Three Outside Up and Down: body edges, color and third closes

Three Outside Up and Down combine an engulfing-body relationship with a comparison of the third and second closes. Confirmation means the selected condition is complete. It does not prove a lasting trend reversal, identify the traders responsible or guarantee a profitable execution.

What are the Three Outside Up and Three Outside Down patterns?

In TA-Lib 0.6.4 CDL3OUTSIDE, Up requires C1 < O1, O2 < C1, C2 > O1 and C3 > C2. Down requires C1 ≥ O1, O2 > C1, C2 < O1 and C3 < C2. The engulfing inequalities are strict. The function imposes no historical body-size threshold or preceding-trend test; record any additional context filter separately.

The bullish and bearish Engulfing pattern compares bodies, the intervals between Open and Close. A second body can cover the first without covering its entire High/Low range. The third-close comparison does not specify the third body’s color: C3 above C2 can coexist with C3 below O3, and the converse can occur for Down.

How does Three Outside relate to Engulfing and to Three Inside?

Three Outside has an engulfing second body, while the Three Inside Up and Down setup has a contained second body. The different third-close and equality rules must also be checked in the selected detector. Combining two generic pattern names is not an exact specification.

The Harami pattern and engulfing geometry select different observations. Their relative body sizes establish no ranking of reliability, decisiveness or profitability. Compare fully stated rules and costs rather than interpreting the larger candle as proof that one side took control.

How do you spot the setup on a chart?

For the fictional Up example, assign earlier closes 1.1060, 1.1040, 1.1020; for Down, 1.0940, 1.0960, 1.0980. This three-close context is an illustrative extra condition, not the library’s trend definition. Use one price series and specify its candle clock; these examples do not describe real sessions or the frequency of interbar price differences.

A strict body condition cannot be replaced by a wick crossing or an equal boundary. O2 = C1 fails the Up rule here; C3 = C2 fails the final comparison. Do not silently relax a condition because the instrument is FX. A later Open is not universally identical to the preceding Close, and a chosen inclusive variant needs its own name and evaluation.

What does it look like in an example?

The following complete fictional EUR/USD bars assume one pip = 0.0001. Both examples strictly engulf the first body while leaving the second full range inside the first. They deliberately separate third-candle direction from the change between closes.

Where do you place the entry, stop-loss and target?

After candle 2 the engulfing condition can be assessed, but the final third-close condition remains unknown until candle 3 closes. A live price above C2 can fall back before that Close. State the decision time and subsequent order separately: the pattern does not mandate a market entry, breakout order or single stop level.

A boundary beyond a candle or the whole formation is an independent invalidation choice. Calculate its distance using the intended execution price, position size and costs; an ordinary stop does not guarantee its trigger or a maximum loss. Target, expiry and optional indicator filters likewise require explicit rules and evaluation. Pattern recognition supplies neither a probability nor a universal reward-to-risk threshold.

What the extra close condition changes

The additional condition changes the selected observations and recognition time. It does not by itself prove fewer losing trades or a better net result. Compare the complete two- and three-candle methods on the same data with realistic submission, fills, exits and costs; do not backdate the three-candle decision.

The morning star pattern uses a different geometric construction. Its label and the word confirmation do not establish a performance hierarchy. Keep body relationships, context and final-price thresholds separate from the later trade outcome.

A concrete scanner counterexample illustrates why Engulfing > 0 plus C3 > C2 is not necessarily CDL3OUTSIDE. Assume three earlier bars (100,110, 98,105), then O/H/L/C (100,101, 96, 98), (O2,102, 95,101), (101,103,100,102). With O2 = 97, Engulfing gives +100 and Three Outside +100. Change only O2 to 98: Engulfing gives +80 because one body edge is equal, while strict Three Outside gives 0. C3 > C2 holds in both. These are hypothetical source-condition checks, not trading outcomes.

TA-Lib’s non-bearish color convention also allows a first doji in Down, but not in Up; a third doji can satisfy either final-close comparison. A custom filter excluding such equality cases must be recorded rather than treated as the same detector.

What to do tomorrow

1. Choose a detector version and test strict and equal body edges explicitly. Record any separate preceding-context or candle-color filter.
2. Calculate C3 relative to C2 and to O3 separately. A falling third body can still close above the preceding candle.
3. Save the feed, time zone and price side. The unchanged diagram shows Bid 104→105 while synchronous Ask 107→106 falls; spreads 3→1 stay positive. This tests one close condition, not the full pattern.
4. Define decision time, executable order, exit, exposure and costs before examining results. Retain rejected, missed and losing cases.
5. Use the technical analysis guide to compare definitions. Geometry exercises need no live trade, and selected charts do not establish an edge.