What the remaining names actually are
Six chapters have taught you the relationships: where the close sits in the range, what a long shadow records, how two bodies can contain or penetrate one another, what a run of sessions does as it tires, and what location does to any of it.
Almost every name left in the catalogue is one of those relationships with an extra condition bolted on. A three-candle run where each session also opens at the previous close. A penetration that stops one third of the way in instead of halfway. A gap that is entered but not closed. The relationship is one you know; the name marks a particular degree of it.
That is why the rest of this chapter is a reference and not a course. There is nothing new to understand in it. There is a list of names, their measured conditions, and — for several of them — the earlier page where you already met the idea under a different word.
- Patterns implemented
- 67
- Core — taught properly, with a page of their own
- 23
- Variant — a stricter or confirmed form of a core pattern
- 16
- Advanced — kept as a reference, not drilled
- 28
- By length
- 12 single-candle, 23 two-candle, 23 three-candle, 3 four, 6 five
- Names in a typical charting app's list
- well over a hundred
The tiers are a claim, and it is written down
`scripts/candle-rules.js` sorts every pattern into core, variant or advanced, and that sorting is an editorial judgement rather than a fact about markets. Core means a beginner should recognise it. Variant means a stricter form of something already core. Advanced means it is here so a name can be looked up, not so it can be memorised.
Granthara publishes the tier with the pattern for the same reason it publishes the thresholds: a reader is entitled to see where the book drew a line and to disagree with it. Another writer would move several of these, and 04.05.04 showed how little separates a name from a near-miss.
What the tiers are not is a ranking of usefulness. An advanced pattern is not a better one. It is usually a rarer one, which means fewer examples, which means the little that can be said about it rests on less.
A detection library answers the question “what is this called?”. It has never once answered “what should I do about it?”
Why ticking every box makes the chart worse
Rahul's screen is now unreadable, and that is not a display problem. 04.05.04 ran this book's grader over a thousand boards built from noise and 732 of them earned a name. Turn on every detector and you are labelling the noise as diligently as the signal, because the detector cannot tell them apart and neither can the label.
There is a second cost. A name arriving unasked stops you reading the board. You see bullish harami before you have noticed that the range halved, that it happened at a price the market argued about last month, or that almost nothing traded. The label answers a question you had not asked yet, and it answers it first.
The reference works the other way round. You read the board, you notice the relationship, and then you come and look up what it is called. In that order the name is a filing label. In the other order it is a conclusion you did not reach.
How to use the five pages after this one
Each one takes a small family, gives the measured conditions the grader actually applies, and says where you met the idea before. They are short on purpose and they are not sequenced — nothing in 04.07.05 depends on 04.07.04.
Read one when you have met the name and want to know what it means. Skim them once now if you like, so you know what is in the index. Do not learn them.
And where a name is a second word for something you already know — kicking for kicker, counterattack for meeting lines, stalled for deliberation — the page says so plainly. A catalogue that has collected two names for one thing has told you something about the catalogue, not about the market.