Why the Same Range Can Have Different Interpretations

23.1 Primary Hypotheses, Alternatives, and Neutrality

Before a trading range is complete, accumulation and redistribution, or distribution and reaccumulation, often look similar. Analysis must begin with observed action and ask which explanation best fits current evidence. Provisionally rank the better fit as the primary hypothesis and retain the other as an alternative. They must make different predictions for the next move: which side price leaves, whether it remains outside, and where a retest should hold. The ranking indicates a current preference, not a guarantee of future direction.

Conditional probability can describe this judgment, but subjective impressions cannot supply invented numbers. Let H_A be the accumulation hypothesis, H_R the redistribution hypothesis, and E a newly observed class of price behavior. Their relationship is: P(H_A|E)/P(H_R|E)=[P(H_A)/P(H_R)]×[P(E|H_A)/P(E|H_R)].

This relationship follows from two conditional decompositions of the same joint probability: P(H_A)×P(E|H_A)=P(E)×P(H_A|E). Applying the same decomposition to H_R and comparing cancels the common P(E), giving the formula above. These are two expansions of the same joint event; reversing the observation order does not reverse market causality.

E must be a predefined, repeatedly recognizable class of behavior, such as 'crossing the upper boundary and returning inside within the specified observation period.' A unique exact price path cannot be designated as the event ad hoc. All probabilities must also condition on the same existing information, and every probability used as a denominator must exceed zero.

The left side is the relative likelihood of the two hypotheses after observing E. The first factor is their relative assessment before E, based on the existing trend and range behavior. The second is the likelihood ratio: the probability of E under accumulation relative to its probability under redistribution. A ratio above 1 shifts judgment toward accumulation; below 1 shifts it toward redistribution. Only when the two conditional probabilities are similar and the ratio is near 1 does E have weak discriminating power. Without probabilities estimated from comparable markets, consistent rules, and sufficient samples, comparisons must remain qualitative. 'Stronger' or 'average' cannot arbitrarily become percentages.

Remain neutral when evidence is close, boundaries unclear, swings not reasonably comparable, or key data missing. This is not absence of analysis; available information is insufficient to bear directional risk. Forcing a side makes ordinary fluctuation look like confirmation. The hypotheses become meaningfully separated when price trades persistently outside one boundary or resumes the breakout direction after a retest.

Test an explanation through four questions: which observed facts support it, which conflict with it, what information remains missing, and what next behavior would make it clearly preferable to the alternative? An explanation that only accounts for the past without observable expectations for the next move is merely a plausible story. Structural assessment is not a trade instruction. Unfilled plans, existing positions, and protective orders remain governed by their respective conditions.

Even if accumulation is temporarily primary, immediate exposure may be unsuitable. In the middle, price may be far from the lower-boundary invalidation reference yet close to overhead resistance, leaving insufficient potential return for the risk. A bullish assessment without a trade is consistent. Conversely, an execution-timeframe opportunity means only that one local move meets trade conditions. A local trade itself adds no evidence to the larger-timeframe structural assessment.

The primary hypothesis organizes the next observations; it does not authorize interpreting every new move favorably. A bullish view must still monitor persistent lower trading outside the lower boundary and shortening rallies. A bearish view must still monitor recovered breakdowns and upper-boundary breaks. If new action invalidates only a candidate Spring or UT, stop using that event to support direction while retaining its historical behavior. Downgrade the entire assessment or adopt the alternative only when price damages the primary hypothesis's main boundary and subsequent-swing expectations.

Neutrality also needs ending conditions. Persistent outside trading, a successful retest, and renewed progress on one side can separate the hypotheses. Continued rapid recoveries on both sides preserve neutrality. If boundaries are too wide, volume sources incomparable, or structural levels confused, even a tentative directional preference may leave no trade until risk boundaries and objective room become clear.

