How to read a price chart

What the drawing is made of 

Underneath every chart is a stream of quotes with timestamps. A firm's systems record a new quote whenever the price it is willing to deal at changes, which in an active market is many times a second. That stream is far too dense to draw, so a chart chops it into intervals of equal duration, one minute, one hour, one day, summarises each interval as a single mark, and lays the marks out left to right in the order the intervals occurred.

Summarising is where information is lost, and the three common chart types differ only in how much of it they agree to lose. Within any interval four prices are worth naming: the first price recorded, called the open, the highest and the lowest recorded during it, and the last price before the interval ended, called the close. A line chart keeps one of them. A bar chart and a candlestick chart keep all four. Nothing beyond those four survives in any of them: the path price took between them, and the order in which the high and the low occurred, are discarded by every chart type equally.

The line chart 

A line chart takes one price per interval, conventionally the close, plots it as a point, and joins consecutive points with a straight segment. The segment is drawn, not observed: no price is claimed to have travelled along it, and prices during the interval that ran above or below either endpoint leave no trace. An interval that opened calmly, ran a long way, reversed and closed near where it began therefore appears as a point almost on top of the previous one. Traditions that prefer line charts treat that as a feature rather than a loss, on the argument that a close is the one price a large number of participants had to agree on, while an extreme may reflect a single quote nobody dealt on. That argument is contested, and the section on disagreement below returns to it.

The bar chart 

A bar chart draws each interval as a single vertical stroke whose top sits at the highest price recorded and whose bottom sits at the lowest, so the stroke is the range. Two short horizontal ticks attach to it: one on the left at the open, one on the right at the close, left being earlier and right later, the same convention as the horizontal axis. All four prices sit in that one mark and their relationship is legible at a glance: a right tick above the left tick means the interval closed above where it opened, and the distance from either tick to the end of the stroke shows how far price ranged beyond them before returning. The format is compact and visually quiet, which is why it survives on screens showing many intervals at once.

The candlestick chart 

A candlestick chart plots exactly the same four prices as a bar chart and differs only in how it encodes them. The distance between the open and the close is drawn as a filled rectangle, called the body, and the remaining travel to the high and the low as thin lines from its ends, called wicks or shadows. The body's fill states which of its two edges was the open: one convention for an interval that closed above its open, the other for an interval that closed below.

No information is added. A candle and a bar drawn from the same interval contain identical data, and either can be reconstructed from the other exactly. What changes is what the eye reaches first: a bar makes the range prominent because the stroke is the range, while a candle makes the open to close distance prominent because it is the only filled area on the mark. The anatomy of a single candle is taken up in the lesson after this one.

Which side of the quote the chart is drawn from 

A quote has two prices at every instant and a chart draws one line, so something has to be chosen, and the choice is a setting rather than a fact about the market. Most CFD charts are drawn from the bid. Some are drawn from the mid, the arithmetic middle of the two quoted prices, which is a number nobody deals at. Some platforms plot the second side as an additional line on request. The platforms YAL provides, MetaTrader 5, each state in their own settings which side a chart is drawn from, and a level read off one side is not a level on the other.

The consequence is concrete. On a chart drawn from the bid, the low of a candle is the lowest bid recorded in that interval. The offer at that instant sat above it, and by how much is invisible, because the offer was never plotted. A price level that appears to have been touched on a bid chart may not have been reached on the offer side at all.

The same reasoning explains why two firms' charts of one instrument do not lie on top of each other. Each draws its own quote from its own sources, so closes differ slightly and extremes differ more, since an extreme is the most unusual quote one firm happened to record. A level identified to the last decimal place on one chart is approximate on another.

Reading the axes 

Four things identify what is on screen, and all four are stated by the chart rather than inferred. The instrument name and the interval sit in the header. The horizontal axis carries time, oldest at the inline start, newest at the inline end, each mark occupying one fixed slice regardless of how much happened during it. The vertical axis carries price. Timestamps carry a timezone, and the one a platform stamps its intervals with is the server's rather than the reader's, so one interval can be labelled with two different clock times on two screens.

The price axis has two properties worth checking before a chart is read as steep or flat. It usually auto-scales, fitting itself to the highest and lowest price in the visible window, so scrolling or zooming changes the apparent gradient of a line that has not changed at all. And its spacing follows one of two rules: on a linear axis equal vertical distances are equal amounts of price, while on a logarithmic axis they are equal percentage changes, so a given amount of price occupies more height at a low level than at a high one. Over a long window covering a wide percentage range the two draw identical data as visibly different shapes.

Two smaller conventions complete the reading. Where a market was closed the chart records nothing, so the weekend and the holidays covered in the sessions module appear as a discontinuity between adjacent marks rather than a flat stretch. And a panel below the price often plots volume, which on a CFD chart is usually tick volume, the number of quote updates in the interval, rather than a quantity of contracts traded, because a bilateral contract has no central record of quantity to count. Two different measurements share one name, and a chart rarely says which it is showing.

Where practitioners disagree 

The first disagreement is whether the extra data in a bar or a candle is information or noise. Systematic traditions frequently work from closes alone, and the argument has a specific technical basis: a bar reports its high and its low but not which came first, so a sequence of bars cannot be replayed as a sequence of prices, and any rule depending on the order in which two levels were reached cannot be examined on bar data without an assumption the data does not contain. Discretionary traditions answer that the extremes are where the interesting information sits, since they mark the points at which the market stopped going one way, and that a cleaner series discards the event along with the noise. Neither position has settled the matter.

The second is the price axis. Logarithmic spacing is conventional in traditions comparing moves across long horizons or across instruments at very different price levels, on the argument that a percentage is comparable where an amount is not. Linear spacing is conventional over short horizons, and its defenders note that a logarithmic axis distorts the geometry of anything drawn as a straight line, which matters for the trend lines taken up later in this module. The axis is a presentation choice: it changes the picture without changing the data, and an observation whose validity depends on which spacing was selected is an observation about the drawing.