Somewhere else, Mike Holmes said this:

Mike Holmes wrote: If, on the other hand, you want to go with a real narrativism based [computer] design, then we're talking about getting into an area of development that's never been tried before. Never. I've postulated such games before, but have never gotten beyond the idea stage.

I don't think it's the binary issue that's the problem. It's that the computer has to predict somehow that the choice being offered the player is somehow meaningful to that player. I'm not sure offhand how a computer would be able to judge that. Even with loads of input at the beginning, it'd be hard to make those choices somehow player driven.

I mean, what if the computer decides that I must choose between freedom and justice, and I don't really care about either, so I just pick one at random. Is that narrativism?

Matt Wilson wrote: I don't think it's the binary issue that's the problem. It's that the computer has to predict somehow that the choice being offered the player is somehow meaningful to that player...

I mean, what if the computer decides that I must choose between freedom and justice, and I don't really care about either, so I just pick one at random. Is that narrativism?

I'm in between these two positions.

Tim, what you've described has been done before, of course. It's the "choose your own adventure" style of game. Though it's mostly done in books, I'm sure that there are computer versions. Moreover, I'm sure that there are games that have done it on computers a tad more complexely.

Here's the thing. Narrativism as a form of play doesn't require any skill in terms of making the the decisions that identify the play as narrativism. Once you've gotten to a narrativism deciding point you can't make a wrong decision. It's like falling off a log. You just make a decision you like, and you're done.

The skill in narrativism is discerning during play what situations are going to make for this sort of decision, and maneuvering into them.

Now, yes, you can make situations that are "asutomatically" of the sort that are likely to be interesting to everyone - this is what the writer does when writing a book. And one can make bangs for characters, yes, before play. But sans a player having some input into the process of deciding what the character is like, and what issues are important, it's all very problematic. Like playing with a really bad GM who has written up characters for you, and decided on all of the conflicts before hand.

Now, what I can see working is having the player allowed to indicate what's interesting to him via the character. The player might, say, choose values for the character and issues. Then the computer uses some algorithm to create conflict around and between these values or issues. The player chooses to maneuver the character through these multiple situations as he sees fit (in fact these become "meta-narrativism" decisions - do I deal with the choice about my brother, or the problem with the slaves?) As the issues are resolved, values change, the computer randomizes, and new situations are generated.

I think it's entirely possible. Not simple, but possible. And It might be really cool. Somewhat like Gamism computer RPGs. Yeah, the computer can't choose really interesting tactics or anything - it's limited to what the interface will allow. But the series of events can be really interesting. Instead of, "And then I fought the dragon and got the sword of Arasis which allowed me to defeat Lord Zerxar" you'd get, "And then I decided I had to sacrifice my brother so that the army would be able to win the war, so he ran off, and then I had to decide if I would stay and face the council or if I would chase after him so that I could catch him and bring him back."

Give any vision?

Mike

I would think that a "choose your own adventure" type of Narrative-style computer game is possible if the player is made aware of what he will have to give up whenever he makes a decision. He might not know where a decision will lead, but the game must force the player to make hard choices to explore a Premise.

An "undo" or "save and restore" feature might be interesting for such a game. Did your choices lead to a conclusion that does not satisfy you? Back up and explore a different path.

I don't get this 'it must customise premise to the players wants' thing. Why?
As Timothy said:

Hmmm, I'm just throwing things out right now, but can't the designer just say, "this game is about freedom and justice, if you're not interested go play something else"?

I agree with you Callan.

My point is merely that this has been done before, and that I think it would be pretty easy to do better. That, basically, you're not tapping the resources of the computer well with the original model. It'll work, it just misses out on what I think computers could do in this area.

I'd like to hear from Walt on this topic if he's reading. I'd bet he'd have something important to say about it.

Mike

I've been thinking about posting in this thread. My thoughts are still a bit unformed, but here's where I currently stand.

I don't think there's any fundamental reason why computer games cannot be created to facilitate Narrativism. But I think there might be a fundamental difficulty with any solo game, computer or otherwise, facilitating Narrativism.

See, half of the Narrativist equation -- am I, the player, addressing (I like to say, expressing myself about) a Premise of emotional import? -- can in theory be satisfied with any game including any computer game. Give me a flight simulator, and I can use it to examine, in my mind, whether it's more important to fly safely or to keep to a schedule that might require pushing the safety margins. This is a question that anyone who drives a car can relate to, and one of literally life and death significance. I can imagine, as I fly through the simulated weather patterns and so forth, that I've got a passenger named Penny who really has to get to Chicago in time to say goodbye to her dying father, and another passenger named Mr. Higgs who's suffering panic attacks every time we hit an air pocket and is demanding that we land immediately, and a co-pilot who's a week away from retirement who doesn't want to take any chances and tempt fate, and a boss back at the airlines who's telling me I'll be fired if I don't fly through the storms to arrive in Chicago on time. What will I do? And what will the results of my decision be?

