With such simple rules, it seems difficult to believe that complicated patterns could emerge from the Game of Life, or any other set of rules for cellular automata. However, it turns out that the outcome of a particular starting configuration has a far from predictable final state. While each automaton does follows the same simple set of rules in Life, those rules call for it to interact with its eight surrounding automata. When considering that for every one cell, there are eight possible interactions that determine the next state, it begins to become evident why an outcome would be difficult to foresee without stepping through all of the periods on the way to reaching the final state.

For example, consider a traffic jam. The simplest explanation for a slow down would be a traffic accident or a slow-moving vehicle that restricts the flow of other cars. However, many times traffic jams have no easy explanation. In reality, the cars cause each other to move slowly because more are trying to move in the same direction than the capacity of the road. Much like in the Game of Life, the movement of one car is affected by the movement of the cars surrounding it. Those cars are in turn affected by cars around them, and so forth. Thus, a seemingly simple phenomenon snowballs into a complicated mess and each lane change, acceleration, or deceleration adds unpredictably to the intertwined confusion.

The unpredictability of the final state in the Game of Life is what allows it to have many applications of a non-trivial nature. When creating the game, Conway carefully considered the rules in order that the game have a delicate balance between having all of the cells die out and having them replicate endlessly. His result is such that it is thus far impossible to come up with an algorithm to accurately predict the final configuration given the initial conditions of the board without going through the intermediate states. This idea will come up again when discussing the computational possibilities with the Game of Life, but is also useful in an area where computers often fall short: generating pseudo-random numbers. By assigning values to the configurations, the intricate patterns of Life can be used to generate numbers with a high degree of randomness.