Who Decides Where a Range Begins and Ends

A range is an interval plus two extremes

Saying that a market has been ranging is an incomplete statement, in the way that saying a car has been fast is incomplete. Over what stretch? The word describes a relationship between price and a window of time, and the window is not in the data. Somebody picked it.

That is not a technicality, because the choice is most of what the description is made of. Take a stretch of price and extend the window backwards far enough and the whole thing becomes one leg of something larger; shorten it and the same prices become two moves with a turn between them. Nothing about the prices changed. What changed is how much of them is being looked at, and that was a decision by the person describing.

The boundaries are whichever extremes the window happens to hold

Once the window is fixed, the top and the bottom of the range are not decided by the market either. They are the highest and lowest prices inside that window, which means an entire boundary can rest on one print that went a little further than the rest.

This is why two descriptions of the same period can put the edges in noticeably different places without either being careless. One takes every extreme at face value; the other treats a single outlying print as unrepresentative and draws to where price spent its time instead. Both are defensible. Neither is more factual than the other, and the choice between them is one more parameter, usually unstated.

A session high is a fact about a convention

The high of a session needs a definition of when the session began and ended, and sessions are conventions rather than events. The trading centres people name are open during local business hours, which move with local holidays and clock changes, and nothing in the price itself marks the boundary.

A daily high has the same character for a more mechanical reason: a platform's day starts at whatever hour its server clock is set to, so two platforms with different server days can report different daily extremes from the same market without either being wrong. And an opening range is this again with a stopwatch — its whole content lives in how long the opening is taken to be, and nobody standardises that. In each case the price is real and the boundary is a decision, which is why a statement about a session or daily high is only checkable when it says whose day and which convention.

A channel adds a slope, and the slope is a fit

A channel is a range that has been allowed to tilt: two lines drawn to contain price, no longer horizontal. Everything above still applies, and one thing is added, because the slope has to be fitted to something.

Which touches count is the fitting decision. Include a shallow touch and the structure tilts one way; treat it as unrepresentative and it tilts another. No procedure settles this, so the tilt is a property of the drawing at least as much as of the market. It is worth noticing when a channel is described as having been drawn by the market. Nothing was drawn by the market. Somebody chose the touches, and the resulting angle is the visible part of that choice.

Quiet is a comparison, and you supply the other half

Consolidation, noise and momentum all sound like properties of price and are all relations. Consolidation means quieter than something — a previous stretch, an ordinary day, whatever the describer had in mind. Momentum means moving faster than something, over a chosen span. Change the reference and calm becomes lively while the prices sit exactly where they were.

Noise is the clearest illustration, because it is defined as a remainder. It is whatever is left once a filter has removed the part being treated as the movement that matters, so it is a property of the filter at least as much as of the market. What one description discards as noise, another with a shorter window describes as the thing that happened. Neither has discovered anything; they used different filters and then named the leftovers.

A measured move is arithmetic on a distance that already happened

Measuring the extent of a completed move is a fact: it went that far, in that time, on that chart. Taking that distance and placing a copy of it somewhere else on the axis is a different kind of act, and the chart does not distinguish the two once both are drawn on it.

The projection's authority comes from the convention that says to do it, not from anything in the price, and the convention supplies no reason for the distance to repeat. This is worth saying plainly, because a projection looks exactly like a level once it is on screen — a line, at a price, extended to the right — and the two have completely different standing. One is a measurement of something recorded. The other is a mark at a price nothing has yet visited.

The parameter belongs in the statement

None of this makes describing price movement a waste of time. It makes a description incomplete when it travels without the settings that produced it, which is nearly always how it travels.

A description worth keeping says the window, the timeframe, whose feed and whose server day, the boundary convention, and what the comparison is against. Stated that way it can be tested by somebody else, and disagreed with in a way that identifies exactly where the disagreement lives. Stated without them it is not false — it is simply not the sort of thing that can be checked, and a description of the market that cannot be checked is indistinguishable from a description of how the market felt to whoever wrote it.