Point and Figure Charts: Price Without Time
Point and figure charts track price without time: rising prices print a column of X's, falling prices print a column of O's, and nothing else on the chart exists. There is no time axis, no volume, no candle bodies. Each X or O fills one box of a fixed size, and a column only turns when price retraces by a set number of boxes. What remains is the purest structure chart in the toolkit: where price turned, how far it ran, and which levels it respected.
Two settings define the chart. The box size is the price distance one X or O represents. The reversal distance is how many boxes price must retrace before the chart abandons the current column and starts drawing the opposite letter. A one-box reversal makes a noisy chart; the conventional three-box reversal ignores moves smaller than three boxes against the direction, which is the point-and-figure version of the renko reversal rule. Between them, the two settings decide what the chart is willing to notice.
Columns, Not Candles
The mechanics are strict. While price rises, the chart keeps filling X's up the same column, one per box of advance. When price falls far enough to satisfy the reversal, the chart steps one column to the right and fills O's downward, starting one box below the X column's top. The same logic mirrors for a turn back up. A tall X column is a completed advance, measured in boxes; the O column beside it is the correction, also measured. The chart's history is a sequence of arguments between the two letters, and the taller column at any moment is the side currently winning.
Because every mark is tied to a price box, support and resistance read directly off the columns. A level where three separate X columns stopped is a resistance with three confirmations. A row of O columns all bottoming at the same box is a floor the chart has tested repeatedly. The point-and-figure chart was drawing horizontal levels decades before the phrase price structure became common, and the reason is structural: the chart literally cannot mark anything except prices that were traded and reversed from.
The chart also ignores gaps in a specific way. An overnight gap that lands price three boxes higher prints as three X's, same as a slow climb would. Sequence survives, magnitude survives, but the violence is gone. A trader who needs to know whether a level was walked or jumped reads the real chart for that answer. Long histories add one more scaling decision: fixed box sizes from years ago can be badly out of proportion to today's prices, so multi-year point and figure studies usually rescale the box as the instrument's price level changes, keeping the chart's proportions comparable across its whole span.
The Count: Turning Columns into Targets
Point and figure has a native way to project targets, and it comes from counting boxes. The vertical count measures the width of a base: count the boxes in a reversal column at the bottom of a range, multiply by the box size, and add the result to the breakout level for a bullish objective. The logic is proportional: the deeper the accumulation, the larger the count, and the taller the column the move can feed. A shallow base projects a small move; a wide base projects a large one.
The horizontal count works the same idea across time. Count the width of a sideways formation in columns, multiply by the reversal distance and the box size, and project from the breakout. Neither count is a promise. Both are arithmetic statements about how much back-and-forth built up before the break, and they give the trader a stated objective to test instead of an open-ended hope.
Breakout signals read directly too. A new X column crossing above the previous X column's top is the chart's buy signal; a new O column crossing below the previous O column's bottom is the sell signal. Everything else, the corrections and retests, fills the space between signals, and the three-box reversal keeps most of that noise off the page.
A Worked Example: One Box at a Time
The following numbers are invented for illustration, a hypothetical instrument with a one-dollar box and a three-box reversal.
The instrument rallies from 100 to 106: the chart prints X's at 100, 101, 102, 103, 104, 105 and 106 in the first column. Then it pulls back to 104. One box against the column: the chart prints nothing. A three-box reversal needs price below 103, and the pullback never gets there. The X column simply waits.
Price resumes and pushes to 109: more X's in the same column, 107, 108, 109. Now the correction comes: price falls to 102, three dollars off the 109 top. The chart steps right and begins an O column: O's at 108, 107, 106, 105, 104, 103 and 102. The reversal was admitted only because it crossed the threshold, and the O column now measures the correction in boxes, seven of them.
The read: the X column topped at 109, the O column is testing the old structure near 102 to 103, and the row of X's at 102 to 103 from the first column, sitting just left of the current O's, is the level the correction is arguing with. If the O column reaches 99, three boxes below that old row, the chart prints a sell signal by its own rules. Until then it prints nothing, no matter how the real chart squirms in between.
| Price event | Boxes against | What the chart printed |
|---|---|---|
| Rally 100 to 106 | None | Seven X's, first column |
| Pullback to 104 | 2 boxes | Nothing: under the reversal |
| Extension to 109 | None | Three more X's, same column |
| Fall to 102 | 7 boxes | New O column, 108 down to 102 |
The invalidation is the signal rule itself. A new O column crossing below the prior O column's low, or here, a break of the 99 threshold, hands the read to the bears by the chart's own definition. Nothing softer than that counts: point and figure does not grade doubts, it counts boxes.
Where Point and Figure Fits
The chart's strengths are structure and patience. Levels built from repeated column tops and bottoms are among the most honest resistance lines a trader can draw, because each mark is a reversal that actually happened at that box. The counts give bases a measurable size. The three-box reversal deletes most of what makes candle charts emotionally loud.
Its limits are the mirror of its strengths. No time means no session logic: the kill-zone behavior this block keeps returning to is invisible on the chart, so the timing must come from elsewhere. No volume means the chart cannot tell a well-attended move from a thin one. And small box sizes on fast instruments reproduce the noise the chart exists to remove, so the box choice carries the same weight as the renko brick choice.
The pairing with the rest of the block is direct: point and figure supplies the levels and the measured objectives, the real chart and the session windows supply the timing, and the bias decides which of the two letters to believe. The next lesson keeps the no-time philosophy but restores the bar: a chart where every bar spans exactly the same price distance.
Point and Figure Questions
What box size should a point and figure chart use?
One that makes the instrument's real legs print and its noise not print, usually a small fraction of the average swing being traded. The test is the same as the renko brick test: most corrections should die before reversal, and real reversals should print within a few boxes of turning.
Why is the three-box reversal the standard?
Because it sits at the balance point between sensitivity and silence. A one-box reversal records every hesitation and clutters the columns; larger reversals ignore genuine turns. Three boxes filters the corrections that never became moves while keeping the ones that did.
How reliable are point and figure count targets?
They are arithmetic projections, not forecasts: they state how much movement the base's size could feed. Treat them as objectives to test against the market's pools and the higher-timeframe direction, and take them off the table when the structure that produced the count is broken.
Can point and figure be used intraday?
It can, with small boxes, but its strengths show best on swings where structure matters more than session timing. Intraday, the missing time axis hides exactly the information the windows and macros exist to provide, so the two tools are better kept apart than merged.
X's and O's strip a chart down to reversals and levels. The next lesson adds one dimension back: bars of a fixed height, where every movement is measured in equal slices.