The circuit-breaker fired on day three

$ eldr --debrief 2CB7

POSITION     MOVE (conditional)
CONFIDENCE   7/10
WINDOW       60d
STOP         Apr 30 · 8 paying customers · else close
TRIGGERED    yes, 3 days after review date
OUTCOME      acted (stopped)
SATISFACTION 4/5
SOURCE       constructed file

This is a constructed worked example, not a customer. Nobody has run a read here yet: the product has not launched and the decision log is empty. The situation and the numbers are written to show how the instrument reads a file of this shape. Real debriefs replace it when there are any.

A circuit-breaker is a decision you make in advance about a state of the world you expect not to be in. This file is the only reason I believe they work, and it is also the file that shows what they cost.

The setup

A side product, built evenings, launched in February. $16,000 of his own money committed to ads and contract design. In the intake he wrote a stop condition unprompted, which is rare, since most men have never been told to write one, and the read did what it always does with a stated stop: it treated it as the window boundary and made the position conditional on it.

The stop, in his words: "If I do not have 8 paying customers by April 30, I close it and I do not spend another dollar."

Three parts, all present: a date, a number he could not argue with, a pre-committed consequence. Most stop conditions are missing at least one of the three, and it is usually the consequence.

What happened on April 30

Five paying customers, two more in a trial that had not converted, and a warm conversation with a firm that would have been number eight if it had signed.

That is the worst possible outcome for a stop condition. Not a clear failure — a near miss, with a plausible story attached about why the real number was almost eight.

The four days

He re-ran the intake on May 1 with the customer count as the only changed input, which is the pattern the multi-read debrief in this library describes. The new read returned MOVE at 5/10 with a 30-day extension, and — this is the part he flagged himself — the counter-case named exactly what he was doing:

You wrote this stop condition in January, when you had no emotional position in the answer, and you are now asking for a second opinion on whether to honor it from a version of yourself who has spent $16,000. The January version of you is a better judge of this than the May version, and he already ruled. If you are going to overturn him, overturn him explicitly: write down what you now believe that you did not believe in January, and write a new stop with a new date and a smaller budget.

He tried to write that sentence and could not make it into a fact. Everything he produced was a version of "it feels close". On May 4 he closed the product, canceled the ad account, and logged the outcome.

What the file establishes

A stop condition works by moving the decision to the version of you with the least invested, and its entire value is realized on the day you want to break it. On any day where the number is clearly met or clearly missed, the stop does nothing. It exists for the near miss, which is the only case that needed it.

Falsifiable prediction attached, and it is a prediction rather than a result: men who write the consequence clause in advance should honor the stop at a materially higher rate than men who write only a date and a number. Nobody has tested that, this product has no corpus to test it against yet, and when it does the sample will be self-reported outcomes with no control group. Take it as a hypothesis with a mechanism.

The cost, stated honestly

The eighth customer may well have signed in June. He closed a product with five paying customers and real conversations in the pipeline, because of a number he picked out of the air in January with no information. That is not discipline, it is an arbitrary constraint executed faithfully, and faithfully executing an arbitrary constraint is how you lose the things that were slow to start.

That objection is correct and it is unresolvable — nobody can know what June contained. What can be said is that the alternative he was choosing between was not "stop versus one more month". It was "stop versus a series of one-more-months decided by whoever he happened to be at the end of each one". He had already watched that film once, in a previous file, with a different product and $30,000.

Six weeks later the file sits at 4 out of 5, with the honest summary being that nobody can know whether stopping was right, only that he is not still in it and that in the previous version of this pattern he would have been.

What made this one hold when the previous one did not

There is a second file behind this one, eighteen months earlier, on a different product. Same shape, no stop condition, $30,000 and a year. The difference between the two is not discipline as a character trait — it is one sentence written in January by a man with nothing at stake yet.

That is the mechanism worth taking away, and it generalizes past money. A circuit-breaker is not willpower and it does not require any. It is a transfer of authority from the version of you who will be tired, invested and looking for a reason, to the version of you who is none of those things right now. You are not being strong in May. You are obeying January, and January was a better judge because January had no sunk cost.

Which also explains why the consequence clause is the part everyone omits. A date and a number feel like planning. A pre-committed consequence feels like betting against yourself, and men do not enjoy writing it down. Write it anyway, in the same session, before the first dollar moves.

The instruction

Write the consequence clause. The date and the number are the easy two-thirds and they are not the part that saves you.

For entertainment and reflective purposes only. Not a substitute for professional financial, legal, medical or mental-health advice.