The Pre-Mortem: Running One on a Polymarket Thesis Before You Size Up

Where the technique comes from

Gary Klein, a psychologist who spent decades studying decision-making under time pressure (firefighters, military commanders, trauma surgeons), introduced the pre-mortem in the late 1990s. Klein's observation: people are much better at explaining a failure that has already happened than at anticipating one that has not (Klein, 2007, "Performing a project premortem," Harvard Business Review).

The trick is to move the question from "what could go wrong?" to "assume this trade lost. Explain why." The framing activates the explanatory circuit instead of the defensive one. Klein's reported result: pre-mortems increased the number of surfaced failure reasons by roughly 30 percent compared to standard risk assessment.

Why it works on a prediction-market position specifically

A Polymarket trader sizing into a $5,000 position on an ETH price contract has three characteristic failure modes: probability is miscalibrated, the contract resolves on mechanics the trader did not fully read, or the position size is so large that any outcome other than the intended one is operationally ruinous. Standard pre-trade checklists catch the first two unevenly. The pre-mortem frame catches all three at once.

The exercise also forces the trader to distinguish thesis risk from size risk. These are different questions. A thesis can be correct on probability grounds and still generate a position that is too large for the bankroll under fat-tail payoffs. Convexly's finding that Hill α on realized Polymarket PnL sits at 1.28 (Edge Score Methodology V1) means variance is not a reliable sizing input. A pre-mortem reveals whether the trader's sizing survives the fat tail, independent of whether the thesis was right.

Worked example: a $5K ETH contract

A trader with $50K in deployable capital has been watching a Polymarket contract on "Will ETH close above $3,500 on March 31?" The contract trades at 42 cents. The trader's view: ETF inflow data and an upcoming protocol upgrade push the fair probability closer to 56 percent. The initial sizing plan is $5,000, or 10 percent of bankroll.

Ten minutes of pre-mortem output, written independently:

Three of the four failure modes are contract-mechanics or tail-shock risks the original thesis did not price. The fourth is a sizing error that is independent of the thesis being right. The trader's post-exercise revision: cut size to $2,000 (4 percent of bankroll), and actually read the resolution oracle before sizing further.

The procedure, step by step

Pairing pre-mortems with the confidence estimate

A useful discipline after the exercise: write down the revised probability estimate. In the ETH example above, the initial 56 percent view drops to roughly 48 percent once the macro-shock and mechanics modes are priced in. At 48 percent on a contract trading at 42 cents, the edge is thin. The position either needs to be smaller, or it is not a position at all.

Convexly's operator takeaway argues for exactly this kind of discipline: the wallets driving the top of the Polymarket leaderboard are running fewer, more concentrated positions, not more ideas (see the operator's takeaway). The pre-mortem is the filter that does the cutting.