Point-and-Figure Charts and Price Target Estimation

A point-and-figure chart (P&F) records movement in predetermined price units. No new box is added until price completes the required movement. Movement in the same direction extends the existing column, while a sufficient reversal starts a new column. Removing calendar time as the equally spaced horizontal unit emphasizes price movement, reversals, and horizontal structure. Wyckoff analysis uses horizontal P&F counts to convert the sideways activity formed within a trading range into a conditional estimate of potential subsequent price distance. Changing construction rules, price inputs, or count boundaries can change both chart and target, so every estimate must specify all parameters.

26.1 Point-and-Figure Charts Versus Time-Based Candlestick Charts

A time-based candlestick chart produces a bar at fixed intervals whether or not price changes materially. A P&F chart updates when price reaches a grid line: it may change many times in one day or add no boxes for several days. A horizontal column does not represent a fixed number of hours or days, and column width cannot directly indicate structural duration. The number of columns occupied by a trading range depends on the number of price reversals, box size, reversal rule, and input data, not simply calendar length.

Candlesticks preserve open, high, low, close, and corresponding volume, making them suitable for assessing price-volume behavior, closing location, gaps, and event sequence. Traditional P&F charts mainly preserve the results of movements across boxes and reversals. They normally do not show volume in each box and compress fluctuations that fail to reach the threshold. This highlights support, resistance, breakouts, and horizontal buildup, but cannot replace supply-and-demand assessment on candlesticks. In practice, first identify a trend or trading range on candlesticks, then check price structure and counts on P&F, and finally return to candlesticks to define the trigger, invalidation, and risk.

A P&F chart marks only movements across one box or the required reversal distance, so it omits smaller fluctuations and does not arrange observations at equal time intervals. This highlights sideways repetition and directional changes but does not show how long a move took or changes smaller than one box. Smaller boxes or fewer reversal boxes usually reveal more local fluctuations; larger parameters usually emphasize major swings. Grid origin, intrabar processing order, and column-switching rules also affect column and reversal counts, so changing parameters does not necessarily change column counts in a fixed direction. The same history can produce several valid P&F charts under different rules and levels of detail. Use identical parameters and price-data types when comparing a group or recounting.

A horizontal count provides potential price distance, not a future price path. Price may approach the target in stages, form new structures en route, encounter supply or demand early, or never move in the expected direction because the original structural judgment was wrong. Target estimates enter trading decisions only in combination with structural direction, market context, and risk boundaries.

Time-based candlesticks aggregate open, high, low, close, and volume over fixed periods, whereas P&F updates only when price movement meets specified conditions. Each additional horizontal column does not mean the same amount of time has elapsed: a fast market can form many columns quickly, and a quiet market can remain unchanged for a long time. P&F is therefore useful for observing price progress, turns, and horizontal column counts, but column spacing cannot reveal elapsed days. Volume normally is not part of classic chart-construction conditions; software overlays should be treated as additional information. In trading, P&F can estimate the price range of a horizontal cause, while candlesticks still establish formation sequence, price-volume behavior, and actual triggers. The two charts cannot replace each other.

26.2 Box Size, Reversal Boxes, X Columns, and O Columns

Fixed box size, denoted b, must be greater than zero and represents the price distance between adjacent vertical grid lines. First establish a grid origin g; all valid grid lines are g+k×b, where k is an integer. Keep the origin fixed thereafter. Resetting the grid after price rises puts the same historical price into different boxes and prevents consistent manual checking. This chapter uses fixed price box sizes and does not mix percentage-based box sizes with fixed-box formulas.

The reversal count, denoted r, must be a positive integer. It specifies the minimum number of boxes price must move against the current column's extreme before recording a reversal. X usually denotes a rising column and O a falling column. An X column extends upward, each new X showing that price reached at least the next complete higher grid line; an O column extends downward, each new O showing that price reached at least the next complete lower grid line. In a multi-box reversal chart, an adverse move smaller than r×b creates no new column and is filtered out. If price subsequently makes a new extreme in the original direction, the reversal distance is recalculated from that updated extreme.