Of course, I have to imagine all those characters and situations on my own. The simulator doesn't help me do this (though it does supply the storms). So the simulator certainly isn't facilitating this "narrativist" play on my part.

But in theory, it could. It wouldn't be difficult to add to the flight simulator a bunch of such characters yelling in the player's ear and reacting to what the player does. Not connected to any scoring mechanism, you understand. No "you get 100 points for arriving on time but lose 40 points if Mr. Higgs freaks out before you get there," which would turn it all into a resource management (Gamist) problem. Just the characters acting and reacting so as to make a meaningful decision (turn back, go around the storms, or push through) necessary. So, would that be a good idea? Would it result in a Narrativist flight sim?

This is where I get into hazier territory. I think there's something still missing. The problem is that the program doesn't actually give a crap what the player does. The player might be addressing (expressing himself about) a Premise but nobody's listening. Can that be Narrativism?

Before anyone pops up with a quick and easy "no" answer, though, consider that computer programs don't actually give a crap about who wins or performs well in a Gamist game either, and yet that hasn't stopped computer games from being acknowledged as providing or facilitating solo Gamist play. Most game programs, or the characters they're portraying, do a pretty good job of pretending to give a crap, and that's apparently good enough.

But it's a little different, when we're talking about traditional computer game scoring and feedback. Because a computer game program can score and adjudicate a game so as to determine e.g. who won or who performed well, it's not unreasonable to say that the computer program actually does "understand" those aspects of the game. "You found eight of the ten hidden mushrooms. Good job!" says the machine. The "good job" part is questionable, but we (the audience) let it slide. Presumably the programmer knows that finding eight will require some effort and therefore merits a "good job!" assessment of the player's performance, but we also know that neither the machine nor the programmer actually cares whether you found the mushrooms this time or not. However, the machine clearly does know that you did in fact find that many mushrooms, so that part of its statement is completely credible.

By contrast, there is no doubt that no computer program can "understand," by any stretch of the meaing of "understand," a Narrativist Premise or a player's actions as they relate to the Premise. By limiting the player's fundamental choices, the designer can provide canned feedback for the choice made ("You decided to fly the safe route, showing that you hold your commitment to your passengers' safety above other concerns -- or perhaps that you lack confidence in your flying skills! Penny is upset that her father passed away before she could get to his side, but at least she's still alive herself. Your boss tried to have you fired but the Union intervened and blah blah blah...") But the player has no opportunity to add nuance and have that nuance be recognized, which I think is where Narrativism really lives (rather than in the instantiated yes-or-no Premise "answer," if any). So can this be facilitation of Narrativism? Still up in the air, so to speak.

For a micro example of this approach, where you make a brief series of "significant" decisions and get canned feedback on the big ideas your decisions seem to be epressing, see this game. Try it two or three times with different answers. Small-scale as it is, I think it might provide a useful example for further discussion.

- Walt

Walt Freitag wrote: But the player has no opportunity to add nuance and have that nuance be recognized, which I think is where Narrativism really lives (rather than in the instantiated yes-or-no Premise "answer," if any). So can this be facilitation of Narrativism? Still up in the air, so to speak.

Mike: Ah, gotcha!

Hi Walt,

Why would someone else have to be listening for it to be narrativism? We might be thinking a little bit too much in our table top roots here, as though we need others present. Tons of other media like movies or soapies present address of premise all over the place for us to observe. We can watch these by ourselves (or with others but with no real discussion afterward on the media) and it doesn't change the nature of that media. In a computer game where we decide the address of premise ourselves, it's the same except its more intimate.

I think what you might find missing is the massive reward of addressing premise in front of other people. I think there are many types of rewards inherant to doing that. And in contrast addressing premise alone may seem...flacid. This sudden contrast is like going from a bright area to a shady area...it makes the shady area seem pitch black at first. Likewise, there may seem to be no reward at first. But I think if one waits for ones 'eyes to adjust' you'll see it there.

As for 'nuance', I think this is where the player doesn't just address premise but addresses it in a particular way they have customised. This adds special meaning to the address. And the probs I think you've identified are: The comp may not accomidate how you want to add nuance and it wont recognise what you do anyway.

To this I say that in the company of other people, are they really absorbing all the nuance you generate or are you just hoping they are? And if they aren't absorbing it all, then was adding nuance special to begin with?

