Range trading tests movements inside previously defined boundaries. A rectangle alone is not a strategy: specify how the bounds become known, decision time, execution and retirement conditions. The following hypothetical models show reproducible rules, without claiming a verified profitable system.

What a systematic approach to consolidation actually means

Define the actual price rule instead of assuming that a sideways description guarantees reversion. The range-trading introduction covers the basic concept; this framework focuses on chronology and execution. The mean-reversion guide explains why a hypothesised return is different from demonstrated stationarity or an edge.

Later boundaries can change an earlier qualification

The same price can look closer to or farther from the lower edge. Reconstruct a decision with the rectangle known at that time.

Scroll sideways to see the full diagram

Price 102 in two rangesIn range 100–110, price 102 is at 20% of width. A later range 95–115 moves its description to 35%. The new scale was unavailable at the earlier evaluation.OLD CLOSE, TWO DIFFERENT RECTANGLESBounds known at t₀102100110Bounds set later10295115Then: (102 − 100)/10 = 20%.After change: (102 − 95)/20 = 35%.Later bounds change the old description.

Our model: at t₀, known L = 100 and U = 110; evaluate Close = 102. Boundaries 95–115 appear only later. The rule admits the lower quarter, x ≤ 25%, where x = (Close − L)/(U − L). We model neither order execution nor a future result. The document TradingView — Repainting confirms that repainting and future-data use can make the historical display differ from information available during observation.

Fictional Close 102 is unchanged. Only the boundaries used to describe it change.

Using data available at t₀, price passes the location check: 20% ≤ 25%. Recalculating the old point with later bounds gives 35%, apparently removing the earlier condition. That changes the description after the fact, not the earlier information. Store the determination time and both bounds at each evaluation. A new rectangle can serve the next decision, but must not replace the previous record.

How to tell a genuine range from a pause in a trend

For our own fictitious 23 September 2026 case, freeze bounds 1.0800–1.0900 before 09:00 UTC using the known extrema of 24 closed H1 Bid bars; retire new-entry qualification at 12:00. These are synthetic inputs, not that date’s market quotes. Width is 100 EUR/USD pips at pip 0.0001. Recalculating old signals with later extrema introduces future information.

Inputs to save before range evaluation
Bounds and sampleSource, price side, interval, fixed window and availability time
TouchesIf used, define distance and spacing; successive bars need not be independent tests
WidthNo universal 30/100/200-pip minimum follows from a range label
FiltersSpecify optional indicator conditions without assuming their value
End stateRecord expiry or retirement separately from live SL/TP execution

The support/resistance guide helps state the bounds. Fidelity describes ADX as trend strength; direction comes from DI comparisons. A chosen 20/25 threshold does not establish calmness, future containment or buyer/seller edge. Bollinger’s official rules distinguish a band tag from a signal, allow walking a band and treat 20 periods/two deviations as starting defaults. For more context, see technical analysis.

Separate observation, qualification and fill

Our lower working zone is 1.0800–1.0820, 20% of the frozen width. Require a closed M15 Bid bar with Low≤1.0800 and inclusive Close 1.0805–1.0820. Parameters are chosen for this experiment, not optimised or validated. Bid bar 09:00–09:15 has O 1.0810/H 1.0818/L 1.0798/C 1.0815 and qualifies only after final data arrive.

  1. At the assumed 09:15:01 quote Bid 1.0815/Ask 1.0817, spread is 2 pips. The bar alone did not supply Ask, and its later Close cannot justify a 09:00 entry.
  2. Allow at most one attempt only if current Ask≤1.0820, spread ≤2 pips, trading is available and no position exists. That pre-order gate does not guarantee its fill; the calculation separately assumes Ask 1.0817 execution.
  3. An RSI or stochastic filter is an additional rule to test, not automatic evidence of a higher bounce probability.
  4. Fix entry, exit, size and costs. Waiting for a closed bar defines information time; it does not eliminate most false signals.

Stop loss and take profit — the arithmetic of the range

For 10,000 EUR on a USD model account, use assumed Ask entry 1.0817, SL 1.0785 and TP 1.0895 checked against Bid under the Forex rule. Pip value is USD 1. With hypothetical USD 2 round-trip commission and zero financing, independent exact target/stop Bid fills give gross USD 78/−32 and net USD 76/−34. Worse stop Bid 1.0780 gives gross −USD 37/net −USD 39. Net 76/34≈2.24 is not a success probability or maximum loss. Fill prices include spread once. No fixed 10–20-pip buffer, 90% capture or 1% input establishes safety.

The most important rule: do not fight a breakout

Our model stops qualifying new entries at the first closed M15 below 1.0795 or above 1.0905, or 12:00 UTC. This is a declared retirement condition, not proof of a lasting breakout. An already open trade follows earlier SL/TP execution; otherwise submit its timed closing order at 12:00, with its actual fill then unknown. Do not suspend a live stop until a candle closes.

Bollinger’s official band rules do not equate a tag with an entry signal. In a separate retirement variant, no position is yet open: Close 09:15=1.0794 disables our model. The next bar Low 1.0798/Close 1.0810 at 09:30 meets the price condition, but the model remains disabled. Restarting on re-entry needs another rule, not a decision added after a favourable bounce.

A step-by-step example — a hypothetical scenario

Keep a separate hypothetical H4 illustration with bounds 1.0800–1.0900, four support touches, five resistance touches and ADX 18 as assumed inputs. They do not validate a tradable range. A 100-pip versus 200-pip width gate would merely select different samples; neither threshold is an established minimum.

For that independent H4 case, assume actual Ask 1.0815, planned Bid SL 1.0785 and target Bid 1.0895, after a chosen closed-bar pattern/RSI condition. Distances are 30 and 80 pips, gross 80/30≈2.67. For 10,000 EUR and USD 2 commission, exact target/stop net outcomes become USD 78/−32, ratio 78/32≈2.44. A confirmed pattern is an input, not proof of a bounce or guaranteed execution.

A bar outside the bounds can end a declared setup even before entry; that is not automatically a small realised loss. Preserve every range’s states and actual order events. If one bar spans SL and TP, OHLC alone does not order them. Use finer data or a declared conservative convention and evaluate a sample not used to select parameters.

What to do tomorrow

  1. Record the source, interval, price side and clock. Save the fixed bounds and any indicator settings at each decision.
  2. Define touch spacing, width and expiry rules explicitly if used. Keep periods that never give an entry and those that retire; do not remove them after their outcome.
  3. Write the complete hypothesis, entry information time, fill assumptions, quantity, costs and exit states. Indicator agreement is not automatically independent confirmation.
  4. Separate new-entry retirement from existing-position management. Define any restart before testing and compare it as a new version.