Chart Pattern Confidence Scores: What the Numbers Mean

What a Confidence Score Measures

A chart pattern confidence score summarizes how closely a detection matches the detector's criteria. It is not a calibrated probability of reaching a target or making money. A score of 80 does not mean an 80% win rate.

Scores help prioritize chart review. The pattern's entry, target, stop, timeframe, liquidity, and current price still need separate inspection.

How the Composite Score Is Calculated

The scoring pipeline starts with the pattern detector's own assessment and combines five components. The current base weights are:

| Component | Base weight | What it measures |

|---|---|---|

| Shape fit | 40% | Detector-specific geometry, with available proportion and price-symmetry checks |

| Volume confirmation | 20% | Available volume evidