Skip to content

Probability Ladder

This ladder covers contest probability where:

  • the randomness is in the statement
  • the state space is finite or bounded
  • the answer is one exact probability or expected value

Use it when a problem asks you to analyze a random process exactly, not to design a randomized algorithm.

  1. bounded-sum distribution DP
  2. expectation from a known PMF
  3. expectation through linearity / independence
  4. harder process models only after the first three feel routine

Anchor Problems

  • Dice Probability: the repo's exact first route, one PMF DP over repeated fair-die throws

Retrieval / Build Path

Compare Points

  • Candy Lottery: expectation from a derived distribution of the maximum
  • Moving Robots: repeated transition probabilities plus independence/complements
  • Randomized Algorithms: anti-cue when the randomness belongs to the algorithm rather than the statement

Exit Criteria

You are ready to leave this ladder when you can:

  • decide whether to build a PMF or compute expectation directly
  • use linearity of expectation without needing the full joint process
  • recognize when simulation is the wrong instinct
  • separate probability-in-the-problem from randomized-algorithm analysis

External Practice