Odds & EndsLAB

Casino mathematics

For adults · 18+
Back to reading notes

Probability notebook

Why “at least one” never means “due”

The complement rule explains repeated chances without giving the next event a memory.

Start with the assumption

Suppose an event has a fixed probability p on each independent trial. The chance that it does not occur is 1 − p. Over n independent trials, the chance of no occurrence is (1 − p) raised to the power n. Subtract that result from one to find the chance of at least one occurrence. Our calculator implements that expression directly.

A worked example

With a hypothetical 10% chance per trial and ten independent trials, the chance of at least one occurrence is approximately 65.13%. It is neither 100% nor a prediction of the particular sequence you will observe. Multiplying ten by 10% answers a different question: the expected number of occurrences, which is one.

What the model leaves out

Independence and a constant probability are requirements, not decorative small print. A game with changing conditions needs a different model. A losing sequence also does not make a future independent event more likely. Use the experiment to understand the assumption, never to decide that a win must be imminent.

Methods & sources

Continue reading

At least one: the probability lab

An average is not a session forecast