A wedge has two converging price boundaries sloping in the same direction. The drawing does not prove that buyers or sellers are exhausted. Check their relative slopes, when the selected points became available, and the exact breach rule before defining a separate trade.

What a wedge pattern actually is

A wedge is built from two converging trendlines but — and this is the crux — both slope in the same direction. That sets it apart from a triangle, where one line falls while the other rises (symmetrical) or one is horizontal (right-angled). Both lines run either up (a rising wedge) or down (a falling wedge) at different speeds, so the space between them keeps shrinking.

Trend context distinguishes a proposed reversal from continuation: a falling wedge within a rising trend may be a correction, while one after a decline may be a reversal hypothesis. Neither classification guarantees the next move. The technical analysis section covers these conventions; geometry alone does not establish trader intent or profitable behavior.

Relative slopes and a calculable width

In a rising wedge the lower boundary rises faster than the upper one; otherwise the interval widens. Assign fictional lines G(t)=1.1100+0.0005×t and D(t)=1.1000+0.0010×t. For a falling wedge use G(t)=1.1200−0.0010×t and D(t)=1.1100−0.0005×t. Here t counts equal intervals and one pip is 0.0001. Both widths are 0.0100−0.0005×t:100, 80, 60, 40 pips at t0, 4, 8, 12.

At t12 the rising endpoints are 1.1160/1.1120 and falling endpoints 1.1080/1.1040. Width has fallen 60%; the lines intersect at t20. That is a mathematical intersection, not a breakout deadline. Beyond it the boundaries reverse order. Lines alone are not complete OHLC or proof of an observed pattern.

In a falling wedge, both boundaries slope downward and converge: the upper resistance line falls faster than the lower support line. A break above resistance is a possible bullish signal, not a guaranteed turn. A rising wedge can form during an uptrend or as a countertrend pause within a decline; interpret either structure in its wider context.

Converging lines define a wedge. The breakout direction, entry price and invalidation must be observed and planned rather than inferred as certain.

How to tell a wedge from a triangle and a pennant

The three formations are easy to confuse, because they all rest on a narrowing range. Look at the slope of the two lines: if they run the same way, you have a wedge; if they run opposite ways, it is a triangle.

Wedge, triangle and pennant — geometry versus hypothesis
WedgeSame-direction slopes and shrinking positive width; possible reversal or continuation context.
Symmetrical triangleOpposite slopes converge; the shape alone supplies no outcome.
Right-angled triangleOne horizontal boundary and one converging sloped boundary.
PennantSmall converging structure considered after an impulse; duration needs a defined rule.
VolumeSpecify the field and filter. Falling activity or a spike is not a required geometric fact.

A fixed timeframe or number of weeks is not part of our line model. Fix a point-selection method and tolerance. If a pivot requires two later bars for confirmation, a point on bar 8 is usable after bar 10 closes, not retrospectively on bar 8. Freeze lines when available; redrawing them with later pivots changes the tested rule.

Crossing mid-bar is not closing beyond the line

The previously known lower line is D(t)=1.1000+0.001×t. At t 12.5,Bid 1.1115 is below D12.5=1.1125. At t 13,final Bid 1.1135 is above D13=1.1130. “Any recorded sample below” passes; “completed bar closes below” fails.

Sample below, close abovePreviously defined D(t)=1.1000+0.001×t. Exact fictitious samples:t 12 Bid 1.1130 above line 1.1120;t 12.5 Bid 1.1115 below 1.1125;t 13 final Bid 1.1135 above 1.1130. Only these samples are shown; OHLC alone does not supply an extreme's time.Compare price and line at the same time1.11301.11151.1135t=12t=12.5t=13Line D(t)=1.1000+0.001×t
Green and red points have exact observation times. Dashed vertical segments compare each sample with the line at that time. We do not join samples into an invented price path.

