Probability magic is a new magical model that looks at the universe as a set of probabilities, and as magic being the ability to increase and decrease these probabilities. This is different from traditional magical theory.
Traditional magical theory worked under a cause and effect theory, or action and reaction. In other words the basic idea of magic was that if you do A then B will occur, and so the point of magic was to assign something to the variable B and then figure out what A is. The problem with this model is that there are a lot of variables that effect the operation. Things like the mood of the practitioner, who is in the room, where the operation is being performed, the time of day, the position of the stars, and the physical condition of the practitioner can all have huge effects on whether or not a spell succeeds. So just because A worked once doesn’t mean it will always work, and just because A didn’t work this time doesn’t mean it will never work. Some magicians who are able to succeed at a spell very rarely succeed at it while others are able to get the same spell to work almost all of the time and rarely ever fail.
Probability magic tries to fix this model. Probability magic begins by supposing that all things are possible however unlikely (this is supported by mathmatics and physics). Next it figures that magic is simply the act of changing these probabilities, and this, not variables which are unaccounted for, is why some magicians can only get a spell to work some of the time. Under the theory of probability magic the spell works every time, but it never brings forth the goal. The spell just makes the goal more likely to happen.