When you build a model of reality, you try to reduce complexity.
Reducing complexity means making choices- and reducing the risk of something unexpected affecting the results of your model.
Actually, it means also reducing the number of parameters- and, therefore, making any evolution in your world more predictable.
But reality is not necessarily limited by your definition: and managing the reality within a model requires more that planning beforehand for what you know, in terms of activities or risks.
You have also to identify what is the “normal” way in which your model will react to unexpected changes in the “reality” surrounding your model.
This post is part of a series, first published in May 2009.