Just call me Mr. O’Brien* 
I do think that this understanding of the dichotomy between feeling Free Will vs. actually exercising it is central to speculative fiction, i.e., Science Fiction that focused on speculating about the Utopian and Dystopian futures that may lay before us.
*the antagonist from Orwell’s Nineteen Eighty-Four
From an article in the NY Times providing a synopsis of the latest episode of Westworld:
[QUOTE=NYTimes]
For visitors to the park, the chief allure of Westworld is freedom: freedom from the strictures of legality and social norms; freedom to indulge in whatever violent and sexual fantasies they can imagine; freedom of the kind settlers might have experienced in the wide-open terrain of the Old West, before civilization came along and posted the guardrails. The paradox of Westworld is that freedom isn’t possible, because the park is monitored more closely than Disneyland and the hosts are synthetic constructs, acting on their loops. Comparisons to “sandbox” video games like “Grand Theft Auto” are apt: Players may enjoy wandering around between missions, shooting at random passers-by and looking for “Easter eggs,” but they’re penned in by the limits of this heavily monitored world.
[/QUOTE]
Bolding mine.
I think you might enjoy some Designer’s Notebook articles on Gamasutra. If you are considering free will in terms of player choice, two opposing terms you’ll run into are railroad (straight game path, no real choices) vs sandbox (players can do whatever they want whenever they want, mostly). There’s also a middle ground of real but limited choices, or false but real-looking choices that bring you around to the same ending. The issue here is player freedom, and how much your game is willing and able to grant it.
Nope. You’re making the classical mistake of demanding that the null hypothesis be confirmed. We don’t require that. The null hypothesis can be rejected, and that’s all that is required for a test to be passed. We never confirm the null hypothesis, only accept that in some tests it is not rejected.
If you’re testing to see if a new drug successfully treats a disease, you will get a few cases in trials where the patient wasn’t given the drug…but got better from the disease anyway. This tells us nothing about the efficacy of the drug, and it certainly doesn’t mean “the test is absurd.” That’s absurd!
Bookmarked; thanks.
Trinopus as I have said, I don’t believe the stuff I am looking to group-noodle over is not contingent upon the definition of Turing Test.
We are engaging a Today that has Gaming, stories like Westworld, the Simulation Hypothesis, etc. All include an element of Limitlessness, dialed up or down based the scenarios being discussed. I am interested in understanding what is being discussed about Humans’ ability to accept somewhat-to-very bounded Worlds.
Um…huh? I lost you somewhere in there. What should I be addressing?
It might be more useful to talk about the ability to make meaningful choices within a game (or at least the appearance of such an ability) instead of using the term “free will”. All games place artificial restrictions on what kinds of actions a player can take, but nearly all games still allow players to choose between multiple possible actions.
A notable exception is a game many of us probably loved as small children but isn’t much fun for adults – Candyland. When discussing game design I’ve joked that Candyland is basically Determinism: The Board Game because the winner has been decided as soon as the cards are shuffled. There is no choice or skill involved, players just take turns drawing cards that tell them which square to advance to. I’d guess that one reason why small children enjoy this game is that it’s one where they can regularly beat their older siblings, and that they’re too young to feel that a victory based on pure chance is a hollow one.
Most other games fall somewhere on the “railroad” to “sandbox” spectrum that xnylder described. In a “railroad” game there’s one correct path to victory, and while player can choose to do other things they will either fail to make progress or “die” and have to start over. This may not be much choice, but it does allow players to feel that their progress is dependent upon their own actions (their choices are meaningful), and in a well-designed game figuring out the one correct solution to a puzzle can be very satisfying.
xnylder - I read it; really good. Given the Three Problems, it seems like the gap between Games and a true feeling of limitlessness is some time off.
Trinopus if I am missing something, let me know. Are you applying the Turing Test not to a Thing you are communicating with, but a World such as in a Game or Reality Simulator?
Lamia - love the Candyland analogy; that’ll bake my noodle for a bit ;). With the gamasutra article, it gave me perspective about the gap between a game and a plausible World. I think the Westworld, uh, world also points out stuff we still need to figure out. I see that Free Will is too specific vs a term like meaningful choice. Thanks.
I can’t figure out what you’re saying here.
I’m applying the Turing test to a thing that humans are communicating with, like chatbots and other programs that emulate human conversation. If they can do that, why shouldn’t they also be able to emulate “free will,” at least within the confines of a game universe?