Box size determines vertical resolution; reversal count determines the movement required to change column direction. Reducing b or r usually increases symbols and columns, exposing finer local structure; increasing them combines smaller swings into a simpler chart. Parameters change both the column count and multiplier in horizontal counting, so an old target cannot simply be rescaled after a parameter change. Every chart should specify the instrument, data period, box size, reversal count, grid origin, whether it uses closes, OHLC, or tick prices, and the software algorithm.

When price falls between two grid lines, record only the fully reached line. A price that touches a line exactly counts as reaching that box under the predetermined convention. If an instrument has a minimum tick size, box size should be compatible with it to avoid grid lines at untradeable prices. Corporate actions, price adjustments, futures rolls, and continuous-contract construction alter historical price distances, so establish adjustment and contract-splicing methods before constructing the chart.

26.3 Price Data and Construction Rules

The standard column-switching procedure in this section applies to multi-box reversal charts. For a one-box reversal involving a one-step-back move, use the exception in 26.4.

Construction begins by specifying initialization. The first price normally establishes only an initial reference box; it cannot determine an X or O first column by itself. A subsequent first move to a complete grid line above the reference establishes X direction and fills the crossed boxes; a first move to a complete line below establishes O direction. If software initializes from the first price segment or a manually specified direction, record the initial reference price, first-column direction, and rule, because initialization may affect a small section at the far left of the chart.

With tick prices, process trades one by one in their actual sequence. In an X column, first check whether the new price reaches one or more boxes above the highest X. If so, fill successive Xs upward in the same column. Only if the column does not extend and price reverses r boxes from the highest X should an O column start to the right, filling downward from the box immediately below the highest X to the lowest complete box reached. Mirror this procedure for an O column: first check for downward extension; if it cannot extend and price rises by the reversal distance from the lowest O, start an X column to the right, filling upward from the box immediately above the lowest O.

With closing-price data, process only the close of each period. This method is easy to reproduce but ignores intraperiod boxes reached and then retraced before the close. Tick or finer transaction data preserve actual sequence but require more data. Whichever method is chosen, do not directly combine column counts from close-based and high-low-based charts.

When building P&F from OHLC data, the order of the high and low within a bar is usually unknown, so processing order must be fixed beforehand. A common convention is extension before reversal: first check whether price can continue the current column; if it can, do not assess reversal on that bar. For X columns, first use the high to test upward extension; only if there is none use the low to test downward reversal. For O columns, first use the low to test downward extension; only if there is none use the high to test upward reversal. This is a consistent calculation convention, not a reconstruction of actual intrabar order. If the order affects the research conclusion, use finer data.

For manual construction, process each price observation sequentially: establish the current column's direction and extreme box; check whether same-direction price reaches a new complete box; if so, fill all crossed boxes and update the extreme. Otherwise check the reversal distance from the current extreme and, if met, start the opposite column to the right and fill its boxes. If neither condition holds, leave the chart unchanged. Software may use different rounding, initialization, percentage box sizes, or intrabar priority rules, so manually check a short sequence observation by observation before using an automated chart.

26.4 Basic Differences Between One-Box and Multi-Box Reversals

In multi-box reversal charts, X and O columns normally alternate, and every new column expresses at least the specified reversal movement. A reversal smaller than r boxes leaves no mark; once it reaches the threshold, the new column begins next to the prior extreme and fills each crossed box with the opposite symbol. This logic suits standard charts with r greater than one and makes each column's direction clear.

A one-box reversal chart is not produced simply by setting r to one in the same procedure. The traditional One-Box One-Step-Back convention allows Xs and Os to coexist in a special column to handle consecutive one-box reversals. The key is how to handle a new reversal column containing only one mark: suppose a one-box decline has just created a single O to the right of an X column, and price immediately rises one box. Place an X above that single O in the same new column; do not start another column or return to the original X column on the left. The current column thus contains both O and X.

