Cognitive Biases and the Pre-Mortem

By Better Wise

In the last lesson you put your options side by side and priced what each one costs you in alternatives. That comparison is only as good as the inputs, and the inputs are produced by a human brain with predictable shortcuts. This lesson shows you where those shortcuts enter the table, then gives you a 10-minute routine that drags them into the open before you commit.

By the end of this lesson you should be able to:

• Name the six biases that most often distort a comparison of options
• Run a 10-minute pre-mortem that produces concrete, specific failure modes
• Attach a specific bias to each of the top failure modes
• Convert each failure mode into one cheap check you can run before deciding

Note: A clean comparison table built on biased inputs is worse than no table, because it looks rigorous while quietly protecting your preferred option.

Where bias sneaks into your comparison

Bias rarely arrives as an obvious error. It arrives as a reasonable-sounding reason to stop looking. Watch for the asymmetry: if you spent forty minutes listing the downsides of the new role and four minutes on the downsides of staying put, you are not comparing, you are justifying.

Two of these deserve a closer look because they hide best. Availability bias replaces data with stories, so one dramatic failure outweighs twenty quiet successes. Sunk cost disguises itself as loyalty, commitment, or "not wasting what we built". The test is simple: would you choose this option today if you were starting fresh with no history?

Exercise 1: Find your bias fingerprint

Take the decision you framed in Module 1 and have been comparing since. Write one sentence describing the bias fingerprint a neutral observer would see in your last 30 minutes of work on it.

Answer guidance for Exercise 1

A useful answer names an observable behavior, not a feeling. "I rewrote the cost section of the option I like three times and never touched the other two" points at confirmation bias. "I keep bringing up the migration work we already paid for" points at sunk cost. "My confidence range for the timeline is plus or minus one week" points at overconfidence. If your sentence is "I think I am being fairly objective", you have not looked yet.

The 10-minute pre-mortem

A pre-mortem inverts the usual risk question. Instead of "what could go wrong?" you state as fact that the decision failed badly, then explain why. The technique comes from psychologist Gary Klein and rests on a finding called prospective hindsight: explaining a certain outcome produces more specific risks than asking about a possible one.

flowchart TD
    A["🎯<br/>Set the frame"] --> B["✍️<br/>Silent failure modes"]
    B --> C["🗣️<br/>Read aloud, no rebuttals"]
    C --> D["🧩<br/>Group and vote"]
    D --> E["🧠<br/>Name the bias"]
    E --> F["✅<br/>One cheap check"]

1. Set the frame (1 min): "It is twelve months from now and this decision failed badly. Write the story of why."
2. Write silently (3 min): one failure mode per line, no discussion, specifics only.
3. Read around the group (2 min): one item each in turn, no defending the decision yet.
4. Group and vote (2 min): cluster into three to five themes, then mark the two most likely and the two most damaging.
5. Name the bias (1 min): for each top failure mode, write the bias that would make you miss it today.
6. Convert to a check (1 min): one cheap action you can take before committing that would surface the problem early.

Three rules make or break it. Silent writing first, because the first person to speak anchors the whole list. No rebuttals during the listing rounds, because defending the decision is exactly the behavior you are trying to interrupt. And specificity over categories: "the client renews with the incumbent" beats "sales risk".

Exercise 2: Run it on your own decision

Set a timer for five minutes. Produce at least six specific failure modes for your decision. Then pick the two most likely and complete this line for each: "The failure is X. The bias that would hide it is Y. The cheap check is Z."

Answer guidance for Exercise 2

A strong pair might read: "The failure is that hiring takes four months instead of six weeks, hidden by planning fallacy, checked by asking two people who made a similar hire how long it actually took." Or: "The failure is that we stay on the old tool because of migration work already done, hidden by sunk cost, checked by asking whether we would pick it starting today with no history." Weak answers stay abstract ("the failure is poor execution") or name a check that costs more than the decision itself.

Turning failure modes into bias-to-check pairs

The pairing step is what separates a pre-mortem from a worry list. Each failure mode earns its place only when you can say which bias would let it through and what you will do about it this week.

You can ask an AI to suggest failure modes you missed, and it is often good at that. It cannot tell you which bias is yours. That judgment stays with you, and the next lesson builds directly on it by matching how much analysis a decision deserves to how reversible it is.

Field rules for catching your own shortcuts

• Run the pre-mortem before the comparison is locked. Once you have emotionally committed, the exercise becomes theater: the list gets shorter and the checks get softer.
• Write the failure modes before anyone speaks. Group pre-mortems fail through politeness much more often than through bad ideas.
• Treat "we already invested" as a red flag, not a reason. Past cost is evidence about the past, not about the next twelve months.
• Ask what would change your mind, in writing. If no finding could, you are not deciding, you are announcing.
• Keep the checks cheap. A check that costs more than the downside it prevents will not get run.
• Do not let a long risk list feel like progress. Three specific failure modes with named biases and one check each beat thirty abstract worries.

What to carry forward

Bias enters through the inputs, not the arithmetic, so a tidy comparison table can still be a justification. A pre-mortem forces specificity by assuming failure and asking why, and its output is only useful when each top failure mode carries a named bias and one cheap check. Name the bias, then buy the check.

Next up: you have stress-tested the options, so the next lesson asks how much of this work a decision actually deserves. That depends on reversibility, on the speed the choice allows, and on how well your stated confidence matches your track record.

Front: What is a pre-mortem? Back: A technique that assumes the decision already failed and asks people to explain why, producing specific failure modes instead of vague risks.

Front: What psychological effect makes pre-mortems work? Back: Prospective hindsight: explaining a certain outcome yields more concrete risks than asking about a possible one.

Front: Why write failure modes silently before discussing? Back: It stops the first or loudest speaker from anchoring the group's list.

Front: How does anchoring distort a comparison? Back: The first number or option sets the range, so everything else is judged relative to it, often the status quo.

Front: What is the tell for confirmation bias in decision work? Back: Your research questions are worded to support the option you already prefer.

Front: What does availability bias do? Back: Makes vivid, recent, memorable cases feel more probable than base rates.

Front: How do you detect sunk cost in an option? Back: Past investment keeps it alive: ask whether you would pick it today starting fresh with no history.

Front: What does the planning fallacy look like on paper? Back: Optimistic timelines and costs, with the best-case plan presented as the plan.

Front: What is the overconfidence tell in a forecast? Back: A confidence range that is too narrow, for example plus or minus one week on a four-month project.

Front: What must each top failure mode produce? Back: A named bias that would hide it plus one cheap check to run before committing.

Front: How does a pre-mortem differ from a risk list? Back: A pre-mortem starts from certain failure and explains it; a risk list starts from what might go wrong and stays abstract.

Front: What rule keeps a pre-mortem from collapsing into defense? Back: No rebuttals or debate during the listing rounds; criticism comes only after the list is complete.

Front: When does a pre-mortem become theater? Back: When it runs after the decision is effectively made and no check could change the outcome.

A pre-mortem assumes the decision _____ and asks people to explain why, rather than asking what could go wrong. Writing failure modes silently first prevents _____ by the first or loudest voice in the room. Sunk cost keeps an option alive because of _____, even when it no longer wins on future value. The planning fallacy shows up as _____ timelines and costs, with the best case presented as the plan. Availability bias makes _____ cases feel more probable than base rates. The useful output of a pre-mortem is not a longer risk list but _____ you can run before committing.