Making Decisions You Can't Take Back

Most decisions can be revisited: a bad hire gets let go, a wrong product bet gets killed, a bad vendor gets dropped. Some decisions don't work that way. A cofounder relationship ended badly, a house bought in the wrong market, an acquisition signed, these don't have a quiet undo. The stakes of getting it wrong are entirely different, and so is the thinking that should go into it.

Why does generic decision-making advice fail here?

"Trust your gut" and "sleep on it" are both reasonable defaults for low-stakes, reversible decisions, where the cost of a wrong call is small and correctable. Applied to an irreversible decision, they're dangerously inadequate, because gut instinct is exactly where the most well-documented cognitive biases live, and a single night's sleep does nothing to surface the specific blind spots a high-stakes decision needs checked.

Generic AI prompting tends to make this worse by sounding more rigorous than it is. Ask an assistant "should I take this deal" and it will often produce a confident-sounding pro-and-con list that mimics careful analysis without actually surfacing which specific cognitive bias might be distorting your read of the situation, or whose outside view might disagree with yours and why. The structure looks like diligence. It isn't the same thing.

The danger is specific to irreversibility: a wrong call on a reversible decision is a correctable mistake, but a wrong call on an irreversible one is a permanent outcome dressed up, in the moment, as a confident and well-considered choice. The extra weight these decisions deserve isn't more anxiety, it's a different and more deliberate process than the one that works fine for everyday calls.

What do three documented thinkers say to do instead?

Daniel Kahneman: slow down specifically where fast thinking is known to fail

Kahneman's System 1 and System 2 framework, documented in Thinking, Fast and Slow (2011), distinguishes fast, automatic, intuitive judgment from slow, deliberate, effortful reasoning. The well-documented failure mode is that System 1 produces a confident answer quickly, and most people mistake that confidence for accuracy, especially under time pressure or emotional stakes.

Applied: Before an irreversible decision, explicitly ask which of the well-documented biases might be distorting this specific call, loss aversion making you overweight a small downside, sunk cost making you reluctant to abandon a path you've already invested in, overconfidence from a recent unrelated success. Naming the specific bias forces System 2 engagement in a way that generic "think it over" never does.

This works because System 1 doesn't announce itself as fast or potentially wrong, it simply feels like clear judgment. The only reliable way to catch its errors is to deliberately interrogate a decision against the known list of failure patterns, rather than hoping extra time alone will surface the distortion. Kahneman's own research is explicit that expertise doesn't immunize against this; experienced professionals are demonstrably as susceptible to these specific biases as novices, which is why the check has to be structural rather than relying on seniority or track record alone.

Charlie Munger: invert the question and ask what guarantees failure

Munger's inversion technique, documented across his published talks and Poor Charlie's Almanack, reframes a hard decision by asking not "how do I succeed" but "what would guarantee failure here, and am I doing any of that." Avoiding known failure modes is often more tractable than directly pursuing an uncertain success.

Applied: Before committing to an irreversible decision, list every way it could go badly, overpaying, misjudging a key person's character, misreading the market's actual size, and check which of those failure modes you have the least real information about. The decision isn't ready until the biggest unknowns on that list have been addressed, not just acknowledged.

Munger's own emphasis is that inversion works precisely because avoiding stupidity is more reliable than pursuing brilliance; most catastrophic, irreversible mistakes come from a known, nameable failure pattern that was visible in advance and simply wasn't checked against directly.

Ray Dalio: weight other views by actual track record, not confidence or seniority

Dalio's Principles (2017) documents believability-weighted decision-making: when a decision is consequential and uncertain, actively seek out views that disagree with your own, and weight each one by that person's actual demonstrated track record on similar decisions, not by how senior or confident they sound.

Applied: Before finalizing an irreversible call, identify someone whose judgment you respect who would likely disagree with your current lean, and ask them directly why. Weight their answer based on whether they've actually navigated a similar decision successfully before, not on how persuasively they argue in the moment.

Dalio's framework is explicitly a corrective against a common failure mode in high-stakes decisions: surrounding yourself with people who agree, or deferring to whoever in the room sounds most certain. Believability-weighting forces a different question, not "who agrees with me" or "who sounds confident" but "who has actually been right about something like this before."

Which approach fits which kind of decision?

What can you run before you decide?

Why it works: naming the specific bias forces a deliberate, slow-thinking check rather than a generic "have you considered the downsides" prompt that doesn't target any particular failure pattern.

Why it works: inversion surfaces concrete risks a forward-looking "will this succeed" question tends to skip past, and ranking by information gap tells you exactly where to focus before deciding.

Why it works: simulating a genuine, well-informed disagreement surfaces weaknesses in your reasoning that a simple pro-and-con list tends to miss, because it forces engagement with the strongest counter-case rather than a weak strawman.

What's the honest limit of any framework here?

These frameworks structure how you think about an irreversible decision. None of them has access to the specific facts only you know, the actual character of the person you're going into business with, the real state of your finances, the parts of the situation that don't fit neatly into a framework. The thinking gets sharper. The decision, and the risk of being wrong, is still entirely yours to own.

None of this is financial, legal, or professional advice, and a decision with real money or legal exposure attached deserves a qualified professional's input alongside, not instead of, this kind of thinking, especially when the decision is the kind that can't be quietly reversed later.

Kahneman's full System 1 and System 2 framework sits in the Behavioural Science category, while Munger and Dalio sit in the Investor category alongside Buffett, useful to load together when a decision spans both psychological and financial judgment. For running a structured decision brief before committing, Decision Brief is a $79 tool built for exactly that, and for checking afterward whether a past call was actually good or just lucky, What Actually Happened does the reverse.

All the copy-paste prompts from this guide, and the rest of the Situations series, live at /prompts too, free to copy without reading the guide first.