Sunday, February 15, 2009

Balanced Choices

In my last post I glossed over an area because it would have taken a whole essay to explain. An anonymous poster immediately noticed.

So here's the essay.

The anonymous poster... ugh, I hate anonymic, so we'll call him "Herbert". Herbert says
"Not enough deeply disturbing dilemmas, I think. I want to choose the light side but it will mean sacrificing something very important. Usually it's sacrificing a tempting weapon or maybe a nice sum of money."

Okay, this is the "balanced choices" theory of design. This is the same theory of design that KotOR and friends use. But they are starting to discover its limits, and you can clearly see this in Mass Effect. It's not a very good way to design things, although it is useful for another reason.

As you read this essay, please consider the difference in choices between Mass Effect and a KotOR game. I think you will see that they have followed the same logical path I am outlining here.

Let me explain what I'm talking about. In fantastabulous detail, yes.

First we have to point out that we're not talking about a skill challenge. A choice of this nature should never be about choosing the option with greater utility. These choices don't train up a skill (like jumping, shooting, or placing plumbing) and they don't happen often enough to form a complex "terrain" of choices. Furthermore, they are heavily warped by the player's play style and current situation, so any "balancing" you do between the options is going to be totally ruined by what the player brings to the table.

It is possible to make the choice between a long-term gain and a short-term gain, in which case it could really tax the player's skill at analyzing these sorts of things. But frankly, you'll only choose the short-term option if you're doing badly, and it will always feel like you've failed, like you've had to ask for help. It will make you weaker in the long run, too, so that's a nasty positive feedback loop. No, not a good situation.

These kinds of choices are not useful as skill challenges. They are, instead, offered mostly as expressive choices. Allowing the player to choose between a gun and a sword is a very effective choice, even if they are the same, utility wise. Hell, even if one is a bit better or worse.

This is because most players have a strong preference on that front: I'll always take the gun. While it's not a "challenging choice" - the answer is obvious - it does allow me to express myself and start on a path I prefer. Of course, if the game continues to offer me the choice between sword and gun over and over for the rest of the game, they don't count. I've already expressed myself on that front, and unless something has significantly changed the utility of one or the other, I'll just go the same way over and over again.

Self expression is the only thing these sorts of "A or B or C" choices are useful for. But, once you've self-expressed a particular sentiment, there is only a little value in expressing it again unless something big has changed.

Now, to get back to Herbert. Herbert says he wants his preferred side (light side, same as me) to be balanced against a tempting equivalent on the other option (guns and money). Let's quickly examine why this is not a good way to think, although the same end result may be reached from other lines of thought from time to time.

Let's remove the flavor for just a moment. Instead of light side, we'll just make it a "power point". Which it is: you get a point or two of light side for choosing the light side path.

Those points are very personal and long-term statistical improvements. They are very valuable and keep their value forever. Choosing a gun or money, on the other hand, is a short-to-medium-term benefit that will go away as you cycle through various equipment over time. (Actually, money is usually completely worthless in KotOR games, but let's skip that fact.)

This means that, without the flavor, we're talking about that old bugaboo we mentioned, long term choice vs short term choice. Unless we have decided that the points are worthless, choosing for short term is going to feel like failure. It's admitting that we need outside help instead of being self-sufficient with our steadily growing point-power.

Now, adding the flavor back in, choosing light side is emphasized even more.

It may make us feel good that we're "sacrificing" for the sake of our light side choice, but the fact is quite the opposite and, as before, it's not usually an actual choice. It's usually asking us to re-express the same thing we've expressed before: yeah, we're STILL light side.

We could, instead, reverse the flavor. This is probably what Herbert was trying to get at. The idea that we choose flavor or reward.

I think that flavor is, essentially, a long-term reward. No matter how long the game goes on, the fact that we saved that child will never be "replaced" by a "better" act of good. They are cumulative, unlike which gun you equip. So flavor is a long-term reward.

Choosing guns over flavor is, therefore, admitting that you need help bad enough that you're willing to give up on the cool shit in the game to get a boost.

This works in both directions, of course - light side and dark side flavors are equally long-term and valuable. So it's not simply that bad guys get more stuff. In order to get the cool "I burned all their faces off" sequence, you have to sacrifice stuff, too.

