In a game there are usually some rules to winning:

* The last event which won a game has to complete the winning pattern.
	... This makes sense, because the last person wins a game by performing the last move
* The pattern is a pattern which the loser does not have.
	... This also makes sense, as if the loser had the winning pattern, he or she would have won already.
* The pattern will almost certianly work in places where each piece in the pattern could have the exact same attributes.
	... If in the game of chess, the board was mirrored down the vertical axis, the position of the pieces would be different,
	but the attributes would be broadly the same, therefore it would be a winning pattern, given the same attributes.

Defining characteristics of a winning pattern can include:
	* The amount of pieces in the pattern
	* The positioning of the pieces in the pattern

To write an efficient algorithm, the computer will have to recognise transformations,symmetry ect. by itself. This would be a form
of generalisation in artificial intelligence.