University of Helsinki Ask Questions Dont be shy, I like to answer - - PowerPoint PPT Presentation

Zach Laster University of Helsinki

 Ask Questions

 Don’t be shy, I like to answer them

 I tell lots of stories (usually Tales from

the Pixel Mines I’ve picked up)

 Bored of the content? Get me off on a

• tangent. ;)

 Please silence your phone

Who am I, and why am I here?

 I have recently completed my Master’s

degree in Computer Science, specializing in Artificial Intelligence and Algorithms for Games.

 I am (hopefully) starting a PhD focusing

• n Procedural Terrain Generation in

Massively Networked Games this Fall.

 I have been working as an instructor

since 2010, and this is my second class I’ve designed and lectured.

 I’m lecturing because I really enjoy

• teaching. It’s fun for me and I love it.

 My goal is to educate. However I can

best do that, let me know.

Ask Questions!

 Yes, I know I already said this.  I specialize in this. I have difficulty telling what

is obvious vs what is arcane.

 I actually allocate some time in my lectures for

you to ask things, and for me to discuss them.

 I know a lot of stuff, and sometimes I just

forget to put things into the slides! If you are interested in a concept ask!

 Worst case scenarios:

 I don’t actually know, and will get back to you

tomorrow.

 I will say “we’ll cover that on day X”

Slides

 I try to keep the information on the slides

informative and direct, and then to present new and more information by speaking.

 I try not to read off my slides.  I try to keep all of the most important

information on the slides.

 Half the time, the slides are there so I

know what to come back to. ;)

Why this course exists

 HY has several courses on Game

Programming, but nothing on game design

 Game Engine Architecture  Game Architecture  Introduction to Game Programming

 We’ve also had courses on specific

topics

 AI for Games 1 & 2  At least one seminar on Game AI

 The absence of actual game design and

balancing classes means that the course projects from these classes rarely play well as games themselves.

 It is an odd hole in the existing teaching

around the department.

 Especially since a number of people

were interested in such a course, when asked.

Intro to goals of the course

 Look at games and what makes a game

fun and interesting

 Look at how balance affects those qualities  Ability to evaluate the balance of a

mechanic in a game

 Ability to construct metrics and flow paths

for balance

 Ability to construct balanced content for

games

 Each lecture, I’d like to include a bit of “homework.”

Largely this will be things you can use immediately to improve balance in your own project(s).

 Rather than cover widely different topics each day, I

will begin with basic concepts and then progress to deeper things each day, spreading the various concepts across all the days.

 By the end of day one you’ll hopefully have a number

• f basic concepts, which we’ll expand on the next

day.

 Some days will focus more on certain avenues than

• thers, like we’ll have a lot of discussion on

prototyping on one of the days, while most of the time I’ll just mention it or make some comments on it.

Grading and Project

 The course does not have an exam  You will work in teams of 2 to 5  It is recommended that you have a team.

Teammates means more personhours and points of view.

 Max allowed group size is 7, and that’s

• nly if you walk into the class as a group.

If you are forming groups here, max is 5.

Report

 The actual grading for the projects will

be conducted using a report system. Each group will write a project report containing:

 Game Concept  Balance Goals  Game Design  How the balance goals were achieved and

any known issues

 Postmortem

Individual Reports

 In addition to the group report, each

member of a group will return an individual report with:

 Personal impressions and lessons  Commented contribution breakdown out of

100%

○ From a total of 100%, give each member of

the group a value for how much work they contributed.

○ Explain why for each percentage. Large

discrepancies especially need to be justified.

Game Definitions

 Math (Game Theory)  Gaming Theory  Software Dev  Intuitive Understanding

Game Theory

 There exists a mathematical definition of

games for the purposes of game theory

 Commonly, this context is interested in two

player games which have certain payoffs for various strategies

 While knowing some things in this area can

be useful to conceptualizing game balance and design as we mean it, it’s not a very good definition for us.

 We’ll draw on a few concepts from here,

primarily in solving games.

 Game theory is a study of strategic decision making.

