Regret Minimisation as a Decision Framework
Asking which outcome I would regret more produces better decisions than asking which is more likely, because I am bad at probabilities and good at anticipating regret.
I cannot estimate probabilities for this asset class with any accuracy. My scored record of predictions establishes that clearly.
What I can do is anticipate which error I would find harder to live with, and for several decisions that turns out to be a better question.
The framework
For any decision, describe the two errors.
The error of commission: I act and it turns out to be wrong.
The error of omission: I do not act and that turns out to be wrong.
Then ask which one I would find harder to live with in ten years.
Where it produces clear answers
Initial position size. The regret of holding too little of something that appreciated substantially is real and survivable. The regret of holding too much of something that collapsed, in a way that affected my life, is not.
That asymmetry argues for a smaller position than a probability-weighted calculation would suggest, and I think it is the correct adjustment.
Custody. The regret of having done the restore test unnecessarily is fifteen minutes. The regret of not having done it, discovered at the moment it matters, is total.
Trivially asymmetric, and it explains why operational safeguards should be over-provisioned relative to what the probabilities suggest.
Inheritance documentation. One evening against a permanent loss to my family at a moment they are already dealing with something worse.
Leaving funds at a venue. The regret of having moved to self-custody unnecessarily is a transaction fee. The regret of not having, in the case that repeats every few years in this sector, is the balance.
Where it produces worse answers
Timing. Regret about missing an advance is vivid and available. Regret about buying before a decline is equally real and harder to imagine in advance, because the decline has not happened yet.
Applied to timing, this framework has a bias toward acting, which is the opposite of what my record says I should do.
Selling. Regret about selling something that then appreciated is the most vivid financial emotion I have experienced. Weighting it heavily produces a bias toward never trimming, which is exactly the failure that leaves people holding position sizes they never chose.
So I use it for structural decisions and not for market ones.
The version that works
Apply it to decisions about arrangement: how much, where held, what safeguards, what documented.
Do not apply it to decisions about timing or price, where the vividness of one regret is a function of recency rather than of magnitude.
The practical output
Every operational safeguard I run is over-provisioned relative to its probability, because the asymmetry justifies it. Annual restore tests, a second backup location, documented inheritance, a working balance at the venue I buy through kept deliberately small.
None of those is optimal under a probability-weighted calculation. All of them are obvious under a regret-weighted one, and the difference between the two frameworks is largest exactly where the downside is irreversible.
This is a personal account of holding through market cycles. It describes what one person did and why. It is not a recommendation, and past cycles do not predict future ones.