Game Theory · Lesson 07 · ← Lesson 06

Why giving up options can make you stronger

Commitment turns an empty threat into a credible one — and credibility is power.

~13 min · one sitting Skill: commitment & credibility Builds on: backward induction

First — 20-second recall from Lesson 06

Without scrolling back: backward induction means…?

01 · THE PARADOXFewer options, better outcome

In Lesson 06 you learned backward induction — working from the end of the tree to the beginning. And you saw a specific failure mode: a player threatens something they won't actually do when the moment comes. The rival ignores it, and the threat achieves nothing.

The insight that unlocks this lesson is counterintuitive: the fix is to remove your own future options. Not add them. Remove them. When you commit — by burning the fallback, signing a contract, making a public pledge — you change what a rational opponent expects you to do. The threat becomes credible not because you want to carry it out, but because you've made it your best move if they call you on it.

The paradox of commitment: voluntarily restricting your choices can make you better off — because it changes the game everyone is playing.

This is not about stubbornness or bluster. It's structural. You change the game so that what you've threatened is now the rational thing to do. The threat becomes self-enforcing — and the other side, knowing that, won't force you to carry it out.

The foundational treatment of threats, promises, and commitment is Thomas Schelling, The Strategy of Conflict (1960). The formal game-theoretic account appears in Open Yale ECON 159 (Polak), lectures on sequential games and credibility.

02 · THE ENTRY GAME, FLIPPEDHow commitment changes the tree

Recall the entry-game from Lesson 06. You're a potential entrant; the incumbent runs a monopoly. If you Enter, the incumbent chooses to Fight (price war) or Accommodate (share the market). Without commitment, their payoffs were:

Entry game — no commitment (Lesson 06 baseline)
	Fight	Accommodate
You Enter	You: −2 · Inc: 2	You: 4 · Inc: 5
You Stay Out	You: 0 · Inc: 8	You: 0 · Inc: 8

Backward-induct: if you enter, incumbent compares Fight (2) vs Accommodate (5). Accommodate wins → they accommodate. Knowing that, you enter and earn 4. Incumbent's threat to "fight anyone who enters" is cheap talk.

Now the incumbent makes a commitment before you decide — say, they publicly pre-invest in excess capacity and lock in a cost structure that makes fighting cheaper, or they sign a visible, automatic price-match contract with all their patients. Either way, it changes the Fight payoff from 2 to 6.

Entry game — after incumbent's commitment (Fight payoff changed)
	Fight	Accommodate
You Enter	You: −2 · Inc: 6	You: 4 · Inc: 5
You Stay Out	You: 0 · Inc: 8	You: 0 · Inc: 8

Backward-induct again: if you enter, incumbent compares Fight (6) vs Accommodate (5). Now Fight wins. You foresee that — entering yields −2, staying out yields 0. You Stay Out. The incumbent earns the monopoly payoff of 8.

The commitment flip

The incumbent is strictly better off (8 vs 5) having thrown away the option to accommodate. By pre-committing — giving up flexibility — they turned an empty threat into a credible one and deterred entry entirely.

Same reflex you already have A pre-declared intra-op stop rule that you genuinely cannot override mid-case gets its power from being irreversible and visible to the team — exactly what makes a strategic commitment credible. Strategy is the same move applied outside the theatre.

This is the whole mechanism in compact form. You will see it in excess capacity, no-discount policies, money-back guarantees, published "best & final" prices, and escalation clauses — every case where a player locks themselves in to make the other side back down.

The entry-deterrence commitment game is a standard example in Open Yale ECON 159 and SEP — Game Theory (§ on credibility and strategic moves). See also Dixit & Nalebuff, The Art of Strategy (Norton, 2008), chapters on commitment.

03 · INTERACTIVEWhich threats are credible?

Credibility isn't about sincerity — it's about structure. A commitment is credible when it's observable (the other side can see it) and costly or hard to reverse (so backing out is genuinely painful), which means carrying through is now in your interest. Everything else is cheap talk.

Which of these is a CREDIBLE commitment?

Run the credibility test on any threat or promise:

Observable? The other side must be able to verify the commitment exists.
Hard to reverse? Backing out must be costly — reputationally, financially, or structurally.
Changes your best move? The commitment must make following through your rational choice if challenged.

If all three hold, the commitment works. If any fails, you've only made noise — and your opponent knows it.

04 · THE TOOLKITSix commitment devices — with your examples

These are the mechanisms that manufacture credibility. Each works by the same logic: it changes your future best move, which changes what the other side rationally expects.

Burn the bridge / close the fallback. Remove the option you'd otherwise retreat to. Hernán Cortés scuttling ships. For you: publishing a no-refund course policy removes the "just cancel" fallback and forces you to deliver — and signals the same to sponsors.
Binding contract. Legally enforceable terms make reneging genuinely costly. A signed sponsorship agreement with a penalty clause turns "we'll keep the price" into a credible promise. The contract does the work, not your word.
Public pledge. Once announced to a large audience, backing down costs reputation. A posted "best & final" price for your skull-base course sponsorship, shared with all potential sponsors simultaneously, is credible because the reputational cost of secret deals is visible.
Delegate to a rule or third party. "My policy is X — I can't make exceptions" is credible when the rule is real and the other side can verify it. A published, fixed pricing page your team enforces is more credible than a number you quote verbally. The decision is no longer yours to override in the moment.
Sunk investment. Pre-spending money changes your future incentives. A rival clinic that pre-invests in new imaging equipment has committed to competing on throughput — they have to fill the machine. Your own investment in a paid Google Ads campaign changes your incentive structure visibly.
Reputation across repeated dealings. In a one-shot game reputation doesn't bind you — but when you play the same opponents repeatedly (referrers, hospital administration, course sponsors year after year), a track record of following through is a commitment device. We'll return to this in Lesson 08.