Specifically, it is "the study of mathematical models of conflict and cooperation between intelligent rational decision-makers".[1] An alternative term suggested "as a more descriptive name for the discipline" is interactive decision theory.[2] Game theory is mainly used in economics, political science, and psychology, as well as logic, computer science, and biology. The subject first addressed zero-sum games, such that one person's gains exactly equal net losses of the other participant or

• participants. Today, however, game theory applies to a

wide range of behavioral relations, and has developed into an umbrella term for the logical side of decision science, including both humans and non-humans (e.g. computers).

• Wikipedia

Gaming Theory / Game Studies

 Essentially, the focus of Gaming Theory is

• n players and player enjoyment. It’s more

the study of why we play games and what roles they fill, where as the mathematical game theory is interested in the solution to games.

 Here we’d define a game as any “social”

interaction with one or more players. Basically it’s something fun to play, and builds from an intuitive sense of games. See also: Gamification

 Gaming theory is commonly used in

 computer science  psychology  sociology  anthropology  philosophy  arts and literature  media studies  communication  theology  etc

Game Development

 When defining games as software, we

are interested primarily in certain

• aspects. Games are a

soft real-time interactive agent-based temporal simulation

 Which is kind of fancy sounding, but a

very good description of games as software

 This definition is taken our Game Engine

Architecture course

Intuitive Understanding

 Games are games, right?  Generally, most people can tell you

what’s a game simply by looking at it. At least, most games.

 The intuitive understanding of games

isn’t strictly useful, as it only lets us say “this is a game” and doesn’t tell us anything about that game.

So what are we looking at?

 For the purposes of this course, we’re kind of

in the middle of these definitions.

 While our games fit better with Gaming

Theory, Game Theory is certainly a strong way to approach balance.

 If you are building your game in software, then

you care about that definition as well.

 From the intuitive understanding, in our case

we really only care that the game is playable. It needs to have some form of defined interactions, but otherwise it’s not even technically required to be fun. Of course, we probably want it to be…

Because games are fun Because games are social Because games permit exploration Because games allow escape Some can even turn a profit (if you are good)

Immersion

 One of the key elements for players of

games is to be immersed.

 Immersion is a state of being “in” the

• game. Immersive games “suck you in.”

 Generally, as players are more

immersed they have more fun.

 If immersion is broken or the player

cannot be immersed, they generally will not play for long.

 Obviously, then, immersion is a

necessary trait in games.

 They improve the fun of games  Lack of immersion will prevent players form

playing even if the game is “fun.”

 Generally, we’ll want to maintain

immersion however we can.

 There are four levels of immersion  Unimmersed, avatar, character, and persona  These terms come from the perspective of story

immersion, but apply fairly generally

 When a player is unimmersed they are not enjoying

the game. They feel no connection to it and don’t care about it.

 The next level is when the player’s in-game self

seems like an avatar to them. A body which accomplishes the goals the player sets forth.

 This is a disconnected attachment. The player is enjoying

the game, but nothing in the game happens to them, and they interact with the game only indirectly.

 Deeper is when the avatar becomes a

• character. The player feels they are

controlling something a bit more alive; with more purpose.

 Last is the persona. At this level, the

player is fully immersed and feels they are part of the game world.

 What do these levels mean outside of

games with controllable characters?

 These concepts map fairly well, though the

terms themselves don’t work well

 Some games may never really aim for such

deep levels of immersion. Some are happy with the player being at the level of avatar

• r character.

 What does it mean to be immersed in a

card game? In a game of chess?

 What produces immersion?

Flowchannel

 The flowchannel is the ideal difficulty to player

skill ratio

 If a player is too challenged for their level of

skill, they will eventually become frustrated or angry, and will probably stop playing.

 Conversely, if a player isn’t challenged enough

for their level of skill, they will often become bored and find something else to do.

 They may continue playing, but there are strong

chances they won’t play normally, possibly trying to break the game.

 This is a serious consideration to game

balance

 When balancing a game, we want to be

able to keep the player in the flowchannel most of the time.

