Pre-Commitment: Deciding Before the Moment

Why the moment is the wrong time

Every bias in this subject has the same structure: it operates in a moment, on a person with a position and a result to explain. Loss aversion makes a realised loss feel twice as large, anchoring makes an entry price feel like value, confirmation makes the search return the answer the position needs, and tilt makes the next order louder than the plan. None of them can be argued with at the moment they are operating, because the argument is being conducted by them. That is why the countermeasure that generalises is not insight but pre-commitment — moving the decision to a time when the state is not the one that will be present when it matters. Pre-commitment has a specific form, and the form is what makes it work. A rule needs a trigger that can be observed without judgement — a percentage, a date, a close below a level, a count of consecutive losses — and an action that is unambiguous, ideally a change in size or a change in activity rather than a change in opinion. “Be more careful after a drawdown” has neither. “Halve the unit after a 6% drawdown from the month’s high, restoring it after two planned trades are followed” has both, and the second form can be executed by someone else, which is the test of a real rule. The reason a written rule beats a valued one is not that the trader lacks discipline. It is that the cost of the action is most vivid exactly when the trigger arrives: halving the size during a drawdown feels like accepting that the method has stopped working, and the case for waiting another week is always plausible in the moment. Written in advance, the same action costs nothing emotionally, because it was paid for by the trader who wrote it — and that trader had no drawdown, no open position, and no reason to argue. A preference and a rule — “I will be careful after a bad run”: No trigger, no action, no way to check compliance · “After two consecutive losses, no new trades this session”: A count as the trigger, an action, checkable by anyone ← · “I will trade smaller in a drawdown”: No percentage, so it can be satisfied by any size · “Halve the unit after a 6% drawdown from the month’s high”: A number as the trigger, a size as the action, and a restore condition ← The test of a rule is whether someone else could execute it from the written form alone. If it needs the author to interpret it in the moment, the author will be the one interpreting it in the moment.

The pre-mortem, and what to write down

The pre-mortem is the instrument that produces the rules before the position exists. The form is a single question asked before any capital is committed: it is a year from now and this decision went badly — write the story of why. The exercise works because it supplies prospective hindsight: imagining an outcome as having happened increases the number of causes a person can generate, and it does so without the defensiveness that arrives once the position is real. The output is not pessimism; it is a list of the specific ways this decision could fail, each of which can be turned into a trigger. The rules that come out of it cluster into four kinds, and it is worth writing one of each. A size rule, which states the risk per position and the total risk at once. A drawdown rule, which states the percentage at which activity changes and what the change is. A timing rule, which states when trades may be taken and when the session ends. And a review rule, which states the date on which the process is examined rather than the results. Each is a trigger, an action and a date, and each addresses a different part of the psychology: size addresses tilt, the drawdown rule addresses loss aversion and escalation, the timing rule addresses revenge, and the review rule addresses confirmation by putting the analysis on a schedule. Two properties are worth keeping in mind about the whole set. The first is that rules reduce the number of in-the-moment decisions rather than the amount of thought: a trader with four written rules spends the day deciding which setup is in front of them, which is the decision their expertise is actually for. The second is that a rule which is never triggered is not evidence that it was unnecessary — it is the state in which the rule was designed to be forgotten. The measure of the set is not how often it binds but what happens on the day it does. • A size rule: risk per position and total risk at once. • A drawdown rule: the percentage at which activity changes, and the change. • A timing rule: when trades may be taken, and when the session ends. • A review rule: a date for examining the process rather than the results. • Test each one: could somebody else execute it from the written form? A rule that is routinely re-interpreted is worse than no rule, because it produces the feeling of discipline without the effect. If a trigger has been met and the action did not follow, that is the finding for the next review — recorded, not explained.

