Supply and Demand Zones: Drawing Rules and Limits

State how the region is selected

Technical analysis commonly discusses support and resistance around past price behavior. Fidelity describes this chart-based framework. Supply/demand-zone approaches often extend a line into a region based on a consolidation before a move. There is no single universal boundary formula shared by every author or indicator.

A practical definition can specify a maximum number of base candles, their permitted range, a minimum departure distance and the time available for that departure. Such a rule is an illustrative research convention, not proof that the region contains unfilled orders. It should be recorded before later price visits are observed.

An original demand-zone construction

Suppose three completed base candles have lows between 49 and 49.50 and highs between 50.50 and 51. A subsequent two-bar move closes above 54. An illustrative rule marks the complete base range, 49 to 51, after the qualifying departure has occurred.

A body-only variant might produce a narrower region, perhaps 49.75 to 50.50. If price later reaches 50.80, it visits the full-range zone but not that narrower version. The difference is a measurement choice, not evidence that one observer understands hidden supply better.

Assume the declared failure rule is a completed close below 49 within the next twenty observations. A later close of 48.50 records a failure, even if price subsequently recovers to 55. A study must retain the original horizon and outcome rule rather than extending the holding period until a favorable story appears.

Support and resistance: choose a line or a zone before the test

A horizontal support line records one selected price below the market; a support zone records an interval. Resistance uses the corresponding idea above the market. Fidelity explains support, resistance and possible role changes. CME describes previous highs and lows as one way to select levels and notes that reactions need not occur at an exact price. Those descriptions provide context, not a universal drawing algorithm or evidence for the thresholds below.

To compare the drawings fairly, use the same available candles and confirmation time. Otherwise a narrow line selected after a successful reaction is being compared with a zone drawn before that reaction. A tolerance is part of a line’s test definition: allowing 0.25 price units around a line differs from requiring an exact trade at the line.

A repeatable selection rule on the same OHLC sample

The following nine one-hour bars are entirely synthetic, expressed in arbitrary price units. They are not exchange data, dollars, pips or a backtest. Bar numbers show sequence; they do not represent a real calendar or venue session. The first three bars make the 49–51 base used above explicit.

Synthetic OHLC prices for a declared base, confirmation and subsequent failure.

Bar	Open	High	Low	Close
1	50.00	51.00	49.00	50.50
2	50.50	50.75	49.50	49.75
3	49.75	50.50	49.25	50.00
4	50.00	53.00	49.75	52.50
5	52.50	54.50	52.00	54.25
6	54.25	54.50	50.80	51.50
7	51.50	52.00	49.20	50.00
8	50.00	50.50	48.25	48.50
9	49.00	55.25	48.50	55.00

1. At each completed close, once five bars exist, examine only those latest five bars. The first three form the candidate base; the last two form the departure window.
2. Require the base’s maximum high minus minimum low to be no more than 2.00 units. Require the fifth bar’s close to exceed the base high by more than 3.00 units. Use strict greater-than for departure; equality does not qualify.
3. Select the first qualifying five-bar window. Freeze its base low as the support line and its full low-to-high range as the demand zone. Do not substitute a later, more attractive base while this candidate is active.
4. Activate both drawings only after confirmation. Examine the next twenty completed bars, ending earlier if the relevant drawing fails. A missing bar leaves coverage incomplete; it is not a zero-price observation or a successful test.

At bar 5, the base width is 51 − 49 = 2.00 and the confirming close is 54.25, above 51 + 3 = 54.00. The frozen line is 49.00 and the zone is 49.00–51.00. Their first eligible observation is bar 6. The low of 49 was already visible at bar 3, but this comparison deliberately waits for bar 5. A standalone support-line method activated earlier would be a different rule.

Define overlap and failure separately

For the line, use an inclusive tolerance band of 48.75–49.25. Record a possible touch when a completed bar’s high–low interval overlaps that band: low ≤ 49.25 and high ≥ 48.75. For the zone, require low ≤ 51.00 and high ≥ 49.00. These are OHLC-range classifications; they do not establish that every intermediate price traded or that an order could fill there.

Declare line failure at a completed close strictly below 48.75 and zone failure at a completed close strictly below 49.00. A close exactly on either threshold does not fail its rule. Preserve overlap and failure as separate fields because one bar can satisfy both. A wick below a boundary does not itself satisfy this close-based failure convention.

Classifications under the frozen rules, without any trade or return calculation.

Observation	Line band	Full zone	Failure at close
Bar 6: low 50.80	No overlap	Overlap	Neither
Bar 7: low 49.20	Overlap	Overlap	Neither
Bar 8: close 48.50	Overlap	Overlap	Both
Bar 9: close 55.00	Post-failure	Post-failure	Earlier failures remain

Count consecutive overlapping bars as one visit episode for each drawing until a non-overlapping bar or its failure ends that episode. Bars 6–8 are one zone episode; bars 7–8 are one line episode. They are not independent successful retests. Bar 9 remains in the record for context, but neither failed drawing restarts automatically when price recovers.

A resistance comparison can mirror the convention: use the base high for the line, require a two-bar departure whose final close is more than 3.00 below the base low, and test subsequent closes above the resistance boundaries. Define line failure above its upper tolerance edge and zone failure above its high. This is a specification, not a claim that bullish and bearish outcomes have equal probabilities.

Keep the rules with the observation record

Download the blank observation worksheet for a manual record of one candidate across sequential bars. Its ten blank rows are marked not_recorded; all prices, timestamps and classifications are blank. It contains no formulas, automatic signals or prefilled performance results. Copy the candidate’s frozen rules consistently across its rows, then record source, timeframe, timezone, OHLC, confirmation time, eligibility, overlaps and failures. Use separate candidate IDs when changing a rule.

For real data, retain the original timestamp and price basis, such as bid, ask, last trade or a documented chart feed. Do not join different feeds or timeframes silently. Check missing bars and gaps before labeling touches. OHLC cannot recover the sequence of every intrabar trade, and a high–low range spanning a level does not prove continuous trading through it. Leave an observation unavailable when its prerequisite is missing.

This exercise compares measurement conventions. Evaluating a trading strategy would additionally require predeclared entries, exits, sizing, spreads, commissions, financing and executable prices on an unseen sample. In this sample, the different overlap counts follow from the chosen interval boundaries and prices. The zone does not contain the entire line-tolerance band, so its greater width does not guarantee more overlaps in every sample. Neither count establishes higher profitability.

Separate a past region from present liquidity

A fast departure records what prices did during a historical interval. It does not reveal whether previously resting orders were filled, canceled, replaced or never present in the assumed quantities. A drawn rectangle is therefore different from an exchange order book.

Even a live order book represents only its available venue coverage and displayed orders at a particular time. Liquidity changes. A historical zone drawn on a broker's FX chart cannot be treated as a complete inventory of global currency demand or supply.

Count visits and overlaps consistently

Specify whether repeated entries into a zone count as one test or several. Decide whether a new nearby region replaces an old one, merges with it or stays separate. Also define how gaps that jump across a region are recorded, because they may provide no executable price inside it.

These details matter when comparing a claimed retest rate. Ten overlapping rectangles surrounding one price event are not ten independent observations. Keep the original detection time and avoid drawing the base as a known signal before the later departure that qualified it.

Compare specific order-block conventions, inspect what an order book measures, and log candidate outcomes with the journal workflow. The useful result is a clearly defined region and testable observation process, with failures included, rather than a claim that price owes the zone a profitable reaction.