You could make it long-term reward versus long-term reward. For example, strictly a flavor choice between light side and dark side (perhaps with a small amount of random statistical crap tossed in so that the player feels like they're affecting their game). But these flavor choices have a dominant side, just like offering me a gun or a sword. I'll always choose the gun and, similarly, I'll always choose light side. It's not a choice except for the very first time.

You can choose between, say, light side flavor and light side points - two tangentially related long-term choices. Except that most players will probably choose in favor of one or the other and always choose that way, just like I'll keep choosing light side after I've starting choosing light side. There may be a little bit of waffling, but I think you'll find it trends towards choosing flavor, because flavor keeps it's value even after the game has ended, so it is longer-term than points. I know I would feel like I was failing if I chose points over flavor. Unless, obviously, the flavor was not enticing to me. Players don't feel bad about not choosing something they find has no merit.


What I'm saying here is that the idea of a scripted choice that is internally balanced is not very good. A player will generally make his choice the first time he sees such a choice, and then continue going that direction forever. If offered choices between long-term and shorter-term rewards, he'll trend towards the long-term rewards. And flavor is a much longer-term reward than any balanced gameplay reward.

The problem is that these choices are always going to be inherently shallow. Compare them to non-shallow gameplay. Non-shallow gameplay iterates rapidly and involves training up your skills. These kinds of scripted choices do not iterate rapidly - they barely iterate at ALL - and they don't involve any kind of skill. So they are not skill challenges.

As ways of allowing the player to express himself, they are good to some extent, but tend to become very repetitive. Even if I have always chosen light side, the scripted choice is always between light side, neutral, and blatant dark side, rather than being chosen to challenge my sense of light side.

Therefore, it is necessary to create a method of extremely long-term options that never ask the same question twice.

That is what I outlined in my last post.


Sigh, I'm trying to figure out how I should have written that first half. It's not very good. Oh, well, does everyone understand what I'm saying? Agree? Disagree?


Ellipsis said...

I agree, although really, in terms of flavor, for most Bioware games the real flavor choice comes at the end, when you choose between the various endings (because you CAN choose any ending at the last minute, no matter how you've been playing so far).

This is something I felt very clearly in Jade Empire - I wish there was some meaningful distinction between the alignments earlier on. Even if you choose the "jerk" dialogue options, NPCs will refuse to not join your party, and people will still insist on giving you quests that seem out of character.

Craig Perko said...

Yes, those are the problems I have with the way these things are currently set up.

Greg Tannahill said...

It's a false dilemma that shouldn't be posed in the first place.

Consider: the choice between short term benefit and long term benefit is not a good way of looking at morality, because it's a skill challenge - there's a right answer and a wrong answer.

So okay, we choose between flavour and power. Which is a bad choice, because the players most inclined to choose flavour are the ones who are going to be most frustrated when the game gets disproportionately hard later. Players shouldn't be asked to choose between flavour and power.

So, okay, we're now choosing between two flavour options that have equal but different power benefits. This is good gameplay but you've taken the sting out of the moral choice; the player gets rewarded no matter what they pick.

So, fine, now we're choosing between two flavour options where the player is also penalised for one option - so now you're back to a skill challenge, where there's a right option.

This is the problem: we don't want what we think we want. We think we want to choose between two difficult paths, but that choice is just as un-fun in games as it is in real life.

The correct way to look at the dilemma is in terms of storytelling. You've been looking at morality in terms of my first two categories of choice: skill challenges and expressive choices. It should actually fall into the third category, the feedback choice. By making morality decisions, a player is telling the game what sort of story they want to be told (or take part in telling). They're saying what sort of situations and interactions they want to be having, and it's up to the game to then deliver. That's why it's important for these choices to come early, be followed through, and be re-asked at key narrative junctions.

Craig Perko said...

I told you that I considered the second and third types to be the same. I believe that an expressive choice and a feedback choice are fundamentally the same.

But, if you must draw that line, then I agree.

Matthew Rundle said...

What if the player has to shift his preference at some point, because the choice he started with stops being the right answer? Is this too painful because it wastes the points he spent on the other option?

Craig Perko said...