Bright lines beat dials

There is a robust finding in the research on self-control, and it is uncomfortable for anyone who prides themselves on judgement: commitment devices work, and they work better when they are **automatic, costly to reverse, and absolute**. A savings commitment that deducts automatically outperforms the same intention held in the head. A rule with a threshold that removes discretion — no trading after a two-percent daily loss, no position without a written stop — holds up far better than an equivalent rule with a range and a judgement call. The reason is not moral. A rule with room for judgement has to be re-decided every time it applies, and each re-decision happens in the least favourable state: the one where the rule is inconvenient. The practical form of a bright line is a discontinuity. “I will not trade after two losses in a day” is a bright line; “I will be careful after two losses” is a dial, and dials get turned. The same principle applies to position sizing (“no single position above X%”), to instruments (“no naked short options, ever”), and to process (“no entry without a written invalidation”). Each of these converts a recurring decision into a one-time one, which is the entire mechanism: the value is in the removal of the negotiation, not in the wisdom of the particular number. The cost of bright lines is that they will occasionally be wrong. There will be a day when the third trade was the good one, and a bright line forbids it. That is the price of the device working at all — a rule you break when it is inconvenient is a dial with extra steps. Which is why the useful question when writing one is not “is this threshold optimal?” but “can I keep this on the day it hurts?”. A threshold you can hold is worth more than a better one you cannot, and the research on commitment consistently finds that the enforcement mechanism matters more than the calibration of the target. • Automatic, costly-to-reverse and absolute beats discretionary — the mechanism is what makes it work. • Turn dials into discontinuities: no trading after a limit, no position without a written stop. • The cost is occasionally being wrong; the alternative is a rule that is re-decided under pressure. • Choose thresholds you can hold on the day they hurt, not ones that look optimal.

What happens the first time you break it

A commitment that has never been broken is untested. The interesting question about a pre-commitment is not how it was written but what happens the first time it fails, because the ordinary outcome — quiet abandonment — means the rule was never real and the trader has learned that their rules are advisory. The sequence that makes a commitment durable has three parts, and skipping the third is what ends most plans. **Notice**: the breach is detected, which normally requires the rule to be written somewhere checkable and the record to be reviewed on a schedule; a rule that lives only in memory is breached without being noticed. **Consequence**: something specific and immediate follows, chosen in advance, and small enough that it will actually be applied — a reduction in size for a period, a return to simulation, a written explanation circulated to somebody. A consequence that is severe and therefore never enforced is worse than a mild one that is, because the unenforced version teaches that the rule is negotiable. And **re-commitment**: the rule is reinstated explicitly, with a line about what made the breach possible. Without that step, the middle step becomes a punishment cycle and the plan dies of it. The evidence on commitment devices points the same way. People will pay to restrict their own future choices when they know they will be tempted, and the devices that work are the ones that make defection costly or impossible rather than the ones that rely on resolve. A risk limit configured at the broker and a risk limit written in a document are different objects, and only one of them can survive a bad afternoon. The specific failure that kills most plans is an all-or-nothing interpretation: one breach means the plan is finished, and a spiral follows. The counter is to define the plan’s tolerance in advance — how many breaches, of what kind, within what period, constitute a real failure rather than an ordinary slip. That treats the plan as a system with a noise level, which is what it is, instead of as a moral test, which is how it fails. And there is the awkward case: sometimes the rule is wrong. A breach can be information that the rule was badly designed, and the pre-committed way to handle that is a scheduled review at which rules may be changed and the change logged. Changing a rule outside that review, and in immediate response to an outcome, is precisely the thing the pre-commitment existed to prevent — which is why the review has to exist as a legitimate outlet, or the pressure to change the rule in the moment finds no legal place to go. • Notice, consequence, re-commitment: skipping the third ends most plans. • A consequence that will not be enforced is worse than no consequence at all. • Define the tolerance for ordinary slips, or one breach becomes the end of the plan. • Give rule changes a scheduled review, or they will happen in the moment instead. The strongest pre-commitments are expensive to reverse — a hard limit at the broker, a contractual constraint, a lock-up. Where a rule cannot be made expensive, it should at least be made visible to somebody who will notice it was broken.

Sources

• Implementation intentions and pre-commitmentBehavioural economics literature on commitment devices
• Pre-mortems and prospective hindsightKlein, “Performing a Project Premortem”
• Rules versus discretion in trading plansSteenbarger, “The Daily Trading Coach”