Take fight games like Mortal Combat (or is that Kombat?) Anyway, like Street Fighter. You don’t know which moves the (computer) opponent is going to use. The opponent has all the appearance of intelligent strategy, just as a chess playing program does. Inside the limits of the game, computer adversaries are, these days, so sophisticated that you cannot be sure you’re not playing against a real human.
Okay, so a Mortal Kombat computer fighter that a player enjoys fighting and feels like it passes as a legit opponent passes a Mortal Kombat Turing Test? Ie we shouldn’t limit the TT to the commonly thought of scenario?
Sorry if I am being slow here.
I don’t really understand this question. If humans ever exercise free will then of course they exercise free will when playing games. Indeed games could be defined as a series of interesting choices.
Of course in games your choices are limited, as they are in the real world. The real world may present you with more choices, depending on the situation. But I don’t see why this quantitative difference would make one free and one not free.
WordMan - I’m glad you liked the articles. Gamasutra’s a good site overall for insider opinions on games and game design.
Like others have stated above, I think we have to separate the concept of “free will” from the more commonly used term “player choice”. “Free will” is probably too broad of a concept for games.
A game like Super Mario Bros. or Dark Souls involves very little player choice: Continue on a straight path (like a railroad). Defeat enemies and try not to die. A sandbox game like Minecraft, Animal Crossing, or Stardew Valley has much more player choice. What do I (the player avatar) feel like doing today? What will I build? Will I interact with NPCs (if possible), and if so, will I help or hurt them?
Even in a well-built and vast sandbox game, however, player choice is only a tiny subset of free will. For example, in the farming simulator Stardew Valley you can choose to grow crops, explore a mine, or talk to villagers among other things. All these options are built in. But you can’t steal tools or kill a villager. In Grand Theft Auto, you can steal weapons or kill NPCs, but you can’t grow crops. Valid player choices are determined by the design of the game, which in turn is informed by the game’s genre and intent. This may reduce overall “free will”, but in turn makes the game playable.
I appreciate everyone’s patience with my word choice. I am totally open to stepping away from Free Will and referring to Player Choice instead.
I also appreciate this thread’s discussion, even if my question isn’t well defined or clear. I am, I believe, kinda trying to work back from the Elon Musk-popularized scenario that we exist in a computer simulation created by a more advanced race, and also working back from a show like Westworld.
In the Musk scenario, we would have to exist in a simulation that is FAR more sophisticated in “allowing Player Choice” if I am using that phrase correctly. And I assume some of that sophistication would be in “enlarging the sandbox” and some of it would be “camouflaging the railroad” (again, assuming I am using those phrases correctly).
As I look at those challenges, I was stepping back and trying to look at the state of Games today vs. that scenario. Based on what I am reading here and at the Gamasutra site, we are pretty darn far away from being able to build a program that can accomplish this.
As for Westworld, again, there is the Game aspect to it - how fun would it really be to step into a reset-every-day set of scenarios featuring robots with fairly limited parameters? So they have built a bigger sandbox vs. what we can really do, and we are expected to look past the ways they camouflage the railroads of each storyline the robots are part of.
The key thing about a game is that a game has a pre-defined goal: capture the flag, defeat the other group of guys, shoot the main boss, whatever.
As such the game has already constrained not what the players *can *do, but what the players *want *to do.
Think of a hypothetical exploration game where the sole purpose is to wander around and see cool sights, fanciful animals, weird plants, etc. That’s it.
In this “game” the game world does not have to be infinite in extent. It just has to be large enough that very few people wander all the way to the edge. We can help achieve that with ever less simulated square mileage by doing things like not starting the player near an edge, putting forbidding terrain near the borders, etc. We are setting up soft limits which tend, not demand, that the user behave in ways the designer wants. Meanwhile the vast majority of users will *not *choose to scale the high overhanging cliff or swim the raging rapids. They’re exercising legitimate free will to explore the easier route despite the fact the obstacles are, unbeknownst to them, placed there deliberately as behavior controls. We can even do subtle things like put more, and more brightly colored, animals in a subtle gradient towards the center of the game space. Folks seeking animals will sensibly turn towards the gradient and unwittingly away from the edge of the world.
Switching gears …
Insane real life humans have truly wild levels of free will. Much more than you or I do. They might be chatting nicely with you on a park bench then suddenly bite your face off while screaming Shakespearean love sonnets. What we call civilized sane behavior is about the “designers” of our collective society limiting absolute free will by imposing soft limits. With the result that sane people don’t think to shout sonnets while face-biting. And non-psycho-criminals don’t think to face-bite, sonnets or no.
