Strategic decisions deal with the way an organization or system of organizations moves into the future through a changing environment. They relate to the very purposes and essence of the organization itself, and of its human, material and financial resources. They involve a choice of objectives, a contrivance of means, and both of these involve an assertion of will rather than responses deterministically derived from what has happened previously. They are purposive thrusts into the future rather than decisions directed by testable logic or continuity of circumstances. Such decisions rest largely on judgement, foresight and imagination.
Game theory provides a unifying viewpoint for all types of conflict situation (whether military, political, or economic), and gives two meanings to strategy. A pure strategy is a move or a specific series of moves by a participant in a conflict in which successive gains are clearly delineated. A grand or mixed strategy is a statistical decision rule for deciding which particular pure strategy a participant should select in a particular situation to improve the probabilities of a desired payoff.
A weak and a strong definition of strategy may be distinguished: rationalization actively effected to order the totality of (internal or external) elements which may play a role in the determination of the general orientation of a social entity through its environment; or the rational organizing and directing function of the totality of forces (resources and systems, which are not all entirely or constantly mobilized) of social entities in their reciprocal negations (more or less focused and not necessarily correlated and equivalent). Essentially strategy is negation through the imposition of constraint on the freedom of action of the opponent.
Strategy is contrasted to tactics, which is a specific scheme for employment of allocated resources. It is also contrasted to policy which is a contingent decision, whereas strategy is a rule for making decisions. Specification of strategy is forced under conditions of partial ignorance, when alternatives cannot be arranged and examined in advance. In the case of policy, under conditions of risk (alternatives and their probabilities known) or uncertainty (alternatives known but not their probabilities), the consequences of different alternatives can be analysed in advance and decision made contingent on their occurrence.