The cup & handle is a bullish base/continuation pattern: a long, rounded U-shaped consolidation (the cup) followed by a short, slightly-downward drift (the handle) that sets up a breakout buy point at the handle's high. It was defined and popularized by William J. O'Neil in How to Make Money in Stocks (McGraw-Hill, 1988) as one of the core base patterns of his CANSLIM methodology, and it is heavily institutionalized through O'Neil's own commercial vehicle, Investor's Business Daily (IBD). The core tension: the geometry is intuitive and the breakout logic is sound (accumulation → shakeout of weak holders → demand overwhelming supply), but a "valid" cup/handle is highly subjective to identify, and its measured base rates — while genuinely good for the breakout itself — are more modest than the IBD reputation implies once you ask for larger gains.

How it's formed

The textbook geometry (O'Neil / IBD, cross-checked against StockCharts ChartSchool and Bulkowski):

• Cup: a smooth, rounded U-shape (explicitly not a sharp V). O'Neil's original specification runs roughly 7 to 65 weeks; ChartSchool frames the typical cup as 1–6 months on daily charts. The two rims (left and right highs) should sit near the same price level, though sources stress flexibility.
• Cup depth: O'Neil's guideline is a retracement of about 12–35% from the left rim, with the cup ideally retracing one-third or less of the prior advance (ChartSchool allows up to 1/2 in volatile markets, 2/3 in extreme cases). Deeper, more jagged cups are lower-quality.
• Handle: forms after the right-side rim, as a short downward drift (often resembling a small flag/pennant slanting down). It must form in the upper half of the cup, retrace only about 5–15% from the recent high, and typically lasts 1–4 weeks (Bulkowski's minimum is 1 week, no maximum). A handle that sags into the lower half of the cup invalidates the setup — a key O'Neil rule.
• Volume: classically dries up through the bottom of the cup and during the handle (supply exhausting), then surges on the breakout — IBD's rule of thumb is volume at least 40–50% above the 50-day average on the breakout day.

How it's used in practice

• Pivot buy point. The actionable trigger is a breakout above the handle's high (the "pivot"). O'Neil's rule is to buy within ~5% above the pivot — chasing further extended raises risk. The handle high (not the cup rim) is the operative resistance because the handle is the final, tightest consolidation.
• Measured target. Both ChartSchool and Bulkowski use the cup depth as the projection: measure the height from the right rim down to the cup's lowest point and add that distance to the breakout price. Bulkowski refines this — the target is met only ~61% of the time, so he scales the projection by that hit rate rather than treating full depth as a promise.
• Quality filtering. Practitioners favor cups that are wide and rounded (not narrow/V-shaped), with rims at similar levels, a shallow handle high in the cup, and clean volume dry-up. In CANSLIM, the chart pattern is only the "N"/timing leg — it is meant to be combined with strong earnings, leadership, and a confirmed market uptrend, not traded as a standalone chart signal.

Swing-specific operational mechanics — exact entry trigger, the 7–8% stop, position sizing, hold period, and add-on rules — live in the Swing Trading branch (see O'Neil CANSLIM & Cup-with-Handle, node #1 of the named-swing-frameworks group, and the broader CAN SLIM strategy node). This doc covers the pattern; that branch covers trading it.

Standing & evidence

This is where honesty matters most, because the pattern's reputation is partly a marketing artifact of O'Neil's IBD ecosystem.

Bulkowski's independent measured stats (the most rigorous public base rates, from his Encyclopedia of Chart Patterns / thepatternsite.com, derived from a large sample of "perfect" upward-breakout cups in bull markets):

• Overall performance rank: 3 of 39 bullish patterns (1 = best).
• Break-even failure rate: ~5% (only ~5% fail to move at least 5% past the breakout).
• Average rise: ~54%.
• Throwback rate: ~62% — the price commonly returns to the breakout level before continuing, which makes a naïve entry feel like a failure.
• Price-target hit rate: ~61% reach the full measured (cup-depth) objective.

The break-even number is genuinely strong, but it is a low bar (just 5% past the pivot). The reputation degrades for bigger gains: in Bulkowski's later sample (300 patterns, 1990–2024) ~47% dropped substantially within two months of the breakout, and roughly 23% rose no more than 15% before reversing. So "95% success" framing (widely repeated from the break-even stat) is technically defensible but materially misleading — it does not mean 95% of trades are profitable holds.

Subjectivity is the central caveat. There is no precise, agreed definition of a "valid" cup/handle; two analysts viewing the same chart can disagree on whether one exists, where the rims are, and whether the handle is acceptable. Bulkowski's flattering stats are drawn from post-hoc, hand-selected "perfect" examples — survivorship/selection bias that the live trader, identifying patterns in real time, will not replicate. Most of the source literature is from O'Neil/IBD's own commercial methodology rather than independent peer-reviewed study; treat efficacy claims accordingly.

Strengths & limitations

• Works best in a confirmed market uptrend, on liquid leading stocks, when the cup is wide/rounded with a shallow handle high in its range and clear volume contraction → breakout expansion. Its appeal is a well-defined, low-ambiguity entry trigger (the pivot) and a built-in stop logic (handle low).
• Fails / misleads when: the cup is too V-shaped or too deep (broken trend, not accumulation); the handle drifts into the lower half (invalid); the breakout comes on weak volume (high false-breakout risk); or the market is in a downtrend. The ~62% throwback rate means a clean entry frequently retests the pivot — premature stops are a common, avoidable loss.
• #1 misuse: treating it as a standalone chart signal divorced from CANSLIM's fundamental and market-context filters, and trusting the "high success rate" headline without recognizing it reflects a low (5%) break-even bar measured on hand-picked perfect examples.

System relevance

Cross-links: this is the definition node for the pattern under Technical Analysis → Chart Patterns. Its swing-trading application (entry/stop/target/hold mechanics, IBD's 7–8% loss-cut rule) belongs to the Swing Trading → Named Swing Frameworks → O'Neil CANSLIM & Cup-with-Handle node, and the full system sits under Strategy Catalog → Growth/Momentum → CAN SLIM (O'Neil). The Augustus agent should consume this as pattern recognition knowledge only, and weight any detected cup/handle by (a) market-regime confirmation and (b) the honest base-rate caveat above — the 5% break-even stat is not a profit probability, and identification is subjective enough that the agent should require corroborating volume/regime evidence before treating a cup/handle as actionable.

Sources

• Thomas Bulkowski, Bulkowski on the Cup with Handle — thepatternsite.com/cup.html (measured rank, break-even, average rise, throwback, target hit rate; later 1990–2024 sample on post-breakout drops)
• StockCharts ChartSchool, Cup with Handle (Continuation) — chartschool.stockcharts.com (cup duration/depth, handle rules, volume, measure-rule target)
• William J. O'Neil, How to Make Money in Stocks (McGraw-Hill, 1988) — original pattern definition / CANSLIM; secondary summaries via IBD, Fidelity Learning Center, and TraderLion for the depth/handle/pivot and 40–50% breakout-volume and 7–8% stop rules
• Investopedia — Cup and Handle (history, subjectivity/limitations cross-check)