 This does not mean that we cannot create

games designed to be challenge (Super Meat Boy)

 In these case, the expectation for the challenge

is higher; the players expect to fail frequently.

Bartle Types



In 1996, Richard Bartle set

• ut to try to understand what

players find fun by observing what they do, particularly in a setting of online games.



Through observation, Bartle formed a hypothesis detailing four “types” of players.



Killers, Achievers, Socializers, and Explorers



These types were on two axes.

○

Acting vs Interacting

○

Players vs World 

Players are only ever one type at a time, though they may (and do) switch types

○

The most interesting part is how they move.

 In 2003, Bartle

extended the types along another axis, implicit vs explicit, which resolved a number of issues with perceiving some of the transitions and roles themselves.

 These types (called Bartle Types) can be

particularly useful for considering why people play and how content will be used.

 If you consider how each type will

approach content, you can better cater the content to the players.

 This also can enable designers to create

content specifically for certain types of players, in order to encourage this style of play or to ensure that everyone has something they enjoy.

 There are of course other systems for

categorizing players.

 Bartle Types are amongst the most

commonly used.

 This doesn’t technically make them the best,

but they are good to know.

 They are commonly used for at least one

• reason. ;)

Fun

 This is an exceptionally difficult concept to

formulate

 Obviously games need to be fun, but exactly

what makes a game fun is difficult to specify exactly

 Generally, maintaining immersion and staying

in the flowchannel do well.

 I tend o think of being immersed and having fun as

the same thing, as they go hand in hand.

 Designing with Bartle Types in mind will help  Balance helps, but games don’t have to be

balanced to be fun

In order to facilitate things, it’s helpful to define a number of terms. These will mostly be used to describe games or properties of games. Several of the terms come up in different areas, so you might be familiar with them already.

Deterministic vs Non- Deterministic

 Game mechanics can be either

deterministic or non-deterministic

 A deterministic mechanic will always work

the same way every time given the same starting state (input).

 Chess is a very deterministic game

 A non-deterministic mechanic will have

varying results based on some (mostly) random factor.

 The order of cards drawn from a shuffled deck  The result of a die roll

Why do these concepts matter?

 When it comes to testing and balancing,

determinism helps a lot.

 A deterministic mechanic is easier to predict, and

balancing them is easier to do

 A non-deterministic mechanic may mean some

content is undertested if it doesn’t come up often enough.

 In software games, determinism also makes the

game easier to test overall and makes network play much easier.

 Non-determinism helps make the game new

each time

 Without the random shuffle mechanic, most card

games would not exist

Solvable

 A game is solvable when we can know the

correct move to make at each step of the game.

 There are a couple levels to this

 Trivial Solvability (Tic-Tac-Toe)

○ Something close to what a human can solve in

their head.

○ It’s easy to know and remember the solution to the

game.  Theoretically (Chess and Go)

○ A game which could potentially be solved (by a

computer)

○ Checkers is a draw

 Solving non-deterministic games is

basically probability maximization (game theory)

 This allows us to make a predictive system for

the best move to make given current game state

 Human Factor

 The human factor in games cannot be solved  This is what makes poker interesting (bluffing)

○ otherwise it could be solved fairly easily, though

that solutions wouldn’t always result in a win  What humans decide to do can be erratic and

not ideal. Or they could be lying.

Intransitive

 Intransitive basically means games like

Rock-Paper-Scissors

 Who wins depends on choices made by

both players simultaneously.

 Technically, these games are still

solvable, we just have to solve them

• predictively. This launches headlong

into Game Theory!

 In Rock-Paper-Scissors, each move wins equally.

That is, each move has an equal chance of winning assuming the opponent is equally likely to throw any

• f the moves.

 This means we should throw each move an equal

percentage of the time in order to win.

 Additionally, preferring a move means your opponent

can adapt to throwing the winning move, which gives them an advantage.

 Here each win has an equal payoff, but if we were to

change the payoffs of winning with one of the throws, we’d completely change the ideal solution. Game Theory provides us with tools for solving that.