Once that column contains at least two recorded symbols, a subsequent qualifying reversal starts a new column to the right under the rules. The falling case is exactly symmetrical: after a single X is created to the right of an O column, an immediate one-box decline adds an O beneath that X in the same column. One-step-back handling accommodates tight price swings and means the statement that a column can never mix X and O is false under this convention.

Software implementations of one-box charts have differed. TradeStation explicitly describes one-step-back handling; TradingView also distinguishes classic one-box construction from merely setting the reversal parameter to one. To check software, construct a short sequence containing a one-box reversal followed by an immediate return, and verify whether symbols occupy the same new column and when the next column begins. One-box column counts and structures must not be combined with multi-box charts without explanation.

26.5 Wyckoff Horizontal Counts: Count Line and Counting Range

Horizontal counting begins with structural assessment. Identify the same trading range, expected direction, and key tests on both candlestick and P&F charts, then select a count row representative of that structure's horizontal development. A trading range is countable not because its columns store measurable physical energy or verifiable inventory, but because comparable upward and downward supply-and-demand swings recur within the same price range. Column count compresses these horizontal movements meeting the reversal threshold into a scale measure under consistent rules. Upward counts commonly trace left from near an LPS or final supply test through columns belonging to the same base or reaccumulation structure. Downward counts commonly trace left from near an LPSY, final demand test, or last weak rally through the top or redistribution structure. The count row must relate to structural location, not simply be the row yielding the most columns.

The counting range has left and right boundaries. The right boundary normally lies near the test initiating the count or the completion of the structure; the left returns to where that portion of the cause began. Explain boundaries through price behavior known at the time, such as a trend-stopping event, a clearly separate subrange, or a structural transition. A later extended move cannot justify moving the left boundary farther back; doing so chooses inputs to fit the outcome.

Denote the raw horizontal column count N. Count every column between the left and right boundaries, including blank columns containing no X or O on the selected count row. The count measures the range's span in the horizontal column sequence, not the number of characters on the row, trading days, volume, or vertical range height. If software displays converted points or target distance directly, first check whether the value already includes box size and reversal count. Do not multiply a converted result again.

A range may have more than one reasonable count row. Check adjacent rows where structure permits and express results as candidate zones, but state the row-selection rule beforehand. Count credibility comes from consistent rules and transparent structural interpretation, not from finding the row that happens to fit later highs and lows.

Horizontal counting first selects a fixed price row, then counts the consecutive columns belonging to the target horizontal structure along that row. The count line should pass through a row representative of horizontal activity, and the left and right boundaries must be determined by structural events identifiable at the time, with reasons retained. Column count is neither candlestick count nor elapsed days; additional time within the same column does not increase the count. If moving the row by one box causes a large difference, report that sensitivity instead of keeping only the largest result. Phase labels help interpret structure but cannot force count boundaries to coincide exactly with a lettered phase.

26.6 Segmented Counts and Their Difference from Phases A–E

Phases A–E describe the developing function of a trading range: stopping the prior trend, building cause within the range, a possible test, leaving the range, and subsequent progress. Segmented counting is an operational division on the P&F chart, dividing one large horizontal structure by structural events into complete column segments along the same count row. One segment can span multiple phases, and one phase can contain several local segments. The two cannot be mapped mechanically.

Segmented horizontal counts within the same large trading range always use the same count row, box size, reversal rule, and grid. Work from right to left, starting nearest the final test. For an upward count, the structurally complete segment between the LPS and Spring can be the first segment, with N1 columns. The complete segment farther left from the Spring to an earlier ST can be the second, with boundary-inclusive count N2. Count the shared boundary column only once, so N cumulative = N1 + N2 - 1. Only if N2 is defined as additional columns excluding the shared boundary is N cumulative = N1 + N2. Mirror this procedure for downward counts using LPSY, UT, or the corresponding top test as anchors.

