Game Theory Discussion: Player Knowledge Vs. Character Knowledge

[Sinitrena]

The following are just some thoughts I had while playing a lot of games lately. I tried to structure them, but ... 
---------------------------------------
---------------------------------------

Game Theory Discussion: Player Knowledge Vs. Character Knowledge

When designing a game, you need to consider various things, including story, puzzles, general looks, and so on.

What people normally don‘ t think about in these terms is the difference between what the character knows and what the player knows. This problem certainly shows up when designing, but I think it might be helpful for some people to look at it a bit more closely.

So, what do I even mean with Player Knowledge and Character Knowledge?

To put it simple:
What does the player know in any given situation during the game?
What does the character know in any given situation during the game?

These questions are similar to a watsonian and a doylistic analysis of literary works.
---------------------------------------


Now, why is this important?

When character and player knowledge do not align, it can easily lead to frustrations. For example, when the player knows how to solve a puzzle but the necessary trigger for the character hasn‘t happened, the player might run around forever trying to find a different solution. Or when the character knows something, and acts accordingly (for example combining inventory items even though the player was just randomly clicking) this might lead to confusion.
-------------------------------------------------


There are generally three possible states:

1. The character knows more than the player:

From a storytelling viewpoint, this might show up when you give your character a specific profession (e.g. Indiana Jones, who is an archaeologist). From the get-go, you should assume that the character knows things the player does not.

This might not seem like much of a problem. You just need to make sure that the internal thoughts of the character (aka speech bubbles for the player to read) convey all these informations.


Another possibility this comes up is when the game takes place somewhere the character is familiar with (e.g. various Maniac Mansion Mania games). In this situation, you have the player explore their surroundings, even though the player certainly would not do this.

I don’t see much of a problem with this in general either. Generally the worst consequence is that there might be a bit of an immersion break.


More knowledge of the character only becomes a problem when it involves the solution to puzzles. I’m sure we’ve all seen situation where we were a bit stuck playing and randomly tried to combine inventory items. Sometimes, this leads to a solution and often to an explanation from the character of the clever trick they are going to play.

The problem here should be obvious: A clear miscommunication between the knowledge of the character and that of the player. The solution would be to give more hints.


2. Character Knowledge and Player Knowledge align:

I would say this is the ideal state, but hardly achievable. You can never know what your players might know.


3. The player knows more than the character:

No player goes into a game as a blank slate. There’s no way to know beforehand what they know, but I think one should consider what can be seen as general knowledge or more specialized knowledge and never use too specialised information to solve a puzzle.


What might come up is that the player knows the solution to a puzzle before they are actually supposed to solve the puzzle (e.g. knowing the code to a safe too early because one played the game once before). There are three ways to react to this: Allowing the player to proceed with their “cheating”, or force them to do all other actions first. Either solution can easily lead to a break in the immersion. I don’t have an opinion here.


The most frustrating situation is what I already mentioned above: A player figures out a puzzle but this one trigger is missing for the character to figure it out. The only way to avoid this frustration is by making sure the triggers are logical.
------------------------------------------

All in all, it seems important to me to align character and player as much as possible.

There are some general steps I think might be helpful:
- make sure your character conveys all knowledge they have to the player (textboxes, detailed descriptions, clear designs, dialogue options)
- avoid the necessity of outside knowledge (a player should never have to look something up to solve a puzzle)
- try to make triggers for events logical
- keep in mind that players might take shortcuts
- a break in the immersion is better than frustrating a player.
---------------------------------------


These are just some thoughts and I’d be interested in yours.

(I avoided specific examples, mainly because this is long enough as is, but I sure wouldn’t mind to read a few.)

[TheFrighter]

Sometime I prefer that he character knows more than the player, it's part of his charisma. Whitout a bit of knowledge often I have the impression of leading a stupid puppet (that could be good for a comedy game, less for a drama).

This make remind me of the game Daughter of Serpent, where you can choose the knowledges of the character and this affect some aspects of the game.

_

[Creamy]

The player knows more than the character:
No player goes into a game as a blank slate. There’s no way to know beforehand what they know, but I think one should consider what can be seen as general knowledge or more specialized knowledge and never use too specialised information to solve a puzzle.

I don't see a reason to forbid the use of a specialized knowledge to solve a puzzle, provided the information can be found somewhere in the game. Learning is part of the fun.

What might come up is that the player knows the solution to a puzzle before they are actually supposed to solve the puzzle (e.g. knowing the code to a safe too early because one played the game once before). There are three ways to react to this: Allowing the player to proceed with their “cheating”, or force them to do all other actions first.

That makes two 
I think the second approach (force them to do all other actions first) is more common nowadays because developers prefer to avoid frustrating the players.
They'll often keep you from trying a puzzle if you can't solve it yet.

The first approach (Allowing the player to proceed with their “cheating”) can be seen in the White Chamber. You can type coordinates on a device. The character must find them somewhere else first but she won't tell you. If you type the correct coordinates from the get-go, it triggers an easter egg.

