Game Theory & Strategic Decision-Making

Derek Muller's account of the Prisoner's Dilemma is built around the same historical spine Unit 1 uses — Axelrod's tournament, TIT FOR TAT, the Cold War arms race as the highest-stakes real instance of the game — and Muller interviewed Axelrod directly for it, so this is closer to primary source than most explainer content gets. Watch for the RAND Corporation section around the 14-minute mark: Cold War strategists formalized the game specifically to reason about nuclear posture, which is the historical hinge Unit 3 uses to argue the dilemma is a recurring structure, not a puzzle about prisoners. The video closes on Axelrod's four properties of TIT FOR TAT (nice, retaliatory, forgiving, clear) in the same order Unit 1's learning objectives list them — use it as a comprehension check after finishing the unit text.

Steve Levitt interviews Robert Axelrod himself about the tournament Unit 1 is built around — the exact source, not a secondhand summary. Axelrod explains in his own words why TIT FOR TAT's forgiveness mattered as much as its retaliation, and extends 'the shadow of the future' (his own coinage, used verbatim in Unit 1 and Unit 6's titles) to cyberweapons and cancer-cell competition, arenas well beyond the unit's ransomware-gang hook. Long-form and conversational rather than produced, this is the closest this library gets to hearing the scholar think out loud about his own fifty-year-old result, including where he thinks the tournament's lesson has been misapplied since.

Six minutes, animated, and structurally exact: two gingerbread men face a fox's version of the dilemma, and the video walks the Nash equilibrium logic Unit 3 requires — why mutual defection is the unique stable outcome even though mutual cooperation pays both players more. The back half turns the one-shot game into an infinite repeated game and shows the payoff structure flip, which is Unit 6's shadow-of-the-future argument compressed into ninety seconds. Assign this first, before the longer Veritasium video — it is the fastest correct account of the core structure the rest of the showcase builds on.

Jonathan Gruber's MIT lecture runs the full adverse-selection machinery — asymmetric information, the death spiral, and the government and market fixes that emerge to counteract it — using health insurance rather than used cars as the running example. That substitution is the point: Unit 9 argues the lemons structure recurs wherever one side of a transaction knows something the other can only guess, and health coverage is a cleaner illustration of the death-spiral dynamic than Akerlof's original car lot, because the stakes of adverse selection compound visibly over the course of the lecture. Pair with the MRU used-car video below for the canonical version, then watch this for the same structure at higher stakes. College-bound students should notice this is what an actual undergraduate economics lecture sounds like — no simplification, full worked logic.

Primer's agent-based simulation is the single best visual argument for Unit 7's central claim: a strategy can win without anyone choosing it, because genes that produce hawk-like or dove-like behavior simply reproduce at different rates depending on what the rest of the population is doing. Watch the population ratio stabilize around the 9-minute mark — that convergence point is the evolutionarily stable strategy the unit defines, rendered as a number the simulation finds rather than a proof a mathematician derives. This is the course's clearest instance of removing the mind from the game, which is the unit's opening move: deer that fight with antlers instead of teeth are running the identical hawk-dove logic without ever knowing they are playing.

Richard Dawkins explains, in his own words, why evolution favors a stable ratio of competing strategies rather than one dominant type — the hawk-dove insight generalized beyond animal combat to sex ratios and beyond. This is a tighter, more conceptual companion to the Primer simulation above: where Primer shows you the population converging, Dawkins explains why convergence to a mixed ratio, not a single winner, is the mathematically expected outcome once you accept that genes are playing a game against copies of themselves. Four minutes, dense, worth pausing on the fisher/pirate example partway through.

Under three minutes and built entirely around Akerlof's original 1970 example: a used-car lot where sellers know the car's true quality and buyers can only guess, so the good cars get driven out of the market and only lemons remain. This is Unit 9's opening case study delivered in its canonical form, including the market solutions (warranties, Carfax, certified pre-owned programs) that emerged specifically to repair the information gap. Watch before the unit text — it installs the base case so that the unit's later moves (diplomas as signals, peacock tails as costly signals, propaganda as a flooding strategy) land as the same structure wearing different costumes, which is the unit's explicit thesis.

Dr. Emily Riehl walks through the Gale-Shapley algorithm using couples proposing and rejecting their way to a stable pairing — the exact matching-market logic Unit 10 uses to explain how residency programs assign doctors to hospitals and how school-choice systems assign students to schools. The video's core insight is the one the unit's capstone argument depends on: 'stable' does not mean 'optimal for everyone,' it means no pair would rather defect to each other than stay matched, which is a Nash-style equilibrium concept applied to a matching problem instead of a payoff table. Alvin Roth's real-world deployments (kidney exchange, school matching) that the unit covers are direct engineering applications of the algorithm this video explains from first principles.

Delivered from the institution that studies strategic conflict as its actual job, not as an academic curiosity — a Senior Military Analyst at the Naval War College works through what makes something a game in the formal sense before turning explicitly to game theory (the segment starting around minute 18, chaptered 'What About Game Theory?') and then to how war games function as decision-support tools rather than entertainment. This is the course's clearest bridge from Unit 2's payoff-table reasoning to the professional register a reader headed toward defense, intelligence, or policy work will actually encounter: the same dominance and equilibrium logic Polak teaches in a Yale classroom, applied by an institution whose job is reasoning about real adversaries under real stakes. Pair with Polak's Lecture 1 above — same underlying structure, two very different rooms.

This is the actual Ben Polak lecture the course's opener is adapted from — the same payoff-table exercise Unit 2 runs on the reader, filmed live on a Yale classroom of students who don't yet know they're about to learn never to play a dominated strategy. Watch the section starting around minute 21 (Chapter 5, 'Strictly Dominant versus Strictly Dominated Strategies') for the exact proverb the unit repeats verbatim, and the coordination-problem material after minute 41 for the setup Unit 4 continues. Long — 68 minutes — but this is what the course's every-concept-inside-a-real-situation method looks like at full classroom length, taught by the voice this entire course borrows its register from.

Polak's formal definition of Nash equilibrium, delivered through a live class investment game before the definition is stated — the same before-you-name-it sequencing every unit in this course follows. This is the primary-source lecture behind Unit 2's central proverb: equilibrium means no regrets, not the best possible outcome for everyone. The bank-run material after minute 32 is directly relevant to Unit 4's coordination-game material — Polak shows a class splitting between a good equilibrium and a bad one with identical incentives to coordinate, which is exactly the standards-war logic (QWERTY, VHS) the unit applies to focal points. Assign after Unit 2's text, as the fuller derivation of the 'no regrets' idea the unit states more briefly.