Each extension must add a complete count segment defined by structural events, rather than updating the target with every additional column. First save the initial boundaries, N1, and projection; then save the second segment's new left boundary, additional columns, overlap adjustment, and cumulative N. Repeat when extending farther left. Calculate every cumulative target using the same count row and projection convention. This produces both a nearer first-segment target and cumulative targets incorporating more cause, while clearly identifying the historical segment responsible for each expansion.

An independent reaccumulation or redistribution platform formed after a trend advances is a different case, often called a stepping-stone confirmation count. The new platform has its own structural boundaries and count row, so determine N, D, and projection independently. Its candidate targets can be compared with the original base targets to see whether they cluster or support trend continuation, but price distances from two independent structures cannot be mechanically added. The former method accumulates segments within one range on one count row; the latter independently counts different horizontal structures within a trend. Distinguish them in the names of your records.

Do not force segments into counts for Phase A, Phase B, and so on. Phase labels may change with new evidence, while P&F boundaries must correspond to actual horizontal price columns. Two parts separated by a clear trend segment should usually be treated independently; dividing ordinary back-and-forth movement within the same range may count the same cause twice.

Short counts offer nearer potential moves, while long counts describe candidate zones for larger structures. After price reaches a nearby zone, reassess supply, demand, and new structure; a distant count is no reason to ignore reversal evidence. Converging targets can make a zone more useful for management, but counts sharing the same price history are not statistically independent evidence.

Segmented counting studies subcauses formed sequentially within a structure: for example, an early stopping zone and a later testing zone may support different segments of progress. Stage counts within the same large structure use one count row and preserve their start and end columns separately. If two stages share columns, subtract the overlap when totaling so the same horizontal activity is not counted twice. Independently formed reaccumulation or redistribution structures may use their own appropriate count rows and projection anchors, but their projections cannot be added arbitrarily. Phases A–E are defined by events and supply-and-demand evolution, while P&F columns arise when price meets box-size and reversal conditions. There is no one-to-one relationship. A column may span multiple events in time, and a phase may form many columns. Segmented targets help plan progress and management locations; they do not guarantee that price will pause at each stage target in order.

26.7 From Count Results to Target Estimates

For fixed-box, multi-box reversal charts, the basic conversion of a raw horizontal count to price distance is:

D = N × b × r

N is the raw number of columns, including blank columns; b is the price distance per box; r is the dimensionless reversal count; and D has units of price. N and r are counts; only b supplies the price unit. This product is a traditional empirical projection rule converting horizontal scale under the current parameters into a candidate vertical distance. It is not a conservation law, does not prove that the market has accumulated an equivalent amount of energy, and cannot identify unsold inventory. If software has already converted its count using b and r, that output is itself a distance and must not be multiplied again. Percentage boxes, logarithmic charts, and other automated target algorithms also cannot directly use this fixed-box formula.

After obtaining D, disclose the projection anchor. Following the Wyckoff Analytics P&F guide, this course uses the range low L and count-row price C to construct an upward candidate zone: the more conservative candidate is L + D, and the other is (L + C)/2 + D. For downward projections, use the range high H: the more conservative candidate is H - D, and the other is (H + C)/2 - D. Adding or subtracting D directly from the count row may provide a more distant reference projection, but must not silently replace the conservative target.

A complete calculation record lists chart parameters, input price data, structural direction, count row, left and right boundaries, raw N, distance D, range high and low, anchor formula, and final candidate zone in sequence. Another analyst can then pinpoint the source of a disagreement over boundaries without guessing how the software produced its result.

Interpret target zones using the instrument's minimum tick, actual support and resistance ahead, and executable conditions. Precise decimals in a calculation do not imply equally precise forecasting. Significant supply overhead or demand below makes the nearer structure a management zone. P&F projections estimate potential distance only; actual prices must move through a continuous auction and may not reach the target directly.

With fixed box size b, reversal count r, and column count n, traditional horizontal counting first gives the price magnitude n×b×r, then adds or subtracts it from the chosen projection anchor according to long or short direction. Units must be consistent: n and r are dimensionless, and b is price units per box, so the result is a price magnitude. If software already outputs a converted magnitude or target, do not multiply again by b or r. Treat targets as candidate zones: discrete boxes, count boundaries, and anchor selection all introduce a range. Structural resistance or demand ahead may make the usable target nearer. Reaching a target means the result of the original cause has found expression; it does not mean price must reverse.

