Use confidence ranges instead of absolute reads
Replace “Player 8 is evil” with a reasoned confidence level and what would change it. This keeps you flexible when new results arrive and prevents one early read from controlling every later interpretation.
Separate process from outcome
A well-supported shot can hit Town because a relevant interaction was unknown. A reckless guess can hit evil by luck. Review whether the evidence justified the action at the time rather than rewarding only successful outcomes. Otherwise, lucky habits become permanent mistakes.
Control tempo deliberately
Fast play is correct when mechanics are solved and dangerous when Town still needs claims. Slow play is correct when another result can resolve the board and dangerous when evil reaches parity tonight. Tempo is a strategic resource, not a personality trait.
Use a short post-game review
After a loss, identify one decision with the biggest expected impact: target choice, claim timing, vote, execution or shot. Reconstruct what was known before the action and write one alternative. Reviewing one decision deeply teaches more than blaming every teammate.
Protect mental bandwidth
Ranked chat can become noisy and emotional. Keep a compact ledger of claims, confirmed roles and open contradictions. Externalizing the board reduces tilt because you do not have to remember every argument while also making the next decision.
Outcome bias example
You execute a player after two independent mechanical contradictions and they unexpectedly flip Town because a Transporter claim was hidden. The result is bad, but the lesson is not “never execute.” The review question is whether checking for transport was reasonably available before the decision and how to add that check next time.
A review that produces one change
| Field | Example |
|---|---|
| Decision | Shot a player whose claim conflicted with a visit |
| Information available then | Visit report; claimed role; one unresolved transport |
| Missing check | Did not ask both transport recipients to confirm |
| Next-match rule | Resolve a relevant redirect before using the shot |
Do not grade the decision using a role you only learned on the final reveal. Review the information you could have obtained before acting and choose one repeatable improvement.
Common mistakes
- Equating a win with good play.
- Hard-locking a read and filtering every new clue through it.
- Playing fast because the lobby is impatient.
- Reviewing ten small mistakes instead of one high-impact decision.
