1 / 6 · 5 min

Findings, not signals: what a pattern really is

A pattern here is an event under a formal rule, not a «buy» tip. The difference sounds boring — and it is exactly what separates the testable from the sellable.

The word «pattern» is sold two ways. First: «here is a shape, price goes up after it» — a promise. Second: «here is a formal rule; here is how many times it fired this week; here is what price did after every single one» — a measurement. Our feed is built the second way, and this whole course rests on that difference. A pattern in the feed is not a forecast. It is a record: «at this moment price did such-and-such by a formal rule». What that event is worth is a separate question, answered not by the course author but by the «Statistics» table under the feed.

Why an honest feed looks humbler than a sold one

A signal seller shows five winning breakouts and stays silent about the rest. The detector records ALL of them: in one week the five-minute level breakouts alone number about two thousand, and after half of them price goes the other way. That is not a defect of the detector — it is the truth about the market: the event «price closed beyond a level» is by itself a weak predictor of the next move. The value appears one step later: when kinds, timeframes and context start to differ — and the differences get measured.

Worked example

One card, read in full

The card: «SOL · Level breakout · 15m · up · level 103.40 · volume ×2.1 · Bybit». In the detector's words: a fifteen-minute bar closed ABOVE 103.40 by more than the tolerance (a quarter of a typical bar), the previous bar was still below the level, and the bar's volume was at least one and a half times the median of the last twenty — here ×2.1. Under the card, four numbers: after 15 minutes +0.4%, after an hour −0.2%, after 4 hours still «pending», «now» −0.1%. The sign is TOWARDS the formation: plus means price went where the breakout pointed. This particular breakout did not hold for an hour — and the card does not hide it.

What the detector guarantees — and what it does not promise

  • Guarantees: every event is recorded by one and the same rule, with no manual curation and no deletion of failures.
  • Guarantees: the 15m/1h/4h outcomes are computed from real minute candles, not reconstructed after the fact.
  • Guarantees: the rule is readable — the thresholds of every kind are described in the guide and in this course.
  • Does NOT promise: that price will follow the formation. That is measured, never promised.
  • Does NOT promise: that a formation is an entry. Entry, stop and size are your job — this course teaches it.
Common mistake

Reading the feed as a list of trades

The costliest beginner mistake: open the feed, see twenty fresh cards and decide these are twenty invitations into the market. The feed is raw material. Out of two thousand weekly events, a handful of trade candidates survive the filters: a kind with a real skew in the statistics, a higher timeframe, a liquid coin, a fresh card rather than yesterday's. Whoever trades everything pays, in fees and spread, for exactly the half that goes the wrong way.

Open the feed and read any three cards aloud in the detector's words: which kind, which timeframe, which level, which volume, what the outcomes showed. If you cannot do it without hints yet — that is normal: the next lessons take every kind apart threshold by threshold.

Open the pattern feed
Check yourself

Under an upward breakout card it says «1 hour: −0.8%». What does that mean?

An hour after the finding's bar closed, price was 0.8% BELOW the close — that is, moving AGAINST the breakout's direction (the sign is counted towards the formation). The card is not «wrong» and not «broken»: it honestly recorded that this breakout did not hold for an hour. Records like this are what make the statistics under the feed truth rather than a shop window.

Check yourself

Why doesn't the detector delete failed findings — the feed would look better?

Because the «Statistics» table would then become a lie: the break and bounce shares are computed over ALL decided findings. Deleting failures is precisely the trick signals are sold with. The value of our feed is not beauty — it is that its numbers can be trusted when you choose which kinds are worth trading at all.