26.8 Parameter Changes, Recounting, and Hindsight Selection of Boundaries

Changing box size changes the threshold for each box; changing reversal count changes when columns switch; changing grid origin changes box placement; changing input fields or timeframe changes intrabar touches and sequence. These changes affect more than b and r in the multiplication: they regenerate the column sequence and change N. Parameter sensitivity checks therefore require rebuilding the chart, identifying the same structure again, and recounting, not merely changing formula inputs.

Choose parameters according to the trading scale before seeing subsequent results, and keep them stable within a research series. A coarse and a fine chart may be maintained simultaneously to observe different scales, but each needs an independent record and their targets must not be selectively mixed. Similar zones from different reasonable parameters suggest methodological stability; large differences call for wider expression of target uncertainty and less weight on the estimate in decisions.

Common recounting biases include extending the left boundary after the event, selecting a row producing the desired target, or repeatedly adding new columns to the old cause after each advance. Prevent them by saving screenshots and parameters before the breakout or execution, fixing the boundaries known then and the conditions under which new counts may be added. A subsequently formed independent, complete horizontal segment may receive a new record; do not overwrite the old one because its outcome is unsatisfactory.

Data revisions may also require recalculation. Changes in stock adjustment methods, re-splicing continuous futures contracts, and removing erroneous trades all alter the historical grid. Label recalculations with version and reason. Failure to reach a target does not automatically prove faulty construction, and reaching it does not prove correct boundary selection. Evaluate whether inputs are transparent, rules consistent, and outdated estimates promptly retired when structure invalidates.

Robustness checks should vary a reasonable box size, reversal count, count row, and boundary at least once to see whether targets remain in similar zones. The purpose is to identify parameter sensitivity, not to choose the preferred answer. When structure expands, genuinely new columns may be added, but preserve the old version and never add trend columns that were originally outside the base after seeing the result. If a fixed total count and stage counts cover the same columns, show both separately but do not add them to claim a larger target. A rule requiring knowledge of a future high or low to select the counting range has no real-time trading value.

26.9 Connecting Target Estimates with Actual Trading Risk

The P&F target first asks how much price room may exist if the structure develops as expected; entry and stop ask how much risk must be taken to test that hypothesis. For a long, potential reward can be calculated as the conservative target minus planned entry, less fees and slippage. Planned risk is the distance from entry to the predefined protective stop, plus fees and slippage. The protective stop should have a clear market basis but need not coincide with the ultimate invalidation of the larger structure. Reverse the direction for shorts. Align contract multipliers, quotation currencies, and exchange rates. The entire distance from target to stop is not potential profit.

The classic nine tests commonly require at least 3:1 potential reward to risk. This is a planning-stage filter, not a win rate, performance guarantee, or basis for proportionally increasing a position. A conservative target too close or a structurally necessary stop too far away means insufficient room at that entry price. The response may be to wait for a better location, reduce research priority, or abandon the trade, rather than move the stop to a price without structural meaning.

Targets also serve position management. As price enters a nearer projection zone, observe progress, volume, opposing supply or demand, and whether a new trading range forms. If movement is healthy and produces a new independently countable structure, the distant target may remain a reference. If clear opposing evidence appears or the original hypothesis invalidates, follow the risk plan even before the target is reached. Targets are conditional estimates; protective boundaries determine when to exit. Their functions differ.

Position size is calculated from planned risk and stop distance, subject to liquidity, gap risk, correlated exposure, and portfolio limits. A distant P&F target does not make a trade safer and cannot compensate for unclear price-data types, adjustment or continuous-contract methods, or a late entry. A complete P&F estimate must state the structure on which it rests, its parameters, its conservative price room, and the price facts that will cause it to be removed from the decision process.