This is a technique where tests are created based on the test developer's understanding of the implementation under test and those tests which are likely to cause failures. The test developers create tests which are based on their best guess (or their own experience) regarding where errors might be found. Using experienced testers, this has been found to be an excellent way of finding errors. For example, experience might show that on particular processors, selecting particular numbers for maths tests may be more likely to generate errors.