Perfect Information

 A game is said to have perfect

information when all players know everything about the game state

 Chess is a game with perfect information

 “Any deterministic game with perfect

information is, at least theoretically, completely solvable”

 It may take a long time, though (chess, go)

Imperfect information

 Many games have incomplete information

from the players’ perspectives

 They may not know what order the cards will

come up in in a deck, or don’t know their

• pponent’s hand.

 Information that all the players share is

common information.

 Everyone knows the cards on the river in Texas

hold’em

 An information no one knows is simply hidden

information

 Most of the time, no one knows the next card from

the deck

Privileged Information

 Any information which only certain players knows is

privileged information

 A player knows their hand, but no one else should  Players of an RTS share LoS with their allies, but not their

• enemies. What they can see is privileged information.

 How many resources a player has in most RTS games  Sometimes game mechanics either give away some

privileged information or force you to

 Some cards in Magic sometimes make you reveal (part of)

your hand

 In cluedo, this is the driving mechanic. The player controls

what of their information they give out, but MUST give something away if they can.

Symmetry

 Symmetry is where all the aspects of a game

are identical for all the players.

 In many RTS games, mirrored maps accomplish

this, if both players play the same faction

 Chess isn’t actually symmetric

 white goes first  could chess be made symmetric?

 Symmetry is kind of automatic balance for

some respects, but is not a cure all for balance needs.

 Sure, no player has an advantage (or disadvantage),

but that doesn’t mean that the possible strategies are balanced

Meta-game

 The meta-game is the culture and game surrounding

the game.

 Deck building/collection  Availability of game components  analysis or opponent strategies  Must also be balanced  If you fail to balance the meta-game, then the main game

can never be balanced

 Example: A really rare card which is very valuable beats

nearly everything in-game

 This isn’t balanced because some people will still have

access to it, and not others

 The meta-game has granted them an advantage in-game

What is balance?

 We talk and hear all the time about

games (not) being balanced

 But what does this actually mean? Why

is the game balanced or not balanced?

 Why do we care if the game is or isn’t

balanced?

 Basically, game balance is getting the

numbers right to produce the desired "feel.“

 We can adjust the numbers of things until

they all play nicely with each other, literally.

 This can be more of an Art than a Science

 Knowing what to change and how is definitely

something one needs to develop a feel for, though some general rules help.

Why care about balance?

 Because your players will  While there are fun games out there

which aren’t balanced, often they would be more fun if they were.

 Generally, players will have more fun

when they feel like they have a fighting chance to win. If they don’t then the game loses a lot of appeal for them.

What are the numbers of the game?

 If we include things like sports games, surely there

aren’t any numbers?

 Actually, regardless of what the game is or how it is

played, there are usually tons of numbers, they may just not be quite what you’d think.

 In any game, everything has stats.  Stats are any number which matters to gameplay  In a card game, one stat may be how much the card

“costs” to play.

 In a sports game, it may be how fast the player can run, or

how much they can lift

○ RPG stats came from somewhere! ○ Ever seen a baseball (or other sports) card?

 Running with our sports stats concept, if we consider

each player as having stats for things like how often they score and how fast they can run, we can build a team out of really good players

 Technically, these stats are actually measurements, rather

than the sources for determining what happens, but the result is the same

 On the flip side, we also have to pay the players’

costs: their salary

 Really good players often cost more! Balance!  There’s actually issues in the meta-game with sports

games.

 They have mostly been solved by various rules  Can you figure them out and their solutions?

No such thing as perfect balance

 In practice, very few games are perfectly

balanced in every way.

 Often, little things just escape notice  Complex combinations of effects can be difficult

to find.. until thousands of players are playing..

 But that’s fine, because perfect balance is

actually kinda dull

 Overbalanced games - no variety, all moves the

same, not enough player choices

 Basically, perfectly balanced games become too

easy to solve

What defines balance?

 So, we know what balance is, basically, but

how do we draw the line?

 If achieving balance is getting all the

numbers to the right places, how do we know where those places are?

 This largely depends on what the game is

meant to be like

 There can be multiple concepts of balance for a

game

 Some people like more brutal games  Others would rather just play casually  Difficulty settings come into play here!

