A head and shoulders top has three successive peaks after an uptrend: the head is above the shoulders. The neckline connects the intervening reaction lows. A rule can wait for a close below it, then enter, await a specified retest or decline the trade. Choose the rule before seeing subsequent prices; a confirmed pattern condition is not a guaranteed reversal.

What the head and shoulders pattern is

A conventional top consists of a left shoulder, higher head and lower right shoulder, with a neckline through the two intervening lows. Similar shoulder heights help describe the geometry but do not force a future decline. Record the swing and neckline definitions. A close below the line can confirm a chosen entry condition; neither that close nor a drawn shape proves profitability.

The inverse variant has three troughs after a decline, with the head below the shoulders and a neckline through the intervening highs. A close above it can satisfy a chosen bullish condition. Changing direction also changes execution sides: a long opens at Ask and closes at Bid, while a short opens at Bid and closes at Ask.

How to draw the neckline and project the target

Draw the neckline through the two selected reaction lows, retaining their candle indices and the time at which each became known. For a sloping line, calculate its value at the assessed candle and project height using a stated convention. The neckline is related to support and resistance, but its position is a geometric reference rather than evidence of actual defending orders.

The target is found with the measured-move method. You measure the vertical height of the pattern — from the top of the head down to the neckline — and then project that same distance downward from the point where price broke the neckline. The level you get is an approximate geometric target for the decline. It is not a minimum guaranteed move or, without a specified study, a statistical forecast: sometimes the market falls short of it, and sometimes it overshoots by a wide margin. In the inverse head and shoulders you do exactly the same thing, only you project the distance upward.

Hypothetical EUR/USD model — horizontal neckline, USD account
Prior contextAssume the peaks and horizontal neckline were already identified
Head / neckline1.2100 / 1.2000 on the Bid chart
Right shoulder1.2040
Signal closeBid 1.1982, followed by assumed sale at Bid 1.1980
Retest alternativeSale at Bid 1.1995 after the stated candle conditions
Geometric target100 pips below the horizontal neckline: Bid 1.1900

Why recalculate a sloping neckline for each candle?

This separate model replaces a horizontal neckline with a rising line. Its anchors are known beforehand:(t0;1.1990) and(t10;1.2000). We do not change the first example’s rules.

Scroll sideways to see the full diagram

A rising line changes the comparisonThe fictional neckline is anchored at t0=1.1990 and t10=1.2000. Closes at t14 and t16 both equal1.2005. At t14 price is1 pip above the line; at t16 it is1 pip below. Close has not fallen.SAME CLOSE, CHANGING NECKLINENeckline: +1 pip/periodClose = 1.200510141620t14: line1.2004; Close1 pip above.t16: line1.2006; Close1 pip below.Crossing need not involve a lower Close.

Hypothetical EUR/USD, pip0.0001, linear axes and equal periods. Historical anchors t0 and t10 are assumed recognised before evaluating t14; the later segment extends the line. IG’s head-and-shoulders guide treats a close below the neckline as a separate observation condition.

The line is a geometric reference, not a future exchange rate.

The neckline equals1.2000+(t−10)×0.0001. With Close1.2005, compare against1.2004 at t14 and1.2006 at t16. The above/below state changes despite equal closes. Do not substitute an old anchor price or attribute every crossing to falling price. Record line value and Close deviation at each index. This proves neither reversal nor an execution price.

The role of volume and the neckline retest

If volume is part of the rule, specify which series you use and measure the proposed condition rather than assuming it validates the shape. MetaTrader’s iVolume returns chart tick volume. A change in that count is not the monetary size of executed trades and does not identify buyers, sellers or a global spot-FX total. Results from another feed or volume definition need separate assessment.

A retest may occur, but waiting for it can also leave no trade. In this hypothetical alternative, wait at most three later candles for Bid High within 1.1998–1.2002 and Bid Close no higher than 1.1996. High 1.2001 and Close 1.1996 satisfy that rule; assume the later sale executes at Bid 1.1995. A sell limit placed earlier at the neckline is a different rule because it could fill before the confirming close.