Any of us *could *do either or both of those things any time we wanted. But we don’t tend to think of ourselves as being *prevented *from doing that; instead we don’t even think of the possibility.
So ISTM as all this relates to actual game design, the ideal design goal is to present enough of a structured reward system that players are “pulled” towards game paths that you provide for and not pulled towards game paths that aren’t. Yes, you do need some room for sandbox free play. And the less goal-directed your game is, the bigger the sandbox must be. In terms not only of geography and variety, but in terms of actions players can take.
As an example:
I’m also not a gamer. But I bet there’s nothing that formally prevents a World of Warcraft raiding party from surprising a defending force with a hail of weapon fire then promptly sitting down in the open to have a picnic with some beers. The game doesn’t need to prevent that *as such *because the players won’t think to do it. And the defenders, be they other people or be they bots, don’t need a special rule for what to do when encountering picnickers. “Pursue and smite visible enemy” is a good enough rule to appear as intelligent behavior to the foolish picnickers as they’re being killed.
Hope some of this is somehow relevant to the area you’re musing about.
I think you’re confused because of the OP’s use of the term “free will”, but **WordMan **is NOT asking about creating an NPC game character that makes (or appears to make) independent decisions. He’s asking about designing the game so that the player feels they have the ability to make a wide variety of different choices within the game world.
Oh! Cool! Wouldn’t that be the “sandbox” model of game?
I remember an old-ish Batman game, where you could follow the clues and solve the big mystery…or you could just drive around Gotham City and beat up burglars and muggers and robbers. It didn’t advance the plot; it was just a fun way to mess around, being Batman and clobbering the wicked.
I don’t game much…but I’m wholly in favor of broad ranges of player choices. We’ve come a long way since the early Infocom games where you had to do things exactly right in order to win.
Thanks all for the posts, and Lamia for the re-state. That is what I am asking about: a Human’s sense of freedom. Free Will, as many have patiently explained :smack:, Is not the correct concept I am noodling on.
Trinopus - yes, but. Meaning: yes a sandbox game is what I have been thinking of, but per xnylder’s linked Gamasutra article - brief, easy to read and really interesting - sandboxes face clear structural issues that will strictly limit their ability to feel limitless.
LSLGuy, yes that’s interesting stuff. We do accept limits all day, every day. but we live in a much more complex, less-structurally-limited sandbox. I wonder if part of what I am thinking about is: we as successful Humans have that socializing, set Good limits tendency. How do simulated realities exploit that tendency while providing a lot less Reality to work with?
ETA: how about this: What does it take to remove the feeling of “falseness” in a simulated reality? There’s a Turing Test aspect to that, but more about bridging the Uncanny Valley, but on a Worlds basis vs. an Individual robot basis. ???
Then I’d put forward that the Telltale game model (and certainly The Walking Dead, which is the one I’m most familiar with) do give you a great sense of freedom, choice and consequences without actually having any. The formula goes such : a story unfolds, and at times you’re given token busywork to do to make it unfold further, making you feel involved in it or that you contribute to the story. At these times you typically also have the freedom to chat with other people and form emotional connections with them. Then you hit the Big Choices in the story : a dramatic event where you’ve got two, maybe three options you can pick, each having ostensibly far-reaching consequences and none standing out as the self-evident winner. So you pick one, and something dramatic happens, and the story keeps unfolding with the shadow of that choice hanging over you. And it does feel like you’ve accomplished something or made meaningful decisions. Very strongly so, because they’re very good at storytelling.
Except when you replay the game and make the other choice(s), you realize that the story unfolds exactly the same save for some cosmetic details. Maybe one character isn’t in the background of the scene, or maybe that one character died this way instead of that way. All inconsequential and trivial. There really was just the one story after all.
In a lot of ways this is a pretty good explication of real life for most people. You have choices. most are trivial and affect the unfolding story only marginally: Which grocery store will I visit today; the expensive close one or the cheaper more distant one?
Then there are the larger choices: take this job, marry that person. The have large consequences in toto, but even then for most people in most environments the range of eventual outcomes is not so large. A vast list of trivial stuff will be different in specifics. The big picture of both futures is pretty similar.
My bottom line, following upon my earlier post: The game environment only needs to provide *slightly *more freedom than a player actually uses in order to feel to that player as if the freedom is limitless. If you never catch sight of the Big Wall, you’ll never know it’s there.