The balance of balance

 Consider the game Super Meat Boy

 is this game balanced?  is it fun?

 Super Meat Boy is a very challenging

(and unforgiving) game

 And yet the mechanics all seem very

balanced

 I’ve not formally checked this. Perhaps try it!

 How about something like Farmville?

 Really, we can set the bar for intended

balance anywhere we like.

 We should just be consistent  We can even have multiple bars, and let the

player chose the style they want (difficulty settings again!)

 We can also choose a form of imbalance

 If the game is imbalanced particular ways, it will

encourage people to take advantage or avoid that area of the game

 We can use this to subtly control the player

populace or to test new mechanics

 It can also be a point of the game!

Numeric relationships

 Since we’ve defined balance as a system of

numbers, we should look at how that system

• f numbers works.

 This is obviously something we’ll get into more

as we go along.

 The basics of it are that all the numbers in the

game relate somehow, generally through the most important game resources.

 We often tie most numbers back to a single

resource, so that we can balance the entire game against something like HP or gold. We can also use invented resources just for this purpose.

Costs/Benefits

 We’ll get into this more tomorrow.

 For now, lets define the concepts we’ll use.

 First, everything can be reduced to either a

cost or a benefit

 These are effectively opposite pulls on a resource;

they either increase or decrease your resources

 If the balance of an element is too far off, we

will either need to increase its costs or reduce its benefits

 Of course, which one we do depends on

whether we wish to keep the benefits or the costs the same.

 There are four terms which we need for discussing

cost/benefit balance

 Overpowered, Underpowered, Overcosted, and

Undercosted

 You are probably familiar with the first two.  An overpowered item is something where the

benefits outweigh any costs. This is to say that the

• bject cannot be balanced without reducing its

benefit.

 This is in contrast with undercosted, where the

benefit outweighs the currently applied costs on the item.

 People will commonly use overpowered to mean

both of these, but we know better. ;)

 Underpowered is then the opposite of

• verpowered, that the benefit just isn’t

powerful enough to be worth anything; the costs just can’t be reduced any more (minimum cost)

 Overcosted is where an element’s benefit isn’t

worth the existing costs and we can reduce them.

 We define the terms overcosted and

undercosted to help us better describe what is wrong and needs to happen in the game’s balance.

Balancing the Meta-game

 Clearly, deck building and card collection

are part of the meta-game in Magic (and similar games)

 What’s the meta-game in say, a sports

game?

 One of the core aspects to it is the building of

teams

 Who plays for what team? Clearly all the teams

want all the best players.

 How do we prevent the team with the most

money from having all the best players?

 Sports games actually balance the meta-

game by having rules set by body.

 These guys clearly understood games

 American Football includes the following:

 Drafts – Weakest teams choose first from

players who left their teams

 Salary Cap: Players can only make so much.

This prevents wealthy teams from buying them

• ut.

 Player limits: You can only have so many

players on your team.

 These limitations prevent teams from

having all the talent available. They cannot hoard all the best players, and even the least wealthy teams can afford some of the best players since the players can’t make more on another team.

 Combined with the draft, this means that

weak teams generally get strong players every year.

Magic: Balance What?

 Richard Garfield obviously made some

errors when creating Magic. This is rather fair, considering it was kind of the first…

 One of these was that the meta-game was

balanced wrongly. It was believed that rarity was actually a balancing mechanic for card value. We can say safely now that this simply isn’t true.

 As Schreiber points out, Garfield can be

• forgiven. Who’d have thought, at that time, that

people would have spent thousands of dollars

• n a card game?

 TCG balance is also made more difficult

in that it is difficult to “patch”

 A computer game can just be patched with

mechanics changes

 Once a card is released into the wild, the

most you can do is ban it or post errata.

Homework

 What do [immersion levels] mean outside

• f games with controllable characters?

Some games may never really aim for such deep levels of immersion. Some are happy with the player being at the level of avatar

• r character.

What does it mean to be immersed in a card game? In a game of chess? What produces immersion?

 How can Chess be made symmetric?