Indicator library · Inputs

Typical Price, (High + Low + Close) ÷ 3

One line of arithmetic that several of the measures on this site quietly depend on. It reduces a whole bar to a single number, and the sessions where it disagrees most with the close are precisely the sessions worth reading.

The calculation

Typical price = (high + low + close) ÷ 3. That is all of it, and the interesting part is not the formula but why anything bothers.

A bar is a summary of a period: it has a range, and trading was distributed across that range. Any measure that wants to weight a bar by its volume has to decide at what price that volume traded, and the close is a poor answer: it is one print at one moment, and on a session that spent the day at one level and spiked at the end it is unrepresentative of almost every transaction that took place.

The typical price is a crude fix for that: three points of the bar, equally weighted. It is not a price anything traded at, and it is a better description of where the session lived than the close is.

How far the close sits from the typical price, by bar shapeA bar chart of the absolute difference between the close and the typical price for four constructed bars. A bar closing at its high and a bar closing at its low both show a difference of about 0.67 points on a two-point range; a bar closing mid-range shows zero, and a narrow bar closing flat shows zero.points between close and typical priceCloses at the highclose 52.00 · typical 51.33The two inputs disagree by more than a third of the bar’s rangeCloses mid-rangeclose 51.00 · typical 51.00Closes at the lowclose 50.00 · typical 50.67The two inputs disagree by more than a third of the bar’s rangeNarrow bar, closes flatclose 51.00 · typical 51.00How far the close sits from the typical price, by bar shapeA bar chart of the absolute difference between the close and the typical price for four constructed bars. A bar closing at its high and a bar closing at its low both show a difference of about 0.67 points on a two-point range; a bar closing mid-range shows zero, and a narrow bar closing flat shows zero.points between close and typical priceCloses at the highclose 52.00 · typical 51.33The two inputs disagree by more than a third of the bar’srangeCloses mid-rangeclose 51.00 · typical 51.00Closes at the lowclose 50.00 · typical 50.67The two inputs disagree by more than a third of the bar’srangeNarrow bar, closes flatclose 51.00 · typical 51.00
Fig. 1: arithmetic, not market dataAll four bars are constructed and computed at build time. The two that close at an extreme of a two-point range disagree with their own typical price by 0.67 points, a third of the range, while the mid-range bar and the narrow flat bar disagree by nothing at all. That is the awkward part: the substitution is invisible on the sessions nobody cares about and material on the sessions everybody reads.

Where it is already doing work

Most people meet this formula without noticing, because it sits inside other measures rather than being plotted itself.

Measures built on the typical price
MeasureHow it uses it
VWAPMultiplies it by each bar's volume and accumulates, so the running average reflects where the shares changed hands.
Money Flow IndexMultiplies it by volume for raw money flow, and takes its direction against the previous bar to classify the flow.
Pivot pointsThe pivot itself is this number, computed from the previous session; every support and resistance level is derived from it.
Commodity Channel IndexMeasures how far the typical price sits from its own moving average, in units of mean deviation.

Two of those, VWAP and the Money Flow Index, are volume measures, and that is not a coincidence. As soon as a formula multiplies by volume it has committed to an opinion about where the volume traded, and the typical price is the cheapest defensible opinion available from a daily bar.

The variants, and why they matter less than consistency

Three alternatives circulate and each is defensible: (H + L + C) ÷ 3 as here, (O + H + L + C) ÷ 4 which includes the open, and (H + L + C + C) ÷ 4 which double-weights the close. They differ by small amounts on ordinary data and by more on the sessions that matter.

Which one you pick matters far less than not switching. Two VWAP lines built on different variants will differ persistently by a small amount, which looks exactly like an implementation bug and is not one, and it is a common reason two platforms cannot be reconciled. If a figure is going to be published or compared, the variant belongs beside it.

Why it is not a weighted average

Equal weighting of the three points is worth one paragraph of scepticism, because it is the assumption everything downstream rests on. A bar's high and low are each a single print, often a single trade, while the close is the level a whole session settled at. Giving all three the same weight treats one extreme tick as being as representative as the settlement.

