How we verify

Why a trading journal is publishing its verification methodology

Most tools in this category ask you to trust a black box. You see an insight, something like “you’re overtrading” or “you cut winners short,” and you’re supposed to just believe it, because an algorithm or an “AI” said so.

We think that’s backwards. If we’re going to tell you something about your own trading, you should be able to see how sure we are, what we checked before we said it, and what we’re honest enough to admit we can’t yet tell.

This page is that account. Not a marketing claim about how rigorous we are, but the actual shape of the process, kept up to date as it runs. We update it as the process itself evolves, and we keep a running log of every time we go back and re-check work we’ve already shipped (linked at the bottom of this page).

The promise

Three rules govern everything on this page and everything we ever say publicly about your trading:

• Every claim we make is true, and we can point to where it’s proven. If we say we tested something at a certain scale, that test exists and produced that result. If we can’t prove a number, we don’t publish it.
• Any scarcity we mention is real. If we say a cohort is limited, or a waitlist has a cap, that limit is genuine, not a countdown timer designed to manufacture urgency.
• We don’t invent evidence. No fabricated testimonials, no manufactured case studies, no cohorts that don’t exist. If we don’t have real users saying something yet, we don’t put words in an imaginary one’s mouth.

That’s the whole promise. It’s boring on purpose. Boring is what “we’re not selling you a story” looks like in practice.

The discipline

Every behavioral pattern we build (the things we might eventually tell you about your own trading, like a tendency to widen a stop instead of taking a loss) goes through a fixed, nine-part discipline before it’s allowed anywhere near your coaching. We don’t publish what each part checks for in operational detail, since that’s the part a competitor could copy, but we publish that all nine exist, and in plain terms, what kind of problem each one guards against:

1. A voice ceiling matched to how sure we actually are. How confidently we’re allowed to talk about a pattern is capped by how well-established its underlying cause is, not by how compelling the story sounds.
2. A pattern-specific honesty caveat. Every pattern has its own particular gap between what the research actually measured and what our number actually computes. That gap gets stated, not smoothed over.
3. A gaming defense, authored fresh for every pattern. Every pattern gets checked against the specific ways a trader could accidentally or deliberately make it fire incorrectly, or fail to fire when it should. There is always more than one independent check, never just one number that could be gamed.
4. A rule for when the underlying cause is genuinely disputed. When the behavioral science itself doesn’t agree on why something happens, we describe what we observed and decline to assert a cause we can’t back up.
5. Apples-to-apples comparison, always. A pattern is only ever measured against a trader’s own comparable trades, never blended across dissimilar setups in a way that would manufacture a signal that isn’t really there.
6. The number and the sentence must match. Whatever figure a pattern computes has to be the literal thing the coaching sentence describes to you, not two related-but-different numbers dressed up as one insight.
7. Immunity to good behavior. A pattern built to catch a harmful habit has to be built so that the opposite, healthy version of that behavior can never accidentally trip it.
8. A resolution rule for thin data. When there isn’t enough of a trader’s history to say something with confidence, the pattern has to know how to go quiet, or go careful, rather than guess.
9. Independence from any one asset class. A pattern about human behavior has to hold up whether you trade forex, futures, or anything else, not secretly depend on the mechanics of one specific market.

That’s the shape. The actual mechanics inside each of the nine (the thresholds, the formulas, the specific gaming checks) are the part we keep to ourselves. Naming the discipline is the flex; handing over the recipe would defeat the point of having one.

How a pattern earns its place

A pattern doesn’t reach your coaching the day someone has an idea for it. Here’s the path, described in shape only:

1. Drafted. Someone on our side proposes a pattern and writes a first version against all nine parts of the discipline above.
2. Adversarially audited by reviewers who didn’t write it. Independent reviewers, sealed off from each other and from the original author, try to find every way the pattern could be wrong, gamed, or overconfident. They don’t see each other’s findings until both are in.
3. Tested against a blind, held-out panel. The pattern is run against a set of trade profiles it’s specifically supposed to stay silent on, to check it doesn’t fire where it shouldn’t. That is a negative test, not just a positive one.
4. Confidence is split into two kinds, and both are checked separately. We check the mechanism (do we understand why this happens) and the threshold (is the specific number we chose actually calibrated) as two different questions, so a well-understood cause doesn’t quietly borrow credibility for an uncalibrated number, or vice versa.
5. Locked. Once a pattern clears every step, it’s version-locked. From that point on, any change to it is a deliberate, tracked revision, not a silent edit.
6. Re-audited on a cadence. Locked doesn’t mean forgotten. We re-check our own already-shipped patterns on a recurring basis, using new or tightened versions of the same discipline, and we publish that we did it. See the audit log linked below.

What we deliberately don’t publish

We assume competitors read this page. That’s fine. That’s the point of drawing the line where we do.

We publish the standard we hold ourselves to. We don’t publish:

• The specific formulas, thresholds, or numbers behind any pattern.
• The internal logic of how a pattern actually gets detected in your data.
• The exact wording and decision rules behind how our coaching talks to you.
• The specific profiles or data we test patterns against.
• Anything that would hand another team a shortcut to rebuilding what took us real, checked work to build.

If a line on this page ever tells you that we check something without telling you how we check it, that’s not an oversight. That’s the line, on purpose.

The numbers we will and won’t show you

What we’ll show you: Our data layer has been tested to 20 million trades. Reads stay under 18 milliseconds at that scale, and isolation and arithmetic are bit-identical from 10,000 rows to 20 million. The one exception: our analytics-speed test slows under a load no real account reaches; it never loses or mixes data. That was a scale and persistence test on a large synthetic dataset built to stress the system the way a large, growing user base eventually would. It is not a claim about live trading results, and we’re not going to blur that line to make the number sound bigger than it is.

What we will never show you: a win-rate promise, a profit projection, or any claim that using this product will make you money. No one can honestly make that promise to you, and anyone who does is selling you a story, not a tool. We don’t do income claims, full stop.

Our sources

We build openly on the behavioral-finance and trading-psychology canon: the real, named, published research and practitioner work on how traders actually behave under pressure. We don’t pretend we invented behavioral finance. When we lean on a well-established idea (loss aversion, the disposition effect, the psychology of overtrading), we’re glad to say so.

What we don’t do is copy anyone’s specific frameworks, worked examples, or proprietary language wholesale. Everything that reaches your coaching is rewritten in our own words, checked against our own discipline, and grounded in your own data, not recycled from someone else’s book with the serial numbers filed off.

Want to see this discipline applied to a real, already-shipped pattern? Read the audit log, a running record of every time we’ve gone back and re-checked our own work.

Curious what this looks like in your own trading? Join the waitlist. No deposit, no card, just an honest look at whether this is built for the way you trade.