining or losing power-ups is much easier than setting up a many-people physical game.
It would become easier, I agree. It would also lose a lot of what makes real, physical things special. Why do people still play actual board games made of cardboard, or bother to meet face to face? A lot of the rationality involved would probably be figuring out through facial expression and vocal intonation whether or not you can trust someone, and this is nearly impossible in a nonphysical context.
I would prefer that it be among relatively many people to allow for alliance-building, betrayal, and exercise of the social aspects of combat epistemology. Also, people who come in with knowledge of chess or boxing would have an unfair advantage.
An interesting idea I had: epistemology sparring. Figure out a way to make model battlefield rationality work, say in some kind of combat-like game—think boffer LARP, paintball, or even lacrosse or dodgeball—just make it immediate and physical. Make success in the game tied directly to the ability to determine the probability of truth of some statement quickly, and more quickly and accurately than your opponents, and make the games short—no more than half an hour each. Do these frequently to allow for averaging out the failures that will definitely result during play, both at the beginning, due to inexperience, and due to bad luck or inconsistent performance. Any suggestions?
I occasionally go airsofting, which is quite similar to paintball. The organizers mainly borrow from First Person Shooters for their rules, but if you adapt the rules to something similar to Quirrell’s games in Methods of Rationality and throw in ways to do easily conceal information, you might be able to make it work.
First of all, have more than two teams. Two teams don’t allow for interesting scenarios such as teaming up, betraying, bluffing… Adding an extra one does.
Don’t have “respawns”. Failure should be meaningful and you can have “wipe out the enemy” as an objective. It also lowers the time needed to complete an objective.
Give incomplete or inaccurate information to your players about the goals or the layout of your arena. This forces them to quickly process new information they encounter on the battlefield.
There are other things you can do, but I’ll have to give it some more thought.
You can still have interesting stuff with more than two teams!
Players compete for points, and get 50 points if their team wins, but get points for other things: negative points for being killed, positive points for sub-objectives that don’t particularly help their team (picking up items hidden throughout the playing field, standing on tall things...).
For extra fun, give each player a random secret objective (including a “you actually play for the other team, and get scored accordingly” card).
Or for a totally different feel, have an “accomplish objectives” game (find hidden items, or bring back water to your jar with a spoon, the one with the most wins), but also give them paintball/airsoft guns, with rules on death/respawning etc.
I used to play airsoft a few years ago. The main problem with that would be the relatively large expense to start playing, and the need for a very large area. I agree that there should be more than two teams, as I’ve said elsewhere, and the lack of respawns would make sense.
Possibly, however, limited-range weaponry would be more fun/useful/easy to deal with—nerf/boffer swords, (which I admit I have no experience with) thrown weaponry, like tennis balls or dodgeballs, and so on.
In a way every game is a rationality game, because in almost every game you have to discover things, predict things, etc. In another way almost no game is one, because domain-specific strategies and skills win out over general ones.
One idea is based on the claim that general rationality skills matter more when it’s a fresh new game that nobody has played yet, since then you have to use your general thinking skills to reason about things in the game and to invent game-speficic strategies. So what if there were “mystery game” competitions where the organizers invented a new set of games for every event and only revealed them some set time before the games started? I don’t know of any that exist, but it would be interesting to see what kinds of skills would lead to consistent winning in these competitions.
There are various other ways you could think of to make it so that the game varies constantly and there’s no way to accumulate game-specific skills, only general ones like quick thinking, teamwork etc. Playing in a different physical place every match like in HPMoR’s battles is one.
Possibly. You’re giving me an idea—have a simple game with a few interacting rules drawn randomly from a larger set of interacting rules, and see who figures out how to take advantage of how the rules interact. Remind me to get back to you on this.
Sounds like a matrix IQ test...
BTW, have people in this thread played the Flash game “This Is The Only Level” http://armorgames.com/play/4309/ ?
What is a matrix IQ test?
I have played that. My idea is slightly similar, I think.
Something like the Raven matrices.
Ok. I just wasted 15 minutes on that game and got up to stage 13. That looks potentially addicting. How many stages are there?
About 30. Fun. Just finished. (16:21, 83 deaths) Edit: uhoh there’s a 31. hmmm.
O. I can’t get past 31. Is that a real level? If so, what does one do?
Mafia as played online works like that—you have broad expectations, but no idea which roles exist in this specific game.
This is also very close to what I’m trying to get at, but I want for it to be both more physical and conducted in real time.
Hi, this is P. from New Jersey, right? (don’t want to give away your personal info if you don’t wish). I spoke to someone at the MeetUp yesterday who brought up this exact issue.
-Ernie
Yes, it is!
Pretty cool. Given what I’ve seen though (not from you), I don’t know if I’ll be posting much here. It was great to meet you guys.
Wow—someone down votes the new guy expressing apprehension at becoming part of this community? That’s a new low in terms of openness.
We’re not all like that, EGarrett!
Hi tgb,
I think it’s a reference to the post at the bottom of the page, that appeared to receive a few downvotes and ad hominems, and the site blocked my attempt to fix the post or reply to clarify my intent. Not the best foot to start off on or feel welcomed. But your reply is much appreciated.
-EG
Things become much easier if you drop the “physical” constraint.
A basic example would be the ability to gain power-ups by demonstrating rationality in some Quake/Unreal FPS environment.
Whenever you become better at executing successful strategies in-game, you’re improving your instrumental rationality concerning your goal of “beating the game”. Already. As is. Probabilities don’t need to be explicitly stated, and typically aren’t.
Yes, of course. But the OP basically wanted to gamify teaching rationality, in particular by providing immediate feedback to decisions in a game setting. What I am saying is that modifying an FPS game so that specific rationality challenges (which reflect what you want to teach) result in gaining or losing power-ups is much easier than setting up a many-people physical game.
It would become easier, I agree. It would also lose a lot of what makes real, physical things special. Why do people still play actual board games made of cardboard, or bother to meet face to face? A lot of the rationality involved would probably be figuring out through facial expression and vocal intonation whether or not you can trust someone, and this is nearly impossible in a nonphysical context.
Poker-boxing.
I would prefer that it be among relatively many people to allow for alliance-building, betrayal, and exercise of the social aspects of combat epistemology. Also, people who come in with knowledge of chess or boxing would have an unfair advantage.