Why Use Simulations?

There are many reasons to use computer simulations rather than real-world experiments to prove things about a situation. Experimenting in the real world may be:

• Expensive, such as testing a proposed new airplane that might fall apart in strong winds. A simulation of the airplane and wind can eliminate some bad designs before building and testing a real airplane.
• Time-consuming, such as testing the effects of a genetic mutation in a species across generations.
• Dangerous, such as testing whether a nuclear reactor will survive an earthquake.
• Unethical, such as giving a population a disease to test how fast it spreads.

Computer simulations rarely capture the full complexity of real situations. For example, simulations of global climate change should account for hundreds of interconnected factors such as wind patterns, the course of rivers, geological fault lines that cause earthquakes, and interactions of local plants and animals. However, because of limitations in computer speed, researchers leave out details to make simulations practical.

There are risks in making simplifications. For example, there might be an interaction between elements of the real situation that isn't obvious but turns out to be crucial. A simplification might lose the details that would have shown that interaction. Another issue is that if you have ideas in mind about what's going on, you are more likely to emphasize details—even unconsciously—that support your ideas. (This kind of distortion is so common that researchers have a name for it: confirmation bias.)