years, single-player games have been built around the limitations of what an AI could
and could not do. Many action games, for example, have balanced their gameplay by
throwing a large number of fairly dumb AI agents at players. When these same games
became death-match online games, designers typically replaced these dumb AI crea-
tures with human players but, in order to balance the game, drastically decreased their
number and made the players fight for themselves. Whereas there were two factions
previously (the player and the enemies), now there were as many adversaries as there
were players. These simple choices managed to rebalance the game experience fairly
well. But these combat games were still not close to exploring the full potential of
multi-player games, because these human players could just as easily still be really
smart ’bots. In Chapter 9, “Artificial Intelligence,” I talked about how an artificial intelli-
gence can be considered to pass the Turing test if players mistake it for a human. For
multi-player games, the goal should be somewhat different: players could never mis-
take their adversaries for anything but other humans.
To create a distinctly multi-player game experience, it can be valuable to take a cue
from the rich history of multi-player non-computer games, particularly games that
require more than two players. For example, many of the more sophisticated board
games involve negotiation between players via trading and bartering or establishing
alliances. Excellent examples of both mechanics can be found in any playing of seminal
multi-player games likeDiplomacy,Cosmic Encounter, andThe Settlers of Catan.In
many three or more player board games, deciding who the various players will gang up
on at any given time is a key part of the play experience. This can help to fix whatever
balancing issues the game design may have, since as soon as one player gets signifi-
cantly ahead of the other players through luck or an exploit, the other players will
quickly gang up on him. In such games, the personalities of the gamers inevitably come
into play, whether through going easy on the less skilled player who gets upset easily or
a hidden vendetta against a spouse that does not manifest itself except during gaming
sessions.Dungeons & Dragonsand other RPGs allow players to play together in a
group, with the Dungeon Master creating a dynamic game experience for the other
players in a way a computer never can. An online game with players located miles away
from each other and who are most likely real-world strangers will probably never recre-
ate the in-person multi-player experience. Nevertheless, the more successful online
games include components that force players to socialize as part of the game mechan-
ics, and thereby make the social interplay in the game that much richer. This includes
seasoned players coordinating their tactics in team-based multi-player games like
Counter-StrikeorBattlefield 1942, or the recreation of the Dungeon Master experience
inNeverwinter Nights.
Though the potential for players to cooperate should be a key part of almost any
multi-player game, getting players to actually do it is another problem entirely.Battle-
field 1942lead programmer Johan Persson has lamented at how infrequently most
players exploit the game’s more cooperative features, such as one player scouting and
providing coordinates to another player who fires artillery. One way to improve the
chances of player cooperation is to make sure that players who work together have a
significant advantage over those who do not. It stands to reason that if there is a com-
plex cooperative option available that does not provide significant benefit over what one
can accomplish alone, most players will choose to just do it all themselves. An early
Chapter 13: Multi-Player 245