Every indicator in the previous three chapters is a transformation of price. Volume is the one genuinely independent measurement available on a standard chart, which is why ‘confirm with volume’ is repeated so often — and why it is worth being precise about what volume can and cannot tell you.

Volume is the number of units traded in the period. Every trade has a buyer and a seller, so ‘buying volume’ and ‘selling volume’ do not exist in the raw number. When someone says volume was buying, they mean either that trades occurred at the ask rather than the bid — which requires trade-level data with quote context, not the volume bar — or they are describing the price direction and calling it volume.

This matters because it separates two very different things. Book data — resting orders — is spoofable and shows intent. Traded data — prints — is commitment but requires inference to assign a side. Conflating them is the most common error in order-flow analysis, and being able to state the distinction cleanly is a strong signal in any market-data or trading conversation.

Volume-weighted average price is the average price of the session weighted by volume at each price. It matters for a reason unrelated to charting: it is a benchmark institutions are measured against. An execution desk filling a large order is frequently judged on whether it beat VWAP, which means real flow is drawn toward it — a genuine, mechanical, non-mystical reason for a level to matter.

• Session VWAP resets daily and is the standard intraday reference.
• Anchored VWAP starts from a chosen event — an earnings date, a major high, the start of a move — and answers a specific question: what is the average price everyone who traded since that event paid? That is a more meaningful reference than an arbitrary moving average, because it corresponds to a real population of positions.
• The honest caveat: the anchor is chosen by you, which reintroduces subjectivity. Anchor to a mechanically-defined event — an earnings date, a 52-week extreme — rather than to a point you liked the look of.

Open interest is the number of derivative contracts currently outstanding. It is not volume: volume counts transactions in the period, OI counts positions that exist. A trade can increase OI (both sides opening), decrease it (both closing), or leave it unchanged (one opening, one closing).

The standard four-way reading is worth knowing, and worth holding loosely:

The caveat that matters: this framework assumes a directional interpretation of positions that may not hold. A large share of Indian option open interest is hedged or part of a spread, and deep in-the-money puts, for instance, function as synthetic long stock rather than as a directional bearish position. Reading raw OI as directional sentiment without accounting for that is a common and expensive misreading.

Indian exchanges publish, per stock per day, what fraction of traded volume resulted in actual delivery to a demat account rather than being squared off intraday. This is genuine information that most markets do not provide.

• High delivery percentage means a larger share of the day's activity was investors taking positions rather than intraday traders churning.
• It is published with a lag — in the end-of-day bhavcopy — so it is not usable intraday. Any strategy claiming to use delivery percentage during the session has a lookahead problem.
• Test it properly. A delivery spike combined with a price move is a reasonable hypothesis and, when tested with a matched control, the effect is typically weak and slow — real enough to be interesting as context, not strong enough to be a standalone signal. That is a perfectly good conclusion and a far more useful one than the confident claims you will find elsewhere.

A volume profile plots volume by price rather than by time, over a chosen window. The useful objects are the point of control (the price with the most volume), high-volume nodes (prices where a lot of business was done), and low-volume nodes (prices price passed through quickly).

The reasoning is sound and mechanical: prices where large volume transacted are prices where many participants have positions, so activity concentrates there again. Prices that were traversed quickly have few positions, so price tends to move through them quickly again. This is a better-founded version of ‘support and resistance’ because the level is computed from actual transactions rather than drawn.

The caveat, as always: the window is a choice. A profile over the last 20 days and one over the last 200 give different levels, and neither is correct. Fix it in advance.