The model
Evidence being likely under a hypothesis does not make that hypothesis certain once the evidence is observed. Prior probabilities and alternative explanations must also be considered.
Bayes' theorem relates conditional probabilities consistently. In the calculator, H and not-H partition the complete probability space. Inputs lie between zero and one; if the total probability of the evidence is zero, the result is undefined.
Worked example
With P(H) = 0.1, P(E|H) = 0.8 and P(E|not-H) = 0.2, the posterior probability is approximately 0.3077.
Scope & assumptions
The origin of the input probabilities matters as much as the arithmetic. This teaching tool explains probability reasoning; it is not designed for medical diagnosis or decisions about a person.
TRY IT YOURSELF
Check your understanding.
Change the prior probability to 0.5 and examine how the same evidence changes the result.
Sign-in required. Calculator interface currently in Turkish.