That is why the weighted-close variant exists, and why the choice is a convention rather than a derivation: nobody has shown that any particular weighting is correct, because "correct" would require knowing where the volume actually traded, which is the thing the formula is standing in for. The honest position is that the typical price is a cheap approximation that is better than the close and worse than a measurement.

Where it misleads

Known failure modes
SituationWhat goes wrong
Treated as a traded priceNothing changed hands at it. Comparing it directly with an actual print or a previous close mixes two kinds of number.
Variants mixedIncluding the open, or double-weighting the close, shifts every downstream figure by a small persistent amount that reads as a bug.
Wide-range sessionsEqual weighting assumes trading was spread across the bar. On a session that gapped and sat still, it was not.
Unreliable extremesThe high and low depend on which venues the feed includes, so two providers give slightly different typical prices.
Smoothed furtherA moving average of typical prices is an average of averages, and two steps from anything that happened.
Used where the close is the conventionRSI, MACD and most oscillators are defined on closes. Substituting this input silently makes your figures incomparable with everyone else's.

What volume adds

The typical price assumes trading was spread evenly across the bar, and volume is what tells you whether that assumption held. A session that traded steadily all day is described well by the average of its three points; a session that gapped, printed its high in the first minute and then sat quietly at the bottom of the range is not, and the volume distribution says so immediately.

This is the argument for a volume profilewhere the data allows one: it replaces the assumption with a measurement, showing which prices inside the bar the volume actually went to. The typical price is what you use when you cannot afford that, which for most daily-bar work is most of the time.

Frequently asked questions

What is the typical price?

The average of a bar’s high, low and close, divided by three. It reduces a whole session to one number in a way that acknowledges the session had a range, unlike the close, which is one print at one moment. It is sometimes called the pivot price, because the same arithmetic is the pivot in the classic pivot-point formula.

Why use it instead of the close?

Because trading happened across the whole bar, not only at the end of it. Any measure that weights a bar by its volume — VWAP, the Money Flow Index — is implicitly asking where that volume traded, and the close is a poor answer to that question: it is the last print, and on a session that spent the day at one level and spiked into the close it is unrepresentative of almost every transaction.

When does the choice actually change an answer?

When the close sits far from the middle of the range, which is exactly the session type technical analysis cares about. On a bar that closes at its high, the close and the typical price differ by a third of the range; on a narrow bar that closes flat they are within a cent of each other. So the substitution is invisible on quiet data and material on the sessions where something happened, which is the wrong way round for anyone hoping it is a detail.

Which indicators use it?

VWAP weights it by volume to build the session’s average transaction price. The Money Flow Index multiplies it by volume to get raw money flow and takes its direction to classify each bar. Commodity Channel Index and several money-flow variants use it as their base. Pivot points use the identical arithmetic under a different name. In every case the reason is the same: the measure needs one number that represents a whole bar.

Is there a version that includes the open?

Yes, the average of open, high, low and close, sometimes called the average price or OHLC/4. It is defensible and slightly steadier, and it is less common because the open was historically unreliable in American equity data and because the extra term dilutes the close, which is the price most participants actually reference. If you use it, say so: it produces a visibly different VWAP.

What is the weighted close?

A variant that gives the close double weight: (high + low + close + close) ÷ 4. It sits between the typical price and the close itself and exists for people who want a bar-representative number that still leans on the settlement price. Like every choice on this page it is a convention, and the only real error is switching between conventions inside one comparison.

Does it matter on intraday bars?

More than on daily bars, because intraday closes are arbitrary in a way daily closes are not. A five-minute bar ends when the clock says so, not when trading settles, so its final print carries no special meaning at all, while its high, low and close together still describe where the five minutes were spent. On intraday data the typical price is the better default rather than merely a defensible alternative.

Can it be used as a series in its own right?

It can, and it is occasionally plotted as a smoother alternative to a closing-price line. The caution is that it is not a price anything traded at. It is an average of three prices, so a moving average of typical prices is an average of averages, and a comparison against an actual traded level such as a pivot or a previous close is comparing two different kinds of number.

Where to go next