
Lecture 22: Expectation
Erik Demaine teaches expectation of a random variable as part of MIT's 6.1200J, Mathematics for Computer Science. The lecture builds up the definition of expected value from a probability distribution, then works through linearity of expectation, the property that lets you compute the expected value of a sum of random variables by adding their individual expectations even when the variables depend on each other. Demaine demonstrates this shortcut on problems that would be painful to solve directly, showing how splitting a complicated random variable into simple indicator variables turns a hard computation into a short one. The eighty-one minute session runs on MIT's standard chalkboard format, with worked examples building toward the kind of counting and probability arguments used throughout computer science. Part of the Spring 2024 offering of 6.1200J, continuing the course's sequence on discrete probability.