[Mandle]

Another case I would like to add is when you have two characters to control who can be in different locations. Imagine a puzzle depends on, let's say, both characters holding down switches in two separate locations at the same time.

In this case, the player knows how to coordinate the timing, but the individual characters do not as they cannot communicate with each other from the two locations.

This really breaks immersion. The obvious solution is to give the characters some in-game means of communicating with each other at a distance via walkie-talkies, phones, mental telepathy, etc.

If this kind of constantly available communication spoils other puzzles or plot points in the game then a convenient intercom next to each switch might work if logical within the story.

Or the characters could have agreed on a predetermined time at which to both pull the switches.

The only other option really is to break the 4th wall and have one or both of the characters say "I have this unexplained feeling that the other guy is pulling the other switch right at this moment. Weird, huh?"... which can also work as a meta-joke if it's that kind of game.

But simply leaving the issue completely unaddressed might make the game designer look lazy and sloppy.

The switches puzzle is just a simple example but the concept applies to all situations where two separated characters, controlled by the same player, are both working towards a goal together.

One more plot-device I just remembered is from the movie "Hudson Hawk" where the two characters start singing the same song when together and then split up, each continuing to sing the song. They know from the timing of the lines in the song when to do separate actions to work together as a team. This is quite charming and also gives their relationship a feeling of having worked together for a long time. And it can probably be stolen because almost nobody saw "Hudson Hawk" LOL.

[Click'd]

I have a rather weird example from an obscure game called Void and Meddler:

At one point you have to play a series of four notes on an organ. Problems as follows: The notes are heard in the background but somewhat masked by a constant hum. Also, the keys make completely different sounds from what you hear, but apparently people who know music can "translate" the stuff.

How does that fit here? The player character is a musician and expected to know such stuff. Who really has to know it, though, is the player, because neither is there any hint within the game, nor is it a small amount of keys so you could brute force it, but an almost full-size keyboard.

Educated guess: Since the developers are musicians first and game makers second, they forgot that not everybody shares their background or ear for sounds.

[Sinitrena]

The player knows more than the character:
No player goes into a game as a blank slate. There’s no way to know beforehand what they know, but I think one should consider what can be seen as general knowledge or more specialized knowledge and never use too specialised information to solve a puzzle.

I don't see a reason to forbid the use of a specialized knowledge to solve a puzzle, provided the information can be found somewhere in the game. Learning is part of the fun.

Yes, it is very important that necessary information can be found in the game.

But I was thinking more of situations like: "I know that I have to magnitize the needle to built a compass. Why do I have to look it up in the in-game library?"
When does something become specialized knowledge that needs the information provided? When do you make it required to aquire the knowledge in-game? When can you leave out information?
These are design questions. I don't think there's a definitive answer. It's just something to keep in mind.

What might come up is that the player knows the solution to a puzzle before they are actually supposed to solve the puzzle (e.g. knowing the code to a safe too early because one played the game once before). There are three ways to react to this: Allowing the player to proceed with their “cheating”, or force them to do all other actions first.

That makes two 
I think the second approach (force them to do all other actions first) is more common nowadays because developers prefer to avoid frustrating the players.
They'll often keep you from trying a puzzle if you can't solve it yet.

Yes, that's two. Sorry, I meant to split "not allowing" into two options:

1. The circumvent method: Not having access to the room with the safe, for example.
This can be very elegant and even engaging for the player.

2. The brute force method: "Hm, the code doesn't work...", despite it being correct.
I would never say this is a good method, but I've seen it in a couple of games. It's frustrating and it often makes you think you have the wrong solution despite it being the right one, just not the right time.

Another case I would like to add is when you have two characters to control who can be in different locations. Imagine a puzzle depends on, let's say, both characters holding down switches in two separate locations at the same time.

Oh, it gets really complicated when you add another character into the mix.
Do you forbid players to use knowledge one character gathered for the other character? Or do you treat them as one pool of knowledge?
Immersion-wise, treating them as two seperate entities is better, but less frustrating for the player is probably a shared knowledge. Very interesting possiblities. 

I have a rather weird example from an obscure game called Void and Meddler:

That's an interesting example and one where the communication between developer (through the characters) and players seems to have failed.

That's always the key aspect: Communication.
You have to let your players know what they need to know without making puzzles too easy or completely breaking the immersion.


Wow, we already have a nice little discussion. I'd love to hear more opinions.

[heltenjon]

Interesting topic! The old Sierra games had quite a few sudden death situations where the only possible way through was to learn from dying and try again until you found the correct way to avoid death. I reckon the idea came from action games, where this kind of learning gives the player an advantage in the next game. In adventure games, however? Compared to the action games, where you could have to beat ten levels before losing all your lives against the same enemy, the Sierra games at least had a save option. But it's clearly an example of the player gathering knowledge the character doesn't know. Not the type of puzzle I prefer.