It would be a problem if the situation was set up to be on this kind of linear, zero-sum scale. Which is a large part of the reason that games such as KotOR let you completely change your alignment whenever you want - it reduces the problem of going too far down a path you don't end up liking. At least, that's the theory.

But in a situation where we are not using a linear, zero-sum progression, going a different direction has a strong feel to it. It's not that you're discarding your past or working against it, it's that you're building off your past in a new direction.

If built correctly, this would allow for a unique and dramatic result when you decide to change directions, rather than "discarding" or "working against" your past.

Matthew Rundle said...

I think I mean, do you want to make it so that the player has to change his mind at some point, in order to progress - stack the game so that this is the case? You want to make it so that you can't just keep making the same decision throughout the game.

Craig Perko said...

Okay, look at it this way: in Quake, you can go left, right, forward, backwards. And you can turn.

But when people play, they don't insist on always going left, or always turning. The way the world is laid out - and the way they are facing - mean that the meaning of their choices always changes.

Matthew Rundle said...

But there's no left-turning skill that you level up with each successive left turn. The choice of whether to go left or right depends wholly on the level design: the player isn't invested in turning left, and feels no specific drive to avoid turning right. The space of possible values the player can express is inherently different, specifically because it's a space of values that appeal to the player or don't.

Craig Perko said...

(This is Craig at Work)

I should probably do another post to be clearer, but what we're doing can be compared to level design vs. level scripting.

The current way we're doing things is to offer scripted choices, very linear, very low number of branches, very coarse-grain choices.

What we should be aiming for is something like a level, where the player makes many fine-grain choices to change their location in a less "boolean" way.

There will ALWAYS have to be some static design involved, even if you abstract it out to the thing that generates the things that generate the levels. The idea is not to get rid of the designer: the idea is to let the player explore the dramatic and personal space in a high-grain way.

It's an alternative to the idea that the designer tells you exactly what you must experience, how.

Ellipsis said...

I think part of the problem here is that the situation is framed in terms of moral choice.

Moral choices are meant to have a correct answer (especially if you're giving them feedback in the form of points), and if it's not clear to the player what that correct answer is, they'll feel like they were tricked into getting dark side/evil points.

So you have two options (or three if you count "neutrality", but I don't), and only one is "correct," and it's supposed to be obvious what the right answer is. That's extremely uninteresting gameplay, though it will often make the player feel good about themselves/their character.

Making it about personalities has a lot more room to be interesting (because there's more axes, and one option might not be clearly better), which seems to be what Craig's getting at.

Craig Perko said...

I guess that is what I'm getting at, but I find that there are moral choices with no "right" answer, or at least with no right answer everyone can agree on. So I don't think it's a requirement to do away with moral choices (and simply have moral fiat).

However, it is certainly important to do away with moral duality. A "moral axis" is just a bad idea.

Anonymous said...

I think ellipsis has the right idea. The problem with this whole argument is that you are trying to dumb down every option into two choices, good or bad, when in actuality the questions are slightly more high level than that. I think a better way to put it is, "how do you want to be percieved?". In Mass Effect, basically every option is asking you how you want the rest of the world to view you, and there they do a, relatively, good job of breaking each suggestion down to a binomial range, either your a badass hero or a badass rebel, but badass nonetheless. I liked this option, but i simply wished there was a better range of options, rather than just heroic leader or aggressive rebel.

There needs to be a system that better reflects the different nuances of a given situation, and the possible notions your choices say about you and how you want to be perceived. I think thats what you are going for with you range of axes, does this choice say i value freedom over captivity, do i value love over hate. Basically your asking the player, what do you want more? Would you be rather viewed as a loving person at the cost of being called gullible? Are you willing to sacrifice your humanity (being percieved as good) for knowledge? I feel like those are broader questions that can be addressed, and I do feel that you can find a way to break down a situation and ask, well what does this force the player to decide between?

Craig Perko said...

I thought that way too, for a while, but I think that it's thinking too simplistically. You'll never manage to script in broad enough choices to satisfy everyone, and you can't script densely enough to make choosing like that interesting from a gameplay perspective.

We need to start thinking of these kinds of personality choices as being a landscape instead of a railroad with switches. Only by reframing the whole situation can you give the player a feeling of freedom. Otherwise, he'll always feel railroaded and not be able to find the exact option he wants, no matter how many options you give him.