Conditional expectation and variance can be computed for a sum of a random (rather than fixed) number of independent random variables by conditioning on the count: the law of total expectation (Adam's Law) gives E(X) = E(N)·μ, and the law of total variance (Eve's Law) decomposes Var(X) into the expected conditional variance plus the variance of the conditional expectation, yielding Var(X) = σ²E(N) + μ²Var(N). The lecture then introduces four fundamental probability inequalities — Cauchy-Schwarz, Jensen's, Markov's, and Chebyshev's — which provide general, distribution-free bounds (as opposed to approximations) on expectations and tail probabilities, contrasting bounds (always true, direction-certain) with approximations (uncertain closeness to truth). This belongs to probability theory, specifically conditional expectation via iterated expectation/variance and the theory of probabilistic inequalities used to bound expectations and tail probabilities without full distributional knowledge.
Conditional expectation and variance can be computed for a sum of a random (rather than fixed) number of independent random variables by conditioning on the count: the law of total expectation (Adam'…