Mid-period difference is−0.0010,or−10 pips. At the end it is+0.0005,or+5 pips. If you have only a bar's Low without its time, do not arbitrarily compare it with the sloping line's final value. You need the sample time or a differently defined rule. Neither check determines future direction.

Assumptions: entirely fictitious synchronous times and Bid values. The lower line is fixed before t 12; t uses a constant unit. The assessed bar ends at t 13. Pip=0.0001. This is no fill-price or trade-result example.

How to trade a wedge breakout step by step

Recognition and execution are separate, as in breakout trading. Consider three unchanged Bid Closes 1.1125 against the rising lower line: D12=1.1120, D13=1.1130, D14=1.1140. Close-minus-line is+5,−5,−15 pips. Close<D(t) starts passing at 13 without price falling. Adding Close(t)<Close(t−1) rejects all three; neither rule is automatically better.

  1. Specify points and timing. Choose confirmed highs/lows, equal-extreme policy and line tolerance. Two points can define a line without establishing a useful trade rule.
  2. Specify activity. MT5 Volumes for Forex counts price changes, not global traded quantity. Compare a numerical filter with no filter; a spike does not prove institutional approval.
  3. Separate sample and Close. Compare each price with the boundary at that same time. A recorded intrabar breach does not imply a completed Close breach, and a bar Low without its time may be insufficient for a sloping-line rule.
  4. Define execution and exits. A subsequent order can fill away from the signal Close, and a retest may never occur. State expiry, stop, target convention and costs; an ordinary stop is not maximum-loss insurance.

A hypothetical example — how to size the target and risk

In a separate fictional GBP/USD model, choose a120-pip initial width and observe a later Close below a previously frozen rising lower boundary. The following trade is an added rule, not the inevitable result of the shape. Assume an actual short Bid fill 1.2600, accompanying Ask 1.2602, stop Ask 1.2640 and target Ask 1.2480:40 and 120 pips from the entry.

For 25,000 GBP on a10,000 USD account, the assigned prices give gross outcomes−100/+300 USD: price-only 1%/3% of that starting base, gross 3:1 and a25% two-outcome break-even rate. With 6 USD round-trip commission and no funding or other charges, outcomes become−106/+294, ratio 294/106 ≈ 2.77 and break-even 106/400=26.5%. These are conditional model calculations, not a pattern win rate or maximum loss. Quote-side fills already include spread.

Recover a separate EUR/USD model from PL: actual short 10,000 EUR at Bid 1.1108 with Ask 1.1110, total commission 4 USD, no other costs. Assigned target Ask 1.1008, stop fill 1.1158 and adverse fill 1.1178 give gross 100/−50/−70 and net 96/−54/−74 USD. A signal Close 1.1110 at t12 is 10 pips below D12, but is distinct from the later Bid fill 1.1108. The simplified+96/−54-only threshold is 54/150=36%, not efficacy.

The 100-pip target in that second model is projected from entry under our test convention. IG discusses initial width, whereas ChartSchool’s Rising Wedge explicitly provides no measuring technique for the decline. Specify measurement and projection origin rather than presenting a universal target.

The most common traps when trading wedges

A later return inside the area remains possible after a closing breach, activity filter or retest. These conditions define different strategies rather than demonstrably removing false breakouts. Retain failed and untriggered candidates; do not infer reliability from only charts that completed the expected move.

Record the chosen trendlines, confirmation rule and invalidation level before entry. Their placement is discretionary and may differ between traders. Size the position from the actual entry-to-stop distance and include costs; an ordinary protective stop sets a planned exit rather than a guaranteed maximum loss.

What to do tomorrow

  1. Reconstruct both line models and check width while the upper boundary remains above the lower. Record when every defining point became known.
  2. Compare the flat-Close counterexample with an added falling-Close rule. For intrabar rules, use synchronized sample times rather than inventing a path from OHLC.
  3. Record source-specific activity, quote-side fills, rejected cases, fees and funding. Compare complete rules on data outside parameter selection; two weeks of demo practice does not establish an edge.