There are two separable issues in the ideal-state view of system administration. The distinction concerns the perceived intelligence behind the changes which lead to a degradation of the ideal state. We may classify changes as either random (stochastic) or as intentional (strategic) depending on the nature of the adversary.
This distinction is partly artificial: all changes can be traced back to the actions of humans at some level, but it is not always pertinent to do so. Not all users act in response to a specific provocation, or with a specific aim in mind. It just happens that their actions lead to a general degradation of the ideal state, no malice intended. This strikes back to the fundamental principle of detail, namely that high level effects wash out the specifics of low-level origins. Thus there is a part of the spectrum of changes which averages out to a kind of faceless background noise. The details of who did what are of no concern. Random influences have been analyzed in ref. [2] and are found to follow a number of well-known statistical distributions. Their study is part of the problem to be solved, but not all of it.
The other part of the problem is the case of actions which may be
regarded as being more carefully calculated, or following a
systematic behavioural pattern. These are caused by conflicts of
interest between system policy and user wishes. A suitable framework
for analyzing conflicts of interest, in a closed system, is the theory
of games[10,11].
Game theory is about introducing players, with goals and
aims, into a scheme of rules and then analyzing how much a player can
win, according to those restrictions. Each move in a game affords the
player a characteristic value, often referred to as the `payoff'.
Game theory has been applied to warfare, to economics (commercial
warfare) and many other situations. In this case, the game takes place
on the n-dimensional board, spanned by the 
vectors.
There are many types or classifications of game. Some games are trivial: one-person games of chance, for example, are not analyzable in terms of strategies, since the actions of the player are irrelevant to the outcome. In a sense, these are related to the first kind of deviation referred to above. Some situations in system administration fit this scenario. More interesting, is the case in which the outcome of the game can be determined by a specific choice of strategy on the part of the players. The most basic model for such a game is that of a two-person zero-sum game, or a game in which there are two players, and where the losses of one player are the gains of the other. `Zero sum' is the law of conservation of currency (current).
Many games can be stated in terms of this basic model, although this is often a simplification of reality. Games in complex systems are rarely true zero-sum games: energy leaks out, money gets burned or printed and thus there is no exact zero-sum conservation.
