note

green small marubozu delicate breaking out
from green tiny inverse hammer in
river
in
cliff bottom alp
in
alp breaking out
in
river

it looks so counterintuitive. you would want to start with big river patterns, and then sort them into alps and whatever else, and only after more sorting look at little dragonflies. But, no, it's not right! And there's a reason we need to sort first by tiny little patterns, but I will only speak of it indirectly. I will only say: they are the most mysterious.

the key

There are two basic classification into which all chart patterns can be divided:

Red bars.

Green bars.

Both subdivide the same way:

tiny bars, short bars, medium bars, long bars, very long bars.

Each bar length subdivides the same way:

solid or hairless bars, hammers, dragonflies, and inverse hammers.

These types also subdivide alike:

ethereal, delicate, sturdy, and stout. Hairless doesn't subdivide.

After giving it a lot of thought I've largely concluded that the classification needs to start with the smallest observable pattern, one bar. After that it could progress to the second smallest observable pattern, two bars, and then to three bars, and so on until an entire chart is classified. Of course that is nonsense, and after two or three bars we are going to start talking about some larger feature, such as tops and bottoms. (Then again, we might find it's beneficial to be very minute in our analysis, sometimes.) I'm mesmerized, because it seems like there's a proof in the data that classification must begin with the smallest observable feature, and ... actually, I think I know what it is ... now.

sponsored.