23.2 Two Interpretations of a Range After a Decline

Imagine a decline stopping and forming upper and lower boundaries. The first rally travels from the lower boundary to the middle; the second downward probe briefly crosses the old low and recovers; the following rally still misses the upper boundary. For accumulation, the stopped decline and lower-boundary recovery support supply encountering buying support, and the old-low breach may be a candidate Spring. The rally missing the upper boundary is contrary evidence because demand has not demonstrated a full strong result.

For redistribution, the preceding decline provides continuation context. A rally missing the upper boundary, forming a lower high, and recovering little shows weak demand. But a rapid return after a lower-boundary break weighs against redistribution because sellers could not sustain outside trading. If the recovery has very heavy activity and markedly better closes, redistribution must also explain why strong downward effort produced so little corresponding decline. If the break left only a brief wick, that contradiction is weaker.

Both explanations remain incomplete. Accumulation lacks SOS and LPS clearing supply upward and holding during a pullback. Redistribution lacks effective SOW, LPSY, and persistent outside trading after breakdown. Focus next on two outcomes: whether a rally clears a key overhead swing and a pullback stays above it, and whether a fresh downward probe sustains trading below the lower boundary and leaves the rally back resisted. Neither has occurred, so they are conditions for judgment, not evidence already credited to either side.

A subsequent upper-boundary break remaining above adds support to accumulation. A later retest holding the former resistance area adds more, weakening redistribution. A rally resisted lower, followed by a lower-boundary break remaining outside, fits redistribution better. Immediate returns after both upward and downward breaks show neither side retaining outside prices. Stay neutral rather than reversing direction with every brief breach.

If the next move only reaches the upper boundary without breaking it, it still adds information: the earlier rally's failure to reach the opposite side has improved, but accumulation's upward departure remains incomplete. Accumulation gains limited support; redistribution faces contradiction without invalidation. This intermediate result is neither confirmation nor rejection. Recognizing limited improvement avoids exaggerating an unfinished swing to preserve a binary conclusion.

23.3 Two Interpretations of a Range After an Advance

After an advance, a range forms. Price first crosses the upper boundary and returns, then falls to the middle, and later makes a low-volume recovery near the lower boundary. Distribution reads the upper action as candidate UT or UTAD and the decline after reentry as potential weakness. But a lower-boundary break returning inside contradicts distribution. Before effective SOW appears, if the predefined window or one complete downward opportunity is unfinished, this remains missing information. Only after the deadline can expected weakness be called absent. Reaccumulation can view the upper breach as an unfinished breakout attempt and the lower recovery as a supply test. Failure to hold the upper breakout and the subsequent middle-reaching decline still weigh against reaccumulation.

Here, an 'upward probe' cannot count simultaneously as four proofs—resistance break, liquidity sweep, UTAD, and demand exhaustion—because all may describe one upper-boundary breach and return. Observe additional follow-through: whether decline after reentry becomes faster and deeper, whether activity matches net progress, whether the rally recovers after the lower test, and whether the next pullback holds higher.

Under distribution, price should reach and cross the lower boundary, with the next rally resisted at old support or lower supply. Under reaccumulation, lower-boundary support should lead to an upward break and sustained trading outside the upper boundary. The prior advance makes reaccumulation consistent with continuation, without excluding distribution. Conversely, one striking upper wick cannot negate continually improving buying support at the lower boundary.

With reliable volume, compare effort and result in the Upthrust and lower recovery. Heavy effort upward without retention supports supply. Heavy effort downward without retention supports demand. Both may coexist, showing strong two-way absorption inside the range. Activity itself then selects no direction. Wait for acceptance beyond one boundary and failed recovery by the opposing side.

The assessments may take turns leading. After candidate UT, a decline reaching the lower boundary supports distribution; immediate recovery of the breakdown then weakens that support. A subsequent upper-boundary break and successful retest progressively complete reaccumulation. If the lower recovery produces only a brief rally before another breakdown and outside trading, the original distribution assessment may still hold. At each step, examine which side ultimately retains price.

