Experience leads us to suppose that events occurs in chains of cause and effect.

A car starts to move because of a chain of events that starts with the key turning in the ignition and ends with the wheels starting to turn. All the intervening steps are directly caused by the key turning, or a consequence of some other event that is itself a consequence of the key turning in the ignition. The wheels starting to turn is a consequence of the clutch engaging which is a consequence of some other step which is itself a consequence through a chain of events started by the key turning in the ignition. Being able to use a cause to predict an effect allows us to build reliable engines, and gives us further confidence in the rightness of the relation of cause and effect.

Sometimes however the key is turned but the car motor will not start. Another factor has intervened, but it does not mean that the relationship between cause and effect is broken, only that is sometimes more complicated than our limited experience suggests. Turning the key in the ignition is necessary for the car to start but not sufficient. All the links in the chain of cause and effect must be operational. If one of them breaks down, the whole chain of cause and effect will not operate to start the car.

Cause and effect operates in only one direction. You may start the car engine by turning the key, but the car engine cannot start and force your hand to turn the key.

One crack in this belief system has been produced by quantum physics and radioactivity. An atom of radioactive substance such as radium will eventually decay, and in the process it will emit energy. But there is no known triggering event that could serve as the cause of this decay event. In a large collection of radium atoms the rate of decay can be accurately predicted, but the identity of the individual atoms that decay cannot be determined in advance. Their decay is random, probabilistic. All radium atoms are identical, and of those identical atoms, some will decay in the near future and some will not.

Another crack in this belief system has been produced by quantum mechanical events such as the pattern produced by electrons passing through a double slit. The same sequence of causal events (or causal factors) regularly produces different effects (i.e. results), because sometimes the electron will hit one spot on the screen, sometimes another: as many electrons pass through the slits, the results may themselves in some random (unknowable) sequence. Furthermore, the probabilities of results of this kind can be calculated and they are highly predictable.

Results of this kind may be seen in the macro world of human beings in games of chance, where the dice are biased so that over a run of games, the probability of one player winning is enhanced, but there is no obvious pattern. The sequence of wins is random. Instead of scoring one 6 in every 6 throws with a fair dice, a player may score one 6 with every five, but the distribution of 6's is still random.