I think adding nuances can be very rewarding with other humans though, even if they don't really get why you put that nuance in. And lacking that reward with the comp means the player is less likely to go to the effort of generating nuance that goes unrewarded (and if he does, comp limmitations may mean he can't add some nuances as well).

Personally, looking at that link you gave, it left me thinking about difficult issues. This isn't in the short term, rewarding. It's difficult. However, if I'd gone through such a thing with others there is the reward of answering such things in front of others and getting a lot of rewarding feedback from them instantly. If that's a problem, I think that's your main prob with C narrativist.

Additionally, in that last question I didn't consider the soul described as what I think of a soul. I considered a thirty percent chance of living and loss of what the question described as a soul and living on as being like living as only 30% of a man. I basically chose euthenasia because of reduced lifestyle, rather than to save this eternal soul described (which I, as another address of premise, decided not to recognise as a soul. Recognised it as perhaps something important to me, but not a soul). Sort of like a nuance, this is a switch in the problematic issue really. And in a C RPG if I had to make a choice like this early on, where I could only address it from another angle, I'm sort of out of kilter with the rest of the game. I can see that prob, but mostly because in that links last question, I feel there were two addresses of premise, not one (the first is, do you believe what they describe is a soul). In writing a CRPG, it might be important to make sure you seperate out your problematic issues (ie, ask if they believe that is a soul...if they don't, they can't really answer that third question as put).

I recently played the second of the Deus Ex games, after recommendations on this baord although I forget by whom. It's reputation as amongst the most intelligent of the first person games is not undeserved IMO. It;s basic format is in no way novel, but it does make the transition to asking relatively tough questions about values, or at least, presenting the player with such dilemmas.

Essentially, in each level the player will receive at least two and probably more conflicting goals. Choosing which goals to fulfill and which to ignore cannot be done without some sort of moral contemplation in most cases. This works unusually well IMO if only by contrast to the multitude of games that only ask questions about How you achieve goals.

It's not really open ended, and not really that free (I hear the first one was better in this regard). It's also not so world-shattering that I want to over-hype it. But I do think that this shows the way forward; presenting options that are about the people in the game and how the character relates to them through the players decisions, rather than merely efficiency measurements.

Sounds cool. Is it completely about these decisions, or is it intermixed with the standard CRPG gamism? That is, do you hack your way from one decision point to another? How do you get from point A to point B in the decision tree? Does this merely "branch" like Tim's original suggestion, or do the situations evolve in a more multivariate fashion as you play?

Can the player have any impact on the situations themselves, other than the choices made at decision nodes? That is, do they have some input on what the situations are going to be like? Perhaps via chargen? Or in-game play? Do they pick up the values in question in play, or are they preset?

Mike

Mike Holmes wrote: Sounds cool. Is it completely about these decisions, or is it intermixed with the standard CRPG gamism? That is, do you hack your way from one decision point to another? How do you get from point A to point B in the decision tree? Does this merely "branch" like Tim's original suggestion, or do the situations evolve in a more multivariate fashion as you play?

Can the player have any impact on the situations themselves, other than the choices made at decision nodes? That is, do they have some input on what the situations are going to be like? Perhaps via chargen? Or in-game play? Do they pick up the values in question in play, or are they preset?

Cool, that's much clearer.

Does the endgame change depending on the selected choices? You say that it makes some statements - are those in any way affected by your choices? Or does the statement come out the same no matter what you chose?

I'm not saying that wouldn't still be narrativism, just that it would be narrativism only in short limited bursts.

Also, this seems to be a case of "partitioned" CA again. That is, you go from narrativism moment, to gamism "shooting", and back and forth like this until you get to the end. This is similar to my claim that TROS is patitioned in a very similar way. (Which, again, I think is cool).

Mike

You need to claim that Mike? I thought it was pretty obvious that TROS was all 'oh, I will do anything to save my love...now watch me kick this guys ass with my leet moves!'

In CRPGs it'll probably become the new black in five or ten years, if they grasp the facts of it (ie, they see nar but manage to avoid implimenting typhoid mary programming). Especially because I think it will draw in that otherwise elusive female demographic.

Mike Holmes wrote:
Does the endgame change depending on the selected choices? You say that it makes some statements - are those in any way affected by your choices? Or does the statement come out the same no matter what you chose?

You know, a lot of this sort of thing might be possible in old-fashioned IF (Interactive Fiction, like "Adventure" and "Zork") formats.