A short after upper-boundary failure first has its objective constrained by demand below. A long after lower-boundary recovery first faces supply above. Post-breakout continuation provides new grounds for holding, potentially converting a range trade into a trend position. If a retest follows, reexamine whether former resistance or support holds. A leading hypothesis does not guarantee a suitable entry distance remains.

23.4 Supporting Behavior and Contrary Evidence

Observed action meeting an assessment's expectations gives it more support. Action damaging expectations for key boundaries and subsequent swings, or clearly fitting an alternative better, provides contrary evidence. Unoccurred events support neither side. As price merely approaches the lower boundary, observe whether a Spring develops; do not credit an unobserved Spring to accumulation. Before price tests the upper boundary, 'no SOS' usually means incomplete information.

Only after a predefined observation window ends can absence become evidence that expected behavior failed to occur. A window may be behavioral—for example, one complete rally opportunity from the lower boundary failing to reach the opposite side of comparable earlier round trips—or a time period set beforehand. Do not extend it after a challenge until the desired result happens by chance.

'Strong buyers' or 'weak sellers' does not explain the basis. If the current rise is larger and faster than the previous one, clears a swing high, and holds above it on a pullback, upward action has become stronger. If activity increases during a decline but a new low extends only slightly and a rally then reclaims it, supply has spent more effort without proportionate downward result. Specify the two swings compared, their range location, and the final outcome.

Contrary evidence varies in weight. One wick or unfinished bar may reduce only local-event confidence. Where stable round trips already exist, a swing missing its expected opposite side shows relative weakening in that direction. Persistent trading outside a key boundary, a successful retest, and renewed progress directly damage the competing explanation's core expectation. Regardless of structural evidence strength, individual orders follow their own stops.

An unfinished candle is observable too. If rules permit intraday decisions, act on the high, low, current price, and cumulative activity formed so far while recognizing these can still change. Close-confirmed rules wait for the close. Choose the method beforehand; do not switch afterward to the one favoring the original view, or bring an unknown final close back into the historical decision during review.

Three Ways to Record Evidence

Missing behavior becomes evidence only if it still has not appeared after the predefined observation period ends.

Figure 23-2. Three ways to record evidence (based on Section 23.4 of this course)

23.5 Persuasive Evidence Versus Repeated Labels

Evidence is more persuasive when it rules out an alternative more effectively; visual prominence does not equal evidential strength. One enormous-volume bar may reflect active supply, demand absorption, or a news shock. Without estimating its probability under each hypothesis, prominence alone cannot establish discrimination. Persistent trading outside after a breakout, a successful retest, and continued progress more directly conflict with a failed-breakout explanation and therefore usually separate the hypotheses better. Read volume, spread, close, and subsequent response together.

One move cannot be counted repeatedly under different names. A high-volume decline, candidate SOW, local-low break, and 'strong sellers' may all come from one swing. They still support one downward event, whose implications for each hypothesis should be assessed separately. A boundary break, the next rally back, and subsequent continuation occur at different times and add later information, but chronological succession does not imply statistical independence. If the latter two are developments of the same breakout, they retain a common source.

Apparently different indicators may be highly correlated. A wide-spread down candle with a large range, low close, and low break contains three descriptions of one event, not three independent proofs. Their separately estimated likelihood ratios especially cannot simply be multiplied.

Dependent evidence can still update judgments sequentially, but later evidence needs a conditional likelihood already including earlier evidence. For example, after E₁, evaluate E₂ using P(E₂|H,E₁). Only if E₁ and E₂ are conditionally independent given the hypothesis may this simplify to P(E₂|H). Less-overlapping information may come from prior trend quality, changes across several same-direction range swings, retention outside after a breakout, continued directional progress after a test, and relative market performance. Even then, an unvalidated scoring table cannot mechanically vote. Explain the supply-demand changes reflected and which interpretation they contradict.

Repetition of the same phenomenon need not create multiple strong proofs. Three upper-boundary wicks in the same congested area without any downward follow-through show only repeated testing of that supply zone. Repeated tests may also absorb supply. Compare how far price falls after each departure, how much effort the return requires, and which side ultimately holds after a breakout and continues progressing.

Evidence correlation also changes with scale. One daily breakdown may comprise several hourly SOWs and LPSYs, but together they form one daily decline. If the task is daily structural analysis, hourly events cannot all be accumulated as separate votes. The hourly chart can explain the daily decline's formation and provide execution locations, while the daily hypothesis still updates through comparable daily swings and whether price breaks and holds beyond boundaries.

23.6 What Observations Distinguish the Two Explanations?

Specify location, direction, analytical timeframe, and expected result for the next observations. 'Watch what happens next' has no diagnostic value. 'If a trading-timeframe pullback holds the upper-boundary zone and the following rise clears the pre-retest high, reaccumulation becomes more likely' states what to await. Also predetermine whether boundaries are lines or zones and whether confirmation uses closes or intraday structure.

Choose conditions that produce different outcomes under the two hypotheses. Near the upper boundary, distribution expects reentry after crossing and subsequent substantial downward progress. Reaccumulation expects persistent outside trading, a pullback holding the old upper boundary, and further rising. Near the lower boundary, accumulation expects recovery of the probe and progress inward; redistribution expects continued outside trading after breakdown and a rally resisted near old support. A condition that preserves the original hypothesis whatever happens has no testing value.

Do not mix observed action with unverified conditions. A smaller additional decline after breaking the old low than at the previous new low shows only weaker supply progress. Accumulation gains further support only when a subsequent rally clears midrange supply and holds during a pullback. This does not establish an 'imminent upward breakout,' and placing an order does not turn hope into fact. If price develops an unanticipated third pattern, return to neutrality and reassess observed supply-demand changes rather than forcing the old binary choice.

Connect these conditions with specific trade handling. Some moves merely make one direction worth closer attention; some permit searching for an entry trigger; others cancel unfilled opposing plans. The distinction depends on predefined rules. By the time structure is clearer, price may be far from invalidation and no longer offer suitable entry. A local trigger also does not eliminate the alternative structural assessment. Structural analysis selects the principal focus; entry conditions decide whether current price and risk justify participation.

23.7 Updating, Deferring, Switching, and Not Trading

When new action fits the primary hypothesis, retain it and use recent highs and lows as new references. With local contradiction, withdraw that event's support for direction or new entries while preserving its historical behavior and reclassifying if necessary. Only opposing action damaging key conditions requires adopting the alternative or becoming neutral. There is no need to reverse direction with every bar, and local-event failure need not immediately overturn the whole range assessment.

Update unfilled plans with their premises. Cancel a Spring-test buy plan when price breaks the lower boundary and persists outside. Cancel an LPSY-resistance sell plan when price establishes itself above its invalidation area. After an order fills, protection follows preestablished rules. An unresolved larger direction does not justify delaying a stop.

Not trading may reflect an absent trigger, boundaries too wide for suitable risk and objectives, hypotheses lacking discriminating conditions, or price far from the planned area. It is a decision combining evidence and execution constraints, not an end to analysis. New boundaries or tests can support a new version, while retaining the old version's information cutoff and cancellation reason.

In review, check whether supporting and contrary evidence actually existed then, whether unobserved moves were counted, whether one event was duplicated, whether judgment conditions were predefined, and whether appropriate action followed contradiction. Phases A–E and labels such as Spring and UTAD often become clearer as the structure unfolds. Retrospective labels organize history but cannot be treated as contemporaneously known information. Final direction is only one outcome and cannot replace scrutiny of the analytical process.

This review reconstructs the information and choices available then, allowing one to determine whether changing conclusions came from subsequent price action or from retrospectively changing comparison objects or confirmation methods.