Long-Game Probability Strategy
Judge repeated decisions by expected value instead of one painful outcome.
- Difficulty
- Advanced
- Time to result
- ~ongoing to results
- Steps
- 5
- Confidence
- 94%
The Long-Game Probability Strategy treats recurring uncertain choices as a series rather than judging the method by one result. First, the decision-maker establishes principles and an expected-value rule using probabilities, costs, and potential returns. Next, each action is sized so that a loss is survivable. The rule is then applied consistently across many comparable opportunities, while process quality is recorded separately from whether any single attempt wins. This mirrors disciplined poker players, who can make a mathematically justified call and still lose the hand. Over time, repeated trials allow the underlying probabilities to express themselves. The framework does not mean ignoring new evidence: assumptions should be recalibrated when a meaningful sample contradicts them. Its output is a more stable strategy that resists fear, excitement, and short-term economic pressure.
Origin
Extracted from Marketing Against The Grain through Kieran Flanigan and Kipp Bodnar's discussion of poker players, expected value, and maintaining long-term beliefs during a downturn.
Core principles
- 01A sound process can produce an unfavorable individual result.
- 02Probabilities become meaningful across repeated comparable decisions.
- 03Short-term pressure should not force choices that violate durable beliefs.
- 04Emotional discipline improves when rules are established before the stakes arrive.
How to run it
- 1
Define the governing rule
State the principles, probabilities, and payoff conditions that make an action worthwhile before encountering a live decision.
Pro tip Write the rule in a form that another person could apply consistently.
Watch out A rule invented after seeing the outcome is rationalization, not strategy.
- 2
Protect against ruin
Size each decision so an unfavorable result does not prevent participation in future trials.
Pro tip Set a maximum exposure before emotions or urgency intensify.
Watch out Positive expected value does not justify a stake capable of destroying the system.
- 3
Repeat comparable decisions
Apply the same rule across a sufficiently long series instead of expecting one attempt to validate the probability.
Pro tip Define what makes two opportunities comparable before combining their results.
Watch out Unrelated decisions cannot be pooled merely to create a larger sample.
- 4
Separate process from outcome
Evaluate whether the action followed the rule and used reasonable inputs independently of whether it won.
Pro tip Log the information available at decision time.
Watch out Rewarding bad decisions that happen to win corrupts the strategy.
- 5
Recalibrate from evidence
Review accumulated results and revise the assumptions when the observed pattern provides credible contrary evidence.
Pro tip Schedule reviews rather than reacting after every loss.
Watch out Consistency should not become rigidity when the environment or probabilities change.
In the wild
A marketing team runs a series of capped experiments selected by the same expected-value criteria. Several fail, but the team evaluates whether targeting, cost, and upside assumptions were sound instead of abandoning the method after each loss. Quarterly review shows that the successful minority more than funds the failed tests.
→ Repeated disciplined bets produce positive portfolio returns without letting one failure derail the strategy.
A player can make a large call even when the opponent is more likely to win if the money already in the pot creates favorable odds. The player accepts that the hand may be lost because the same mathematically sound choice should pay across many similar hands.
→ The decision remains rational even when its individual outcome is unfavorable.
Common mistakes
Betting the system on one trial
A probabilistic edge cannot compound if one loss removes the ability to continue.
Judging only by outcomes
A lucky win can conceal a poor decision, while an unlucky loss can conceal a sound one.
Using patience to excuse bad assumptions
Long-term thinking still requires updating the model when sufficient evidence shows it is wrong.
Is it for you?
Best for
People making recurring decisions whose outcomes contain substantial randomness.
Not ideal for
One-time existential decisions where failure cannot be survived or repeated.
From the transcript
“do I have like a, do I have a belief system that says this is how I should do this, these are the reasons why…”
“They're playing thousands of hands of cards, but it's not, it's the probabilities that they're playing, right?”
“And if you, and if you're playing the probabilities, you have to play a long game.”
From the episode
How The Best Companies Pivot During A Bad Recession