We've had this discussion in our home recently. My kids helped out as playtesters for BYO Hero, and when they restarted, I had to explain why they couldn't do certain actions yet. "You know that, but Coco (the character) doesn't know it yet!" In that game, Coco has a sort of amnesia. The players will probably understand what has happened before she has gathered all the pieces. It's not a problem, though, just makes the kids feel smart.

Mandle makes a good point about controlling different characters, but on the other hand, Maniac Mansion and especially Day of the Tentacle are exactly like that, and I can't really say it's a problem. Perhaps because they are comedies?

Is immersion necessarily the same as identification with the player character? I'd say no. Larry Laffer makes his one-track observations on his own. Without his goofy personality, the games wouldn't be as fun to play. Gabriel Knight misses several of the hints a German-speaking player will catch, but it's still fun to find the clues together with him.

Another option is the unreliable narrator. This opens up for subjective ways of showing the surroundings. In these cases, the player (probably) knows that the world isn't how the character sees it, because we know that he/she is under the influence/infected/hallucinating in the desert/in denial/whatever. This can be pretty great, but it's not a common trope in adventure games, I think.

[Mandle]

Mandle makes a good point about controlling different characters, but on the other hand, Maniac Mansion and especially Day of the Tentacle are exactly like that, and I can't really say it's a problem. Perhaps because they are comedies?

Is immersion necessarily the same as identification with the player character? I'd say no

Interesting! I'd say my personal feelings when playing those games are that I'm watching a comedy show that I'm in control of. So, yes, I feel that's a reason why it doesn't feel weird. I feel that I'm watching something artificial, but entertaining, from the start. I don't feel like I am the character that I'm controlling though.

In The Dig, for example, I feel like I'm Boston Low, and that I'm in the situation myself. If I remember correctly, that game takes measures to avoid the "that character doesn't know that yet" problem by having the characters meet together and discuss what they have independently discovered.

Slightly off topic but I was explaining "immersion" and "becoming the character" with a mostly non-gaming friend. He only plays strategy games and even skips the cutscenes if there is any kind of story involved. He thought the idea of feeling you are the character in the game is ridiculous and was making fun of me a bit so I asked him:

"Imagine you are playing a Mario game and you say "Ah, shit!" and your roommate says "What?". Do you say "I died" or do you say "Mario died"?"

He conceded that I had a point.

[Cassiebsg]

One more plot-device I just remembered is from the movie "Hudson Hawk" where the two characters start singing the same song when together and then split up, each continuing to sing the song. They know from the timing of the lines in the song when to do separate actions to work together as a team. This is quite charming and also gives their relationship a feeling of having worked together for a long time. And it can probably be stolen because almost nobody saw "Hudson Hawk" LOL.

Pretty sure, I've seen it... but don't remember it. 
But I do remember that in Battlestar Galactica (1978  ) there's an episode where Starbuck does exactly that, to coordinate an attack with a bunch of "warrior kids"...  The episode isn't great, even though it has some good scenes and brings in another marvelous IL serie cylon, Spectre (priceless)! I've always hated the stupid song, but it worked well for the situation.

[Danvzare]

Here's my two cents.
You should try to align player knowledge and character knowledge as much as possible, but not at the cost of playability.
For example, on the original Deus Ex, you can learn codes to keypads and use them on future playthroughs. The player can get and use knowledge that the character doesn't have.
But on the PS2 port as well as the the sequel Deus Ex Invisible War, you need to have seen the code for the character to input it automatically (therefore preventing you from skipping sections). The player and the character's knowledge are forced to be aligned.

Now if you've ever played those games, you'll know that the way the original PC version of Deus Ex did it, was more fun (clearly the developers of Deus Ex 3 and 4 agreed). So it's a bit of a balancing act. You can't be too limiting just because you're trying to follow what's considered optimal game design.

Now with that all being said, you should never require player knowledge and character knowledge to be misaligned. For example, one door leads to death the other doesn't. The only way to find out which is which is by saving, trying a door, and if it's the death door, reloading and trying again. There's no way the character could ever know the answer.

[Cassiebsg]

With passwords or codes, it should be possible to always enter them. After all, the player or the character could just decide to brute-force it, or even be lucky. If the dev. doesn't want it to be brute-forced (make the code longer than 4 numbers/characters... or generate a new code at game start). And if the player does enter the right code/password without having gotten the knowledge, one could always add a "what a lucky guess!" comment or something to that effect. 

Otherwise, all specific knowledge that might be required to do certain tasks should be available in game.

If the character is a chemist he/she should know how to mix stuff that we as non-chemics don't. Then the character should have that info ready to provide the player, by commenting.

If it's something the player knows but not the character, then the info should be available for the character to figure it out but still let the player use his knowledge to solve the problem (cause the player is suppose to be (or identify with) the character). It's very frustrating as a player, IMHO, to know the solution to the puzzle and not be able to do it, cause you haven't looked it up/discovered the info in game, yet. Also this will let the player feel good and smart, as he just solved a puzzle without the need of inside help/hints.