The commitment question

For each device: does it change my future best move in a verifiable, hard-to-reverse way? If yes, it's doing real work. If no, it's theatre.

Before you move on — write it: for holding cranial pricing at a floor with no exceptions, which device above makes that floor credible, and why?

Model read: delegate to a rule stacked with a public pledge. A published, fixed price your team enforces removes your own discretion to grant an exception in the room — and stating it as a "best and final" floor to every patient makes any quiet discount a visible, reputational break. A verbal "I don't usually discount" is cheap talk; the same floor, published and enforced by someone other than you in the moment, is a commitment.

05 · THE DOUBLE EDGECommitment is power — and a trap

Everything said so far makes commitment sound like a pure gain. It isn't. The same mechanism that gives you power also locks you in — and the world changes.

Inflexibility. If circumstances shift after you've committed — a better rival offer, a market you didn't anticipate — you're stuck. The commitment that deterred entry also prevents you from pivoting. Commit only when the payoff structure genuinely rewards it.
Mutual over-commitment (brinkmanship). If both sides commit to incompatible positions — both signal they'll fight rather than yield — the game can lock into a collision neither wanted. Cold War nuclear deterrence was brinkmanship at scale. In your world: if you and a rival both publicly commit to "lowest price guaranteed," one of you has to break or both bleed.
Commit to the right thing. Commitment amplifies whatever you've locked in. Committing to a wrong strategy makes you reliably wrong. Check that the payoff structure rewards the committed position before you burn the bridge.

The discipline: commitment is a scalpel, not a default posture. Use it when (a) the payoff structure rewards it, (b) you can make it visible and hard to reverse, and (c) the world you're committing into is stable enough that the lock-in doesn't become a liability.

Schelling, The Strategy of Conflict (1960), is the foundational treatment of brinkmanship and the dangers of mutual commitment. See also SEP — Game Theory on credible threats.

06 · KNOW THE LIMITSWhen a commitment isn't credible

Commitment is powerful precisely because it's structural — which means it fails in specific, predictable ways. Check for these before you burn a bridge.

Contraindications — this isn't a credible commitment

Unobservable or quietly reversible. A commitment nobody can observe, or one you can quietly reverse, changes nothing — credibility requires it be visible and hard to undo.
Over-commitment. Locking yourself in throws away flexibility you may need if the world shifts — a rival offer, a market you didn't anticipate.
The empty threat. An empty threat fools no rational opponent — if carrying it out would hurt you more than them, they know you won't, no matter how loudly you say it.

Naming when a commitment fails is what stops it from becoming theatre. Run the credibility test from Section 03 before you rely on one — or before you lock yourself into one.

07 · LOCK IT INThe commitment recap

A threat or promise is credible only if carrying it out is in your interest when the moment comes — otherwise it's cheap talk.
Commitment makes a threat credible by changing your future best move — usually by removing your option to back down.
The credibility test: observable? hard to reverse? changes your own best move if challenged?
The toolkit: burn the bridge · binding contract · public pledge · delegate to a rule · sunk investment · reputation (→ L08).
The entry game shows the mechanism: commitment changed Fight from 2 to 6 → Fight beats Accommodate → entrant stays out → incumbent earns more by giving up an option.
The double edge: commitment creates inflexibility. Over-commitment (brinkmanship) can trap both sides. Commit to the right thing, visibly, when the payoff structure genuinely rewards it.

Bring it back. Think of a negotiation or competitive move you're facing — a sponsorship ask, a price negotiation with a hospital, a rival clinic's ad spend you want to deter. Ask yourself:

What could you credibly commit to — publicly, bindingly — that would change how the other side responds? What option would you be giving up? Does the payoff structure reward that trade? Run the credibility test: observable, hard to reverse, changes your best move?

That's the rep that makes this stick. Bring me the real case and we'll backward-induct it together.

Even better — do it on a real decision, not a hypothetical one. Open DECISIONS.md and pull D2 — the particular pricing floor: holding cranial ≥ R$30k as a credible no-discount commitment, the "best and final" that only works if you never make the first exception.

Copy learning-records/REP-TEMPLATE.md and fill Phase 1: name the four outcomes, rank them for you, name what patients/referrers are actually optimising (price as a quality signal vs shopping on price), run the credibility test on your own floor, and flag the one payoff you're least sure of.

The gate: this lesson isn't "done" when you finish reading — it's done when one REP-*.md exists with Phase 1 filled. Delivered ≠ learned. One honest rep beats reading the next three lessons.

Primary sources: Open Yale ECON 159 (Polak) · SEP — Game Theory · Thomas Schelling, The Strategy of Conflict (1960) · Dixit & Nalebuff, The Art of Strategy (Norton, 2008).