PBT is an approach to testing in which we test the properties of a function rather than individual examples. Instead of a single test with specific input data, the framework generates hundreds of random cases.

1. Finds hidden bugs that are not covered by specific tests.
2. Expresses the behavior of the system through properties.
3. Automatically generates hundreds of inputs.
4. Reduces the risk of “bias”, when tests check only “convenient” cases.

1. Formulate properties at the domain level.
2. Don't overuse PBT: for complex business logic, you also need regular tests.
3. Use a fixed seed for reproducibility.
4. Combine PBT with random-based and sample tests.