First, some references:

Nick Monfort, "Toward a Theory of Interactive Fiction"
Emily Short, "On Multilinear IF"
Jim Fisher, "Advanced NPCs: Introduction" (and the subsequent articles)

Backing up to these sites' homes will get you lots and lots more -- it goes on forever. Note that the Inform people are more interesting than the TADS or other system people.

One of Short's points is that there is no intrinsic reason you cannot have an IF game that really has no goals or scores as such, or that assigns them dependent on what sort of story you seem to be walking into.

So for example, you could treat the thing as a pure puzzle and try to "beat" it, in which case the game should be written such that it notices a kind of statistical buildup of puzzle-solving and slowly increments the rewards provided by such activity, when successful.

If on the other hand you decided that you were more interested in having elaborate conversations with NPCs and grappling with some Premise (front-loaded, as in Deus Ex), the system again notices this behavior and increments the rewards offered for such interactions. Note that in this case, "success" would be simply a matter of having the conversations and choosing, when offered let's say a number of different things to look at or places to go, the option most compatible with the Premise concern. This would be very difficult to program, but it's possible (especially now that the Glulx engine has lifted pretty much all the memory restrictions).

Note also that, as Fisher discusses, NPC conversations do not have to be "choose your own adventure" types (called CYOA in that community), because the player interacts with the game through text and a complex parser rather than menu lists. So you can develop very intricate trees that are genuinely multidimensional rather than simply giving that illusion.

What I doubt very much is possible is a Nar IF game that doesn't have a frontloaded Premise, because there would be no way for the computer to guess why you were doing what you were doing.

It's of course also possible simply to drop all the scoring stuff and have a series of choices (invisible to the player) available for the conversations and room descriptions and such. The more the player keeps harping on the character's personal problems, let's say, the more the responses match those same concerns.

For example, let's say that the Premise is about what you are willing to give up to save your baby.

Janet says, "I'm Mr. Johnson's secretary. Can I help you?"
>TELL JANET ABOUT BABY

This suggests, if done early in the conversation, that the player is focused on this issue, whereas

>ASK JANET ABOUT JOHNSON

suggest this somewhat less.

If the player keeps focused on the Premise, the descriptions will shift:

David asks, "So... when are you due? Sorry, I don't mean to pry."
>TELL DAVID MAY
"May, huh? That's pretty soon. Gonna be quite a change in your life, you know. Believe me, I thought it wouldn't be a big deal, being a father and all, but it's changed everything."

And so on.

Difficult to program, but not in any sense impossible.

This was precisely what I was imagining when we were speaking of "choose your own adventure" books online. They are functionally similar, yes.

I wonder if anyone from www.skotos.com would have any commentary on this and how they present their games? Though multiplayer, many of their rooms have programmed responses, last time I checked.

Anyone know anybody over there?

Mike

Mike Holmes wrote: This was precisely what I was imagining when we were speaking of "choose your own adventure" books online. They are functionally similar, yes.

Game Theory has described a number of scenarios where the outcome isn't completely random & which force the players to make choices about what is important to them.

A quick glance at the theory of non-zero sum games would help you set up situations.

Moreover, Nigel Howard's Drama Theory combines set theory with game theory to describe six typical (non-zero sum) games where individuals have to reframe their own values, rethink their perception of a situation, negotiate compromise, or even rethink the game itself.

I am making it sound more touchy-feely than it actually is. The math (what I can understand of it) makes sense, and the logical description of the "6 Dilemas" is persuasive.

Check the Dramatec website (http://www.dramatec.com/articles/beginner.html)

There was a good article on text gaming and Drama Theory, but they have pulled it from their new website. I can email you a copy if you are interested.

timfire wrote: Somewhere else, Mike Holmes said this:

Mike Holmes wrote: If, on the other hand, you want to go with a real narrativism based [computer] design, then we're talking about getting into an area of development that's never been tried before. Never. I've postulated such games before, but have never gotten beyond the idea stage.

I've done a little thinking on the topic of Nar facilitating CRPG's lately. I think the form they would have to take is what I'm going to call "Binary" Nar. Basically, the player would be confronted by a string of binary choices. Either you pick A, or you pick B. The designer could then map out a tree of possibilities based on these choices.

Do y'all think this type of binary decision making would satisfy the Narrativist player? At first glance, it sounds like it would work. I think that many bang-type decision points in tradtional RPG are often more "binary" than they first appear. What does everyone think?

OK, let me rephrase what I said: it seems to me that simple adventure games fall under the IF umbrella, and it's those that I see as like CYOA books. Adventure and Zork etc. The originals.

Yeah, I'm old.

