Price, volume and time

Chapter 1 said a chart holds three columns. This chapter is what each one actually contains, because the imprecision starts earlier than people expect.

Which price?

"The price" is ambiguous. Over any interval there are four, and they answer different questions:

	What it is
Open	The first transaction of the interval
High	The highest transaction
Low	The lowest transaction
Close	The last transaction

The close is given a weight it has not earned. Almost every indicator in this subject is computed on closing prices — moving averages, RSI, MACD, all of it. But the close is simply the last trade before a bell rang. It is not an average, not a consensus, and not more informative than the trade a minute earlier.

It is used because it is convenient and conventional: one number per interval, available everywhere, comparable across days. Those are good reasons for a data format and they are not reasons to believe the number is special.

Two consequences follow. An indicator computed on closes is sensitive to what happened in the final minutes of a session, which is often when the least considered trading occurs. And two instruments whose sessions end at different times are not strictly comparable on closes at all.

What volume measures

Volume is the quantity that changed hands in the interval. Not the number of buyers, not the number of orders, and not net buying — every transaction has a buyer and a seller, so volume counts matched quantity.

"There was more buying than selling" is not a thing volume can show. The quantity bought and the quantity sold are identical by construction. What people mean by that phrase is usually that price rose while volume was high, which is a statement about price, with volume as a backdrop.

What volume does indicate is participation — how much of the instrument actually traded. That is genuinely useful, and chapter 7 takes it further with open interest.

Two cautions:

Volume is not comparable across instruments. A million shares of a company with ten crore shares outstanding is a different event from a million shares of one with ten thousand crore.

Volume has structure. It is routinely heavier at the open and close than in the middle of a session, heavier on expiry dates, and heavier when an index is rebalanced. Comparing today's volume with yesterday's without accounting for that structure compares the calendar as much as the interest.

Time, which is a choice

This is the part most people never examine. The time interval of a chart is selected by the viewer, and nothing in the data specifies it.

A one-minute chart, a daily chart and a monthly chart of the same instrument are built from the same transactions, grouped differently. Each grouping produces different opens, highs, lows and closes — and therefore different values for every indicator computed on top.

So the interval is a free parameter, and a free parameter is exactly what chapter 9's discussion of data snooping is about: when a method has adjustable settings and you try several, finding one that worked is not evidence that the method works.

The research literature is explicit that technical trading results depend on assumptions including the data period chosen — which is the same point, arriving from the testing side.

Working the problem

Daily chart falling, weekly chart rising, same stock.

Both are true, and nothing contradictory is happening. The two charts contain the same transactions, grouped into different buckets.

Suppose over ten weeks the stock rose from 100 to 130, but in the last four days it fell from 135 to 130. The weekly chart shows ten bars marching upward — a rise. The daily chart's recent bars show a decline. Each is an accurate summary of a different window.

Which is correct? The question has no answer, and noticing that is the point of the chapter. Neither interval is privileged. "Falling" and "rising" are not properties of the stock; they are properties of the stock plus a window length you chose.

Three things follow:

Any claim about a trend is incomplete without its horizon. "It's in an uptrend" means nothing until you say over what period. Chapter 3 is built on this.

Selecting the interval that shows what you want is trivially easy, and it will not feel like selection. Someone bullish will look at a weekly chart, someone bearish at the daily, and both will feel they are reading the market rather than choosing a lens. Anchoring and framing in Behavioural finance covers why this is so hard to notice from the inside.

The honest practice is to fix the horizon before looking, for the same reason the Measuring your return subject insists you fix a benchmark before you see the result. A horizon chosen after the fact is not a horizon; it is a conclusion.

The point

A price interval has four prices, and the close is used by nearly every indicator not because it is more informative but because it is one convenient number per interval. Volume is matched quantity, so it cannot show more buying than selling, and it is not comparable across instruments or across parts of a session. The time interval is a free parameter chosen by the viewer, so the same transactions grouped weekly and daily can show a rise and a fall at once — meaning "rising" is a property of the data plus your chosen window, not of the stock. Fix the horizon before you look.