Trend lines and the regression channel

Structure first

Before any drawing tool, the question is answered by reading the chart:

• Up — a consistent series of higher highs and higher lows.
• Down — a consistent series of lower lows and lower highs.
• Neither — and «neither» is a real answer that most people refuse to give.

That is a description of what has happened, and it is checkable: two people looking at the same swings should agree. A drawn line is a different kind of object, and it is worth being honest about which one you are relying on.

What a trend line actually is

It is a hypothesis with a visible failure point. You are saying: while price keeps respecting this slope, my reading holds; when it stops, I want to notice.

That is genuinely useful, and it is useful because of the second half. A line whose break would not change your mind about anything is decoration.

The conventions, which are worth following mostly so that your lines are consistent with each other:

1. Find the pivots — the points where direction clearly changed.
2. Two points define the line, and any two points define some line.
3. The third touch is what makes it interesting, because now the market has behaved consistently with something you drew before it happened.
4. Decide once whether you use closes or wicks, and do it the same way every time. Closes are steadier; wicks catch the extremes. What matters is not switching depending on which one supports the line you wanted.

And the honest caveat, which the playbook this site is built on states plainly: sometimes not every point touches, and the line is still describing the direction correctly. That flexibility is legitimate and it is also the door through which wishful drawing arrives.

The regression channel, and the half-truth about it

A regression channel does the same job with arithmetic. It fits the line that sits closest to all the prices in a chosen stretch — the least-squares fit — and then draws parallel bands at a set distance from it, usually measured in standard deviations.

It is often described as the objective version of a trend line, and that is half right in a way worth being precise about:

The calculation is objective. The window is not. A person chose where it starts, and the slope, the width and the position of every band follow from that choice.

What the bands measure

The outer lines describe dispersion: how far price has typically strayed from its own fitted average over that stretch. Price near a band is unusually far from that average, by that stretch's own standards.

That is a description, and it is where a lot of people quietly convert it into an instruction. It is worth stating the limit clearly: being far from an average is not a reason for anything. Markets that are trending strongly spend long periods pinned to one band, and «unusually far» is exactly what the beginning of a large move looks like.

Used as a description, it answers a genuinely useful question — how stretched is this, relative to itself? — and that question belongs in the context layer from the previous stage, not in an entry decision.

Moving averages are the same idea with a simpler shape: a line through recent prices that says roughly where the middle has been. Same use, same limit.

The rules that keep any of this honest

1. Read the structure before you draw anything. If the highs and lows do not agree with your line, the line is wrong, not the market.
2. Draw it once and leave it. A line that gets adjusted until it fits is a record of your opinion, not of the market's behaviour.
3. Write down what would break it before it is tested. Afterwards is too late to be honest.
4. Note the window. If you cannot say where your channel starts and why, you cannot say what it is measuring.

That is stage 3

Four articles about the picture itself. You now know that the bars are a choice, that a candle throws away the order of events, that a pattern is a name rather than a reason, and that a line is a hypothesis you can see.

Everything so far has been about reading a summary of what already happened. The next stage goes to the other side of the screen — what actually happens when somebody presses the button, and the orders that decide whether you are the one waiting or the one paying.