Find what you might be missing.

Step outside the story of uniqueness and find a comparable reference class.

A number is not a base rate without a defined class, denominator, sample size, source, and date.

Compare the timelines and conversion rates of ten genuinely similar launches.

Separate repeated reports from independent evidence and trace claims toward their origin.

Five accounts repeating one report are not five independent observations.

Trace each confident statement back to first-hand evidence, a primary source, or a repeated claim.

Look at rewards, penalties, status, safety, convenience, and constraints before settling on a character story.

Incentives suggest plausible moves; they do not prove what a person will do.

Ask what changes for each actor if option A is chosen instead of option B.

Notice information asymmetry and turn it into a specific fact-finding task.

An unknown is not evidence of deception. Test it without converting uncertainty into suspicion.

Identify the one operating fact the counterparty can see and you cannot.

Create a materially different causal account before acting on the first story.

A useful alternative must predict something different, not merely rename the same explanation.

Write two explanations and one observable result that should differ between them.

Find the smallest discriminating observation before making a larger commitment.

Prefer a test whose result can actually change the decision.

Define the evidence that would favour demand failure over execution failure.

Break the hidden assumption that the current option set is complete.

Generating alternatives does not establish that they are feasible or valuable.

Temporarily remove the favoured option and produce two executable substitutes.

Expose switching costs, habits, coordination work, and operating burdens.

A superior outcome can still fail when the transition cost is ignored.

Map the first week of changed behaviour, not only the promised end state.

Convert optimism into concrete failure mechanisms, early warnings, and preventive action.

Do not let it become an unranked list of fears or a reason for paralysis.

Name the three likeliest failure mechanisms and the earliest signal for each.

Find the bottleneck or leverage point before optimising every visible part.

Not every system has one stable constraint; verify that the candidate truly limits the whole.

Ask whether replacing a person while keeping the structure would reproduce the same result.

Turn a defended belief into a claim that reality can test.

The falsifier must be observable and set before the outcome is known.

State the metric and deadline that would trigger reconsideration.

Expose necessary conditions and identify the most fragile assumption.

Necessary conditions are not guarantees; rank them by uncertainty and impact.

List three conditions for success, then test the least supported one first.