Sure you can hack these, but it's not part of the inteded game play. That's not to say that I'm against more complex versions of this sort of thing, including the sort of thing that you're talking about, Chris. IF in that case might be very good for this.

But it does leave a question. Which is, now I'm sorta understanding the problem of playing with oneself here. That is, if the game is too maleable, and the player is the one creating the dilemmas for themselves, it's not good narrativism, for the same reason the "Chalk Outlines" is problematic. That is, it's just not as fun when you're creating the dilemmas for yourself to solve. After all, don't you really know where you're going to go with it? Or at least have a sense that you may have known?

How can one create a Bang for oneself? I mean, technically, sure. But will it be effective? There's a neccessary element of surprise that I think is key. That, "Oh, I didn't think that I'd have to make that decision for my character."

Now, if the system can be semi-fixed, and semi-open-ended, then maybe - and I stress maybe - this can be made to work out well. But I have no vision of what this would look like.

Mike

This is the set-up I've been thinking about. The character would have a number of invisible stats. Each stat represents a certain Premise response. For example, you could have 2 stats, Justice and Forgiveness. At pre-set decision points, depending upon the choice made the player, one of those stats would go up. Then again at pre-set moments, the computer would analyze the stats, and depending upon the values, would run one of a few pre-programmed bangs.

You could also have a hiearchy of decision points. At the lowest level, you could have little decisions that don't really change the course of the immediate game. But depending on these minor decisions, the computer would run differernt medium-level bangs. And then these medium level bangs would control your options for the Big Decision points (which would decide major plot branches).

How does that sound?

Sounds good to me.

Mike

Why are major plot branches important?

Do you need the game to branch, in order for it to be nar? Don't get me wrong, the branching would make it sweeter. But is it required? We seem to be running along the idea of 'and then we'd have to have branching plot lines'.

It's cool an all, but much like combat sections in an RPG, shouldn't we consider it optional? Unless I haven't noticed something, and nar really needs this?

Otherwise its not something we need to talk about in regards to faciltating nar gaming. In regards to sprucing up nar gaming, yeah, but not facilitating it.

Mike Holmes wrote: That is, it's just not as fun when you're creating the dilemmas for yourself to solve. After all, don't you really know where you're going to go with it? Or at least have a sense that you may have known?

How can one create a Bang for oneself? I mean, technically, sure. But will it be effective? There's a neccessary element of surprise that I think is key. That, "Oh, I didn't think that I'd have to make that decision for my character."

I know of two games that are pretty Nar-ish, although neither of them are very binary.

There's an old computer game, [URL=http://www.the-underdogs.org/game.php?id=2435]Turst and Betrayal: The Legacy of Siboot[/URL], which does a pretty good job of addressing premise. (Link is to a PC version; an old Mac version is available at the [URL=http://www.erasmatazz.com/free.html]author's website[/URL].)

In the game, you are an alien competing with several other aliens for leadership of a small colony. To win, you need to collect psychic auras by fighting other candidates and taking theirs. Since you can predict an opponent's behavior in combat if you know how many auras they have, the game is mostly about information gathering.
To find information about the other candidates, you have to talk with them, befriend them, make deals and promises, and choose when to betray their trust. Deals always involve trading information about other people's auras. Betrayal occurs when you make a promise (not to attack someone that day, or not to reveal information about them) and break it. In order to get the information you need, you will have to betray someone at some point. You just need to choose who you're willing to betray. And you need to make sure that you place your trust in the right people. There are also little "interstitial stories" that pop up where some text is displayed, and then you're given several choices. Occasionally the game's designer pops up to give hints or comment on your progress.
I've started the game, but never finished it. I tried to play it nice (no betrayals) and had an interstitial story pop up where the designer actually told me I was being too nice. ;)

[URL=http://interactivestory.net]Facade[/URL] is a more recent game that uses some innovative AI to create a short, interesting story game. You're the friend of a married couple who's been invited over to dinner. However, their marriage is currently dissolving, and what you do during the night will affect how it turns out.
The game structure is arranged as a series of "beats", which are small story segments that involve some pre-recorded dialogue and a choice of some sort. Every beat changes the state of the game in some way -- it could raise the tension level (which is tracked throughout the game) or increase your affinity towards one of the two spouses. The AI is responsible for selecting what beat to play based on the current game state. As the game progresses, it increasingly forces you to take sides. The spouses will keep asking you loaded questions, and in later stages, will interpret refusal to answer as taking one side or the other. The AI keeps throwing out beats until it determines that the game state warrants a conclusion of some sort, and then it plays one of the endings.
The game can be played six or seven times, each turning out differently, before there's a noticeable repeat in the beats used. Each play-through takes about 10 minutes.