Schedule
:: Lesson 8 :: Track A
Summaries :: Track
B Summaries
Crawford, C., Chapter 6: design techniques and
ideals
Reading summary/quotes:
In this chapter Crawford explains important design techniques to
increase the player’s motivation and desire to play the game. He
discusses AI, or what he terms “artificial smarts,” and explains the
importance of a game producing reasonable behavior and unpredictability
at the same time. He also explains way to balance games. Triangularity
is also discussed as a way to make games more interesting and yet keep
them balanced. Triangulation is a basic form of indirection,
introducing a third party to the game. Crawford gives several examples
of games displaying various levels of indirection. Learning curves are
explained as “a series of related games (p. 12)” which, ideally, should
transition smoothly during game play. The author ends by discussing
winnability, describing it as an illusion because it must appear that
the game is winnable at all times to all player levels, yet the game
must avoid actually being winnable or its appeal will quickly be
lost.
Requirements for an “artificial smarts system”:
- Should produce reasonable behavior, patterns that should be
somewhat predictable (p. 2)
-
- I.E. no shocks, computer should not allow it’s characters to pass
through walls, fly, etc. unless that is part of the game
- Should be unpredictable (p. 2)
-
- While still being reasonable, computer should be hard to
anticipate
- Computer should focus on a single aspect of a larger pattern
- Better algorithms look at decisions in the largest context
possible, which allows them to take in a lot of information in a single
rule
-
- A point system
-
- Give each kind of move a certain amount of points, which no two
moves having the same value. The move with the higher point value is
the one executed
- It is hard to accurately assign points to moves, it requires a lot
of experimentation
- Field analysis (p. 3)
-
- The computer can calculate a large variety of factors of nearby
opponents, obstacles, etc. and then make the analysis as to the best
course of action
- Change the game (p. 4)
-
- If you can’t find some way to make a certain feature work with the
AI, then just take it out. It will save trouble in the end
- Hard to coordinate moves of many units at once
-
- Use an array of “virtual moves” where the computer can plan its
moves until a successful pattern is found. Then the moves are actually
carried out
- To make games more interesting, there should be multiple
algorithms, with the computer transitioning between them. The
transitions should be subtle, however
- The above works great for special systems, but for not special, use
differential equations
-
- Use a damping factor to make values more interesting (p. 5)
- Even with all the algorithms available, the computer must still be
at least twice as powerful to stand up to a human opponent
Limited Information
Limit the amount of information the human player has access to.
Because humans are superior processors of information, this can
handicap them effectively.
- Don’t just randomly limit or this is frustrating.
- If limiting is done well, it can engender the curiosity of the
player
Pace of the game
You can also speed up the pace, i.e. the computer can calculate
faster than the human. This deprives the player of the time they need
to figure things out. Don’t overuse this technique, though. (p. 6)
Summary of above 4 balancing techniques
Use a small amount of each type if possible.
Relationships between opponents
- Symmetric relationships
-
- Each player has the same strengths, attributes, weaknesses,
etc.
- This does provide for great balance
- This can make the game overly simple. Any good strategy will be
used by both sides. Subtle advantages can provide great rewards (i.e.
like having an extra pawn in chess)
- Asymmetric relationships
-
- Each player has a distinct combination of many attributes,
advantages, and disadvantages (p. 6)
- These combinations must be carefully balanced so that both sides
have an equal chance of victory
-
- Can be done with “plastic symmetry”, where the games start
symmetric, but players can choose different combinations that change up
the situation. The player chooses whether the situation is balanced or
not (p. 7)
- Game can start asymmetric
- Indirection (Triangularity) (p. 7)
-
- Non-transitive games, i.e. rock paper scissor. Just because one
combination holds to win, it does not mean that it will hold for other
combinations
- Allows indirect methods of approaching and attacking, i.e. a player
can use a mix of offensive and defensive strategies in a conflict with
an opponent
- Actors and indirect Relationships (p. 8)
-
- An actor is not a direct participant or opponent. It is a
third-party indirect agent that the player can work through to attack
the opponent
- Indirection makes for more interesting games, especially when no
none side is directly associated with the players
Smooth Learning Curves
This refers to a player’s score as a “function of time spent with
the game”, i.e. the score should go higher the longer the player plays
the game (p. 9)
The ideal curve should always slope upwards smoothly. If the curve
ever drops mid-game, it means the game is contradictory (p. 9)
In order to keep this smooth curve, the computer ideally should
adapt to the player’s skill level by increasing difficulty. Another
method is to allow the player to choose their difficulty level (p.
9)
The Illusion of Winnability
The game should appear to be winnable, but not be truly beatable or
winnable, because then players will lose their motivation to play.
Pac-Man is a good example because you appear to win (by eating all of
the dots), but you never actually do because the game just keeps
getting more difficult (p. 10)
A good game encourages players to experiment; it should not give
away the fact that you can’t ultimately win it.
If a player keeps dying in the game, or losing, he or she should
attribute it to their own errors, which are correctable. If this is the
case, the player will want to play again, correcting their past
mistakes. This losing should not be attributed to bad game controls, or
an impossibly hard game. (p. 10)
“Another important trait of any game is the illusion of winnability.
If a game is to provide a continuing challenge to the player, it must
also provide a continuing motivation to play. (p. 12)”
“A smooth learning curve is worked into a game by providing a smooth
progression from the beginner’s level to an expert level. This requires
that the game designer create not one game but a series of related
games. Each game must be intrinsically interesting and challenging to
the level of player for which it is targeted. Ideally, the progression
is automatic; the player starts at the beginner’s level and the
advanced features are brought in as the computer recognizes proficient
play. (p. 12)”
In summary, each game designer must use all of the above techniques
and skills masterfully to come up with their own individual technique.
This is much like artists who develops their own personal style of
technique and form.
Related articles/class discussions:
- Article name (Lesson 7 Track A): This article relates to Crawford’s
last chapter on game design sequence. In this chapter he expands on
some of the essential elements necessary to motivate players and
provide a continuous challenge as the game progresses.
- Chapter 8: The Internal Economy of Games and Game Balancing (Lesson
7 Both Tracks): Intransitive relationships are like the triangular
relationships mentioned above. This article also talks about symmetry
and balancing between opponents.
- Class discussion: This relates to the development of the SRA game.
The focus should be more on design less on the code to end up with a
worthwhile game.
Discussion points/questions:
- Why is so important for the player to perceive that the game is
winnable at all times?
- What are some other ways in which to give the illusion of
winnability?
- What are some techniques for the computer to adapt difficulty level
as the player increases in skill?
Contributors: Tom Caswell, Marion Jensen, Jennifer Jorgensen,
Jon Scoresby, and Tim Stowell