What Is a Point and Figure Chart?
What a point and figure chart is, what the X and O mean, how box size and reversal amount work, and why two P&F charts of the same market can disagree.
A point and figure chart plots price in columns of X's and O's instead of against a clock. X columns are rising prices, O columns are falling prices, each mark is a box worth a set amount of price, and a new column only starts when price moves the other way far enough to trigger a reversal.
Every number on a P&F chart is a setting you chose, which is why two traders can look at the same market and see two different charts. The box size can be set at least four ways, the price it reads can be set three ways, and two vendors name different settings as the one that can make earlier signals disappear.
What do the X and O mean on a point and figure chart?
X columns are rising prices and O columns are falling prices. Each X or O fills one box, and a box is worth a fixed amount of price that you set.
TradingView's help page gives two rules. A column holds only X's or only O's, never both. And X and O columns always alternate: you will never see two X columns or two O columns next to each other. If a chart shows two X columns side by side, something is wrong with the chart or with the description of it.
StockCharts adds: "Time is not a factor in P&F charting; these charts evolve as prices move. No movement in price means no change in the P&F chart."
Box size and reversal amount: the two settings that make the chart
The box size is how much price must move to add one mark. The reversal amount is how many boxes price must move the other way to start a new column. Multiply them and you get the reversal distance.
TradingView: "if the box size is set to $1 and the reversal amount is set to 3, the price must move $3 in the opposite direction to trigger a new column."
StockCharts says the same thing with the arithmetic written out: "A box size of one and the reversal amount of three would require a three-point move to warrant a reversal (1 x 3)."
The reversal amount is written in different units by different publishers. TradingView and StockCharts write it as a count of boxes, 3, and call the $3 the reversal distance. Alchemy Markets, in a page dated October 10, 2025, writes it as a price: "If the box size is $1, and the reversal amount is $3, you will need 3 Xs or Os to generate a reversal". The arithmetic matches, the labels do not, so a setting copied between two platforms needs checking rather than typing.
How do I choose the box size?
The box size is yours to pick, and no standard exists. Alchemy Markets says so: "The p f chart generally doesn't have standardised calculations."
What the main publishers offer:
| Source | Ways to set the box size | Reversal amount | Price it reads |
|---|---|---|---|
| TradingView help | ATR, Traditional, Percentage (LTP) | a count of boxes, example given is 3 | a Source setting, default Close |
| StockCharts ChartSchool | "Traditional, Percentage, Dynamic (ATR), and User-Defined" | its articles are based on the 3-box reversal method | High-Low Method or Close Method |
| Alchemy Markets | no definite formula, the box size can be any amount | written as a price, and it says many traders find 2 or 3 reasonable | High/Low, Closing Price, or Typical Price |
| Domo | fixed increments, or ATR based sizing, "for example, half the 14-day ATR" | "The standard is three boxes. One-box reversal charts exist but generate significantly more columns." | "either close-only or high/low pairs" |
| Highcharts docs | a boxSize API option | a reversalAmount API option | close only: the series "accepts the data only in [x, close] or {x, close} format" |
| Wikipedia | decide the value each X and O represents | traditionally one box, more commonly three | every price change in principle, end of day summary prices in practice |
Writers disagree about what that choice does to the chart's standing. TrendSpider lists "Objective analysis: P&F charts are objective and not influenced by personal biases or emotions" as a benefit. WallStreetMojo, updated Jun 18, 2025, lists the opposite as a limitation: "Subjectivity: The selection of box size and reversal criteria for a P&F chart can be subjective, leading to different interpretations and results among traders."
Both are describing different stages. Once the box size and reversal amount are fixed, the drawing rules are mechanical and leave nothing to interpret, which is TrendSpider's point. Picking those two numbers is a judgement, which is WallStreetMojo's, and that judgement comes first.
Which setting can change what you already saw?
Two of them, and TradingView and StockCharts name different ones.
TradingView, on its Percentage (LTP) method: "this method is subject to repainting, meaning it can change retroactively as new data comes in."
StockCharts, on its Dynamic (ATR) method: "The default is set at 20 days. However, one should take care with this setting because it changes as the ATR changes. Prior signals may disappear as ATR changes the box size."
Neither warns about the other's method. And the same named method is not the same setting on both platforms: TradingView's ATR length defaults to 14, StockCharts' Dynamic (ATR) defaults to 20 days, and Domo suggests half the 14-day ATR. A chart described as "ATR box size" tells you almost nothing until the look-back is written down too.
Does a point and figure chart use the close, or the high and the low?
Three answers, depending on the platform.
StockCharts: "There are two pricing methods available: the High-Low Method and the Close Method. Each method only uses one price point." It then gives the full high-low rules and the cost of them: "With only one data point used in the High-Low Method, it is possible to leave some key data on the table, or rather off the chart."
Alchemy Markets adds a third, typical price, "calculated as (high + low + close) / 3", which it calls somewhat rare.
TradingView makes it a setting rather than a method: Source selects which data point will be used for chart calculations, and the default is the close.
Highcharts, a charting library rather than a trading platform, settles it by not offering the choice: its point and figure series accepts the data only in [x, close] or {x, close} format.
Domo, writing for dashboard teams rather than traders, treats the input as something to record: "Document whether you're using close-only or high/low, how you handle splits, and what timezone your timestamps use."
The same price history produces different charts under the two methods, so the chart is not reproducible unless the choice travels with it.
Is time really gone from a point and figure chart?
No. Time is gone from the horizontal axis, and it is still in the data the chart is built from.
StockCharts: "Time is not a factor in P&F charting; these charts evolve as prices move." WallStreetMojo: "timeless as they do not factor in time." TrendSpider: "focus solely on price movements, with no regard for time", and it lists "Lack of time dimension" among the drawbacks because you cannot see the time spent at a price level.
StockCharts, on those same pages, marks months on the chart itself, with numbers and letters where each new month began, and says P&F charts do not show time linearly. Its scaling page says where the chart comes from: "The amount of past data shown on a P&F chart is not fixed. Instead, this look-back period is subject to the number of price reversals, which are influenced by box size, reversal amount, and data interval." On the same page: a 60-minute chart will use 7 data points per day, and using intraday periods with the high-low method captures more price action than daily data does. A 3-box reversal chart with a 1% box size, it says, will generally extend for 9 to 24 months.
Wikipedia says where the gap comes from: "The correct way to draw a point and figure chart is to plot every price change but practicality has rendered this difficult to do for a large quantity of stocks so many point and figure chartists use the summary prices at the end of each day."
TradingView has bar types that exist only because of the clock: projected up and down bars "Represent potential up moves during intraday timeframes, based on current prices before the close."
So a P&F chart does not show when things happened, and it is still cut from time-sliced data.
How is a point and figure price target worked out?
With a vertical count: count the filled boxes in the column that matters, multiply by the box size and the reversal amount, then apply the result to that column's high or low.
StockCharts gives both the rule and a worked case: "a column of 10 X's on a 1 x 3 P&F chart would yield 30 (10 x 1 x 3 = 30)", and "the fixed Count Column measures 14 boxes. Each box is 0.50, and the reversal amount is 3. The total Extension is 21 (14 x 0.5 x 3 = 21)."
StockCharts says the rules were never settled: "it is unclear who created P&F charting and the exact rules. As such, there are different interpretations involving counting methods, the counting column, the projection point, and the need for activation." It then names three published variants, as StockCharts describes them:
- Tom Dorsey, in Point & Figure Charting, uses a smaller multiplier for bearish targets, the box size and then 2/3 of the reversal amount, which is 2 on a 3-box chart. StockCharts says it uses that version for its own bearish counts.
- Jeremy du Plessis, in The Definitive Guide to Point & Figure, adds an activation point: the high of the count column has to be broken before the count counts.
- A.W. Cohen started the count from the high or low of the pattern rather than of the column.
StockCharts states the limit: "These Price Objectives are rough estimates based on P&F charting techniques. There is no guarantee prices will reach the objective."
How do I read a point and figure chart, step by step?
Pick the settings first, then walk the price.
This example uses round numbers. It is an illustration, not a real instrument and not a trade. Box size 10 points, reversal amount 3, so the reversal distance is 30 points, close method.
- Price is at 1000 and climbs to 1041. X's fill the 1010, 1020, 1030 and 1040 boxes. One column, four X's.
- Price slips back to 1012. That is 28 points below the top box, less than the 30 point reversal distance, so nothing is drawn. No new column, no new X, no mark at all, however long it takes.
- Price falls to 1005. That is past the reversal distance, so a new column of O's starts one box below the top of the X column and fills down.
A box is a range, not a price. StockCharts: "A P&F Box does not represent a single price, but rather a price range that depends on the box scaling method chosen, as well as the direction of the current column", and the range differs between a rising and a falling column. A column is also not finished until it reverses: "The number of filled boxes in a column is not fixed until there is a 3-box reversal", which makes any count taken before that preliminary.
Check a published example against those two rules before you copy it. WallStreetMojo's own text states the alternation rule, "every X column precedes an O column", and its Example #2 then describes two X columns in a row, the second of them covering a falling price range. The rule and the example on that page do not agree.
Point and figure versus Renko and candlesticks
All of these drop the clock in some way, and they drop it differently.
A candlestick chart gives you one candle per time period whether price moved or not. A P&F column grows while price keeps going one way, and only ends when price turns by the box size multiplied by the reversal amount. A Renko brick is placed by a set distance in the direction price is already moving, so the trigger is one number rather than two multiplied together. A range bar closes after a fixed amount of price movement. Heikin Ashi keeps the clock but averages the prices. TradingView groups the time-free ones: "PnF charts are similar to Renko, Kagi, and line break charts."
Wikipedia dates the family: "The technique is over 100 years old." It credits an anonymous writer it calls "Hoyle", who "was the first to write about it and showed charts in his 1898 book, The Game in Wall Street", and says the first dedicated manual came from Victor Devilliers in 1933. StockCharts tells the same story and credits A.W. Cohen with the classic 3-box reversal charts using X's and O's. TrendSpider tells a different one: P&F charts "were first used by Charles Dow, the co-founder of Dow Jones & Company, and his colleague, E. George Davis".
Can I use point and figure charts in ChartLabs?
No. ChartLabs has no point and figure charts. Its charts are time-based candles you replay bar by bar, plus one chart type that builds bars from price instead of the clock: range bars, where each bar closes after a fixed range, with entries and exits filled at the exact tick that crossed your level. Range bars currently run on NAS100, and the sizes offered come from a per-symbol list rather than a fixed menu. Every trade you take is logged against the chart you were looking at when you took it.
What does this change if I am testing a strategy bar by bar?
It means the settings are part of the test, not a display preference.
The box size has at least four methods and no standard. The reversal amount is three in most places and one in others. The price input can be the close, the high and low, or an average of three prices. The data interval behind the chart changes how far back it reaches. And two vendors warn that a box size which moves can erase signals that were on the chart earlier.
So a P&F signal is only reproducible if the box size method, the reversal amount, the price input and the data interval are written down beside the result. Domo, a business intelligence vendor rather than a trading one, says: "backtest your settings before trading live."
Doing that means stepping through the data and taking the trades yourself, with every setting recorded. Our own article on backtesting accuracy makes the general case about prices nobody could have traded, and a chart built from your own settings is a sharp version of the same problem.
If you want to test that way, with your rules and your settings, start a free trial. You get the full platform for 7 days.
Tags
Related Articles
Backtesting a Trading Strategy: A Beginner's Guide
What backtesting is, why it matters, and how to run one step by step: defining your rules, choosing clean historical data, stepping through the chart trade by trade, logging results, and reading the metrics that tell you whether your strategy has a real edge.
Backtesting Accuracy: How Slippage and Data Quality Matter
Why backtest results and live results drift apart, and what actually closes the gap: clean tick data, honest spread assumptions, and commission and slippage modelled directly on every trade you log.