Why point-and-figure looks different

Point-and-figure (P&F) charts record price movement in columns of Xs for rising prices and Os for falling prices. New columns require a defined reversal. A uniform time step is not the horizontal axis; the chart updates when price moves enough under its box and reversal settings. The 2026 CMT Program Guide includes P&F construction and basic patterns in Level I public objectives. Level II objectives add complex patterns and price projections on arithmetic and log scales. That establishes the topic's scope, not a promise that a particular count method will be tested.

Before interpreting any P&F chart, write its box size, reversal amount, price field (close or high-low), and scaling rule. A “buy signal” on one setting can be absent on another. See the StockCharts ChartSchool vertical-count method for one documented implementation; alternative conventions exist.

Construct a simple 1×3 chart

Take a fixed box size of 1 price unit and a three-box reversal. An X column contains Xs at 10, 11, 12, 13, and 14. Price then moves down to 12. That is only a two-box move from 14, so the X column remains. When price reaches 11, the decline is three boxes; a new O column forms with Os at 13, 12, and 11. The former X column does not erase its Xs.

If price later rises from 11 to 14, a new X column begins with Xs at 12, 13, and 14. A later rise to 15 adds another X in the same column. These moves illustrate how the reversal filter suppresses smaller counter-moves. The exact plotting of a bar that spans both continuation and reversal thresholds depends on a platform's high-low ordering rule; close-only data avoids some ambiguity but loses intrabar information. Document the data mode.

With a 0.5 box and a three-box reversal, the minimum opposite move is 0.5 × 3 = 1.5 price units. Smaller boxes make the chart more sensitive. A larger reversal amount suppresses more direction changes. Neither choice is universally superior; an analyst can accidentally tune settings to make historical patterns look cleaner.

Basic pattern logic

A double-top breakout occurs when an X column rises above the previous X-column high under the platform's box convention. If a previous X high is 20 and the current X column reaches the 21 box with one-point boxes, the new column exceeds it by one box. That is a P&F price event, not a guarantee of continuation. A double-bottom breakdown similarly occurs when an O column undercuts a prior O-column low.

A three-box reversal does not mean a “three-day” reversal; the price might move three boxes in one session or over many. Nor does a P&F chart show the elapsed waiting time evenly. If timing matters, inspect a conventional chart alongside it.

P&F support and resistance often appear as repeated columns topping or bottoming near the same box. A 45-degree line in a square-box display can be a structural reference, but its appearance depends on chart scaling. A bar-chart trendline and a P&F trendline are not interchangeable.

A vertical-count objective

One documented vertical-count convention counts the boxes in a completed X column, multiplies by box size and reversal amount, and adds that extension to the low of the count column. Suppose a completed breakout X column spans boxes 20 through 25 inclusive: six Xs. With box size 1 and reversal amount 3, extension = 6 × 1 × 3 = 18. Adding to the count-column low of 20 gives an illustrative bullish objective of 38. The inclusive count matters: 20, 21, 22, 23, 24, and 25 are six boxes.

The count column is not fixed until a qualifying reversal produces a new O column. Before that, more Xs can raise the objective. Other practitioners choose different count columns, multipliers, or activation rules. The StockCharts method documents this variation. A target at 38 is a projection method, not a minimum move or a forecast probability.

A horizontal count uses the width of a congestion pattern instead of the height of an X or O column. Because the count line, starting level, and active-signal conventions vary, do not mix a horizontal count from one method with a vertical count from another. State the chosen method whenever reporting a number.

Decide what would refute the signal

Suppose a bullish double-top signal appears at the 21 box, with the prior O-column low at 16. One conditional plan is to treat a later O breakdown below 16 as invalidation. A tighter risk plan might use a closer support box, but then it should expect more whipsaws. A P&F objective of 38 without the 16 failure level omits the most important risk information. If entering near 21 and using 16 as a rough price invalidation, the distance is 5 units per share before gaps; the projected distance to 38 is 17. The ratio 17/5 = 3.4 is descriptive and does not establish expected profit.

Also check the original price data. Thin markets, splits, or a futures roll can produce artificial columns. Percentage box scaling differs from fixed arithmetic boxes, so a constant “one-point” count does not apply.

Common mistakes and self-check

Do not read P&F columns as equally spaced days. Do not change box settings after seeing a signal and then compare performance fairly. Do not count an unfinished X column as fixed. Do not assume all software uses the same vertical-count convention. P&F removes some visual noise but not market risk.

Self-check: With a one-point box and three-box reversal, an X column tops at 30. Price falls to 28, then 27. When does an O column begin, and what boxes are plotted? Answer: A fall to 28 is two boxes and does not reverse. At 27 the three-box threshold is met, so Os are placed at 29, 28, and 27 under the simple fixed-box convention.

Next, hand-plot ten closing prices with one fixed box setting, then change only the reversal amount and explain which signals disappeared and why.

Charts load as you approach this section.