How to run the setup step by step

Step 1 — confirm the context and that the pattern is complete

Fix the instrument, interval, price series and swing rules. Evaluate only peaks and closes already known at the decision time. A higher interval is a different dataset, not proof that its signals contain less randomness. Keep rejected candidates and record why the entry condition was not met.

Step 2 — wait for the break and enter

For the chosen model, a close below the neckline precedes the order decision; it does not guarantee a fill at that close. Compare short entries Bid 1.1980 after the break and Bid 1.1995 after the defined retest. Both sell EUR 10,000, giving USD 1 per 0.0001 pip. Assume total round-trip commission USD 2, no financing and no other fees. Alternative closing fills Ask 1.1902 or Ask 1.2044 give net outcomes +USD 76/−USD 66 for the first entry and +USD 91/−USD 51 for the second. Use 10,000 × (entry Bid − closing Ask) − 2; spread is already in those prices.

The net reward-to-loss ratios are about 1.15 and 1.78 for these assigned alternatives. No outcome probabilities were supplied, so they are not expected returns. Assume neither target nor stop was reached before the possible retest entry. In a separate path without retest, the first policy could gain USD 76 while the second makes no trade and has zero trading P/L from this setup.

Step 3 — set the stop and the target

State the stop and expiry in the selected rule. The model’s alternative Ask 1.2044 stop fill lies above the right shoulder, but an ordinary stop does not cap loss there. A worse Ask 1.2047 fill would give −USD 69 and −USD 54 at the same EUR 10,000 quantity. Do not subtract spread again or treat these as maximum losses.

A separate USD 100 price-loss budget gives theoretical quantities 100/(1.2044−1.1980)=EUR 15,625 and 100/(1.2044−1.1995)≈EUR 20,408.16. With an assumed EUR 1,000 quantity step, rounding down gives EUR 15,000 and EUR 20,000. Their gross stop losses are USD 96 and USD 98, and gross target results USD 117 and USD 186. After the separately assumed USD 2 commission, stop results are −USD 98/−USD 100 and target results +USD 115/+USD 184. The budget excluded fees and slippage; changed quantity changes exposure. This is a different comparison from equal notional, not a recommendation to enlarge the retest trade.

When the pattern fails

A break can be followed by a return above the neckline and a loss under the selected rule. Do not import an unspecified stock-pattern failure percentage into FX. Compare every qualifying case, including missed retests and rejected orders, on a stated period. A hypothetical spread filter of 4 pips would reject Bid/Ask 1.1965/1.1971 because the spread is 6 pips; that threshold is an assumed rule, not a universal safe cost limit.

Head and shoulders versus other reversal patterns

The double top and double bottom use two primary extremes rather than three; each definition still needs its own neckline and confirmation conditions. The preceding trend context helps state what is being called a reversal without proving it will happen. Compare complete rules and net results rather than an unsupported “works best” ranking. The broader technical-analysis overview puts these constructions in context.

What to do tomorrow to get comfortable with the pattern

  1. Choose a chart series and swing rule, then review observations chronologically. Record candidates before seeing later outcomes instead of searching only where a trend is already known to have ended.
  2. On every candidate you find, draw the neckline through the two reaction lows and measure the vertical height from the top of the head to that line, then project the same distance down from the break, so that you can see with your own eyes how the measured-move target actually works.
  3. Compare the tick volume on the left shoulder, the head and the right shoulder, as well as at the break — note in how many cases volume really did fade on the right shoulder and rise on the break, because only those numbers will show you how often the theory matches practice on your own market.
  4. Set a price alert just below the neckline on a current, forming pattern instead of staring at the chart for hours — when the alert fires, you can calmly judge whether the candle truly closed below the line and whether it is worth waiting for a retest.
  5. Test a frozen entry and exit rule on a separate period, including costs, unfilled orders and no-trade cases. A fixed number of demo trades does not establish a durable edge or readiness for live risk.