Elliott Wave theory: impulse rules, diagonals and invalid counts

Elliott Wave theory proposes nested price structures, with smaller waves forming larger ones. A number on a chart is an interpretation of available prices, not a message from the market or proof that future prices follow a predictable psychological rhythm. Use the method’s definitions to state what would reject a particular labeling, then evaluate any trading rule separately.

A wave label is not a candle or timeframe

The basic directional pattern is labeled 1–2–3–4–5, with a simple correction A–B–C. Waves need not have equal duration or contain equal numbers of candles. Degree identifies a place in a larger structure, not a prescribed interval. A motive pattern can move downward as well as upward; five-wave does not mean “rising,” and three-wave does not mean “falling.”

EWI’s motive category includes impulses and diagonals. A standard impulse subdivides 5–3–5–3–5: parts 2 and 4 are corrective, and subwave 3 must itself be an impulse. Connecting five external line segments does not demonstrate those internal subdivisions. Check the proposed degree and context instead of label every visible bend as a complete wave.

The anatomy of five impulse waves and three corrective waves

A five-part motive sequence followed by a simple three-part correction is a basic schematic cycle, not a requirement that every finished correction contains only three visible moves. In an upward version, 1, 3 and 5 advance while 2 and 4 retrace; reverse the direction for a downward version. Corrective patterns and combinations require their own subdivisions, so a visible A–B–C can belong to a larger correction.

In a possible impulse, waves 1, 3 and 5 run with the larger direction while 2 and 4 retrace it. Descriptions of crowd psychology and common Fibonacci proportions are interpretive guidelines. Wave 3 need not be the longest; the rule is that it must not be the shortest of waves 1, 3 and 5. Price labels and proportions do not establish participants' motives or predict the next wave.

The three iron rules of Elliott

For the strict standard-impulse variant examined here, check its price rules and required subwaves. A breach rejects the specified labeling; it does not automatically prove another pattern or dictate a trade. EWI distinguishes rules from guidelines and discusses rare short-term overlap exceptions in futures; those exceptions are not assumed in our strict teaching model.

• Wave 2 retraces less than all of wave 1. Reaching or passing its origin fails this stated condition. An invalid label alone does not establish a complete larger correction or an executed stop.
• Wave 3 passes the end of wave 1 and is not the shortest of 1, 3 and 5. EWI discusses percentage movement; compare like units and keep the final wave-5 boundary in view.
• The strict impulse excludes wave-4 overlap into wave-1 territory. Inspect the full path, not only its final price. A diagonal is a separate construction, with its own position and subdivisions, not a label that automatically repairs an overlap.

The extended wave and Fibonacci proportions

EWI describes an extension as an elongated impulse with exaggerated subdivisions. It is not defined solely by touching 161.8% or 261.8% of a prior segment. Typical proportions and expectations about which subwave extends are guidelines, not universal FX probabilities. Reaching one ratio does not require the next ratio to be reached.

Fibonacci retracement measures a chosen correction; extension tools measure projections. Missing an exact 61.8% retracement does not alone violate an impulse rule. Harmonic pattern trading is a related use of specified geometric ratios, but neither a ratio nor agreement between drawings establishes a profitable forecast.

Diagonals and corrections need separate definitions

An ending diagonal is considered in position 5 or C; a leading diagonal can occur in position 1 of an impulse or A of a zigzag. Overlap is characteristic of diagonals, but not every wedge is one. Position, price relations and internal structure matter. EWI acknowledges ambiguity in a strict leading-diagonal subdivision definition; do not impose one simplified subdivision as uncontested.

Practical takeaway: Distinguish the rules of a wave count from interpretive proportions and crowd-psychology descriptions.

EWI lists zigzags (5–3–5), flats (3–3–5), triangles with five corrective parts A–B–C–D–E and combinations. Five corrective parts do not turn a triangle into an impulse: the internal character and nesting differ. These sources document a method promoted by their authors; their selected examples and marketing statements do not replace an independent test of a complete trading rule.

How to count waves in practice — a hypothetical example

For hypothetical EUR/USD with one pip = 0.0001, propose endpoints 1.1000 → 1.1100 → 1.1040 → 1.1160 → 1.1110 → 1.1240. Segments 1–5 are 100, 60, 120, 50 and 130 pips. The three directional lengths are 100/120/130: wave 3 is between the other two, not the longest. Their percentage moves are about 0.9091%, 1.0870% and 1.1701%, giving the same ordering. Wave 2 retraces 60%; wave 4 retraces 50/120 ≈ 41.67% yet remains above wave 1’s end. The endpoints are proposed boundaries, without full history or subwaves, so these checks do not confirm a complete impulse.

Independently change one condition at a time. A wave-2 end of 1.0990 retraces 110 pips against wave 1’s 100 and fails. A wave-4 low of 1.1090 overlaps below 1.1100 and rejects this strict variant without proving a diagonal. A wave-3 end of 1.1120 gives directional lengths 100/80/130 pips: 3 becomes shortest. Its percentage rise is about 0.7246%, also below those of 1 and 5. Keep the earlier version and recorded reasoning when labels change. The broader workshop is in the ForexMechanics technical analysis section.

An unfinished wave can change the length check

In a separate hypothetical endpoint model 100 → 110 → 106 → 114 → 111 → 117, directional lengths are 10, 8 and 6. Wave 3 is shorter than 1 but longer than 5, so this selected condition is not breached. If the proposed fifth continues to 120 with earlier anchors unchanged, lengths become 10, 8 and 9: 3 becomes shortest and that label fails. Percentage moves agree: 10%, about 7.5472%, and either 5.4054% or 8.1081% for the fifth. Passing the check before 5 ends does not establish its final validity; the endpoints still do not verify subwaves.

• Version the labeling. Record observation time, instrument, feed, interval, degree and proposed invalidation levels.
• Preserve availability. Keep an image with the last then-available bar; a pivot may need later bars before recognition.
• Preserve old outcomes. A later relabeling does not erase a prior forecast made with the old record.
• Keep claims separate. Rule compatibility, alternative interpretations and measured net trading results are different questions.

A level that invalidates an interpretation is not automatically an order fill or a maximum-loss promise. A strategy needs separate entry, exit, size, costs and execution rules, with evaluation on unused data. Support and resistance analysis and trend-following systems describe related approaches; their names do not turn a wave label into an independently confirmed profitable method.

What to do tomorrow if you want to learn wave counting

1. Recalculate the endpoint model. Check directional lengths and percentage changes without treating the five external segments as verified subwaves.
2. Separate ratios from rules. A 41.67% wave-4 correction exceeds 38.2% but leaves the stated no-overlap condition intact; a Fibonacci ratio is not a promised target.
3. Check the 100% boundary. Change only wave 2’s end to 1.1000. Its 100-pip retracement equals all of wave 1 and fails the stated less-than-100% condition, without proving another pattern.
4. Keep dated alternatives. Record each version, data availability and rejection condition. Several compatible labels do not acquire equal probabilities merely by being counted; a dozen selected examples does not establish trading efficacy.