Socratic Stress Test

Layer 1 · Principles

• Your position is only as strong as the best version of the opposing argument. Weak steelmanning produces false confidence. If you cannot articulate the opposition's strongest case in a form they would endorse, you have not tested your position. You have only tested a caricature of it.
• Most beliefs rest on one or two hidden assumptions that do all the heavy lifting. Beliefs feel whole but are usually load-bearing on a small number of unstated premises. Surface those premises and the belief either holds up or collapses. Either outcome is useful.
• Falsifiability is the test that separates opinions from claims. If you cannot name the specific evidence that would make you change your mind, you are holding an opinion, not a claim. Opinions are fine. Mistaking one for the other is expensive.

Layer 2 · Systematic Approach

Five-step pressure test

1. State the claim. Write the belief or decision in one sentence. No hedging. If you can't commit to a clean sentence, the belief is not yet formed enough to test.
2. Steelman the opposite. Write the strongest version of the opposing argument in a form someone who holds it would endorse. If you cannot, find someone who can, or admit you haven't engaged with the opposition yet.
3. Surface assumptions. Name the one or two hidden premises that must be true for your claim to hold. List them explicitly. If any feel shaky, that is the leverage point.
4. Falsifiability check. Write the specific evidence or outcome that would make you change your mind. If nothing would change your mind, note that honestly and continue.
5. Confidence recalibration. Rate your confidence before (step 1) and after the test. Record both numbers. Over time the gap between them becomes a measure of how often your initial confidence was inflated.

Layer 3 · Force Multipliers

• Compounding calibration Logging before-and-after confidence on every test builds a personal calibration curve. Over time you learn exactly how much to discount your own initial certainty on different types of questions. The skill transfers to judgment at large. Users running this regularly for three months report significantly reduced post-decision regret and faster recognition of their own overconfidence tells.
• Adversarial thinking as default Forcing yourself to steelman opposition repeatedly rewires the first-response habit. After enough runs, you hear your own claims and automatically start building the counterargument in parallel. The meta-skill is the return, not any single test.
• Assumption surfacing as a transferable skill The step-three habit of naming hidden premises starts firing outside the formal framework. You begin noticing load-bearing assumptions in other people's arguments, in product strategies, in plans. The framework trains pattern recognition that operates everywhere.

Layer 4 · Success Metrics

Leading indicators

• Completed tests include a steelman the opposition would actually endorse, not a caricature.
• Every test names at least one specific piece of evidence that would change the conclusion.
• Before-and-after confidence is recorded.

Lagging indicators

• Measurable reduction in post-decision regret over a month or more.
• User catches their own overconfidence before acting, not after.

Failure modes

• Fake steelmanning
• Unfalsifiable retreat

Do not use when

• Time-critical decisions where a ten-minute pause is more expensive than a slightly suboptimal call.
• Purely aesthetic or values-based preferences where 'stress testing' misses the point.

Layer 5 · Implementation

Required to start

• A specific belief, claim, or decision to test.
• Eight to twelve uninterrupted minutes.
• Willingness to record a lower confidence number at the end than at the start.

Works best with

• A second person who can push back on your steelman.
• A notebook or log where the before-and-after confidence scores accumulate over time.

You are done when

• A clean single-sentence claim.
• A steelman of the opposition the opposition would endorse.
• One or two named load-bearing assumptions.
• A falsifiability criterion written down.
• A recalibrated confidence number.

Where to use it

• High-stakes decisions Before committing to a bet, hire, investment, or strategic pivot. The ten-minute cost is small relative to the decision cost.
• Emotionally loaded beliefs When you notice you are defending a position with more heat than evidence. The framework slows you down enough to separate the two.
• Debate preparation Before advocating publicly. If you cannot steelman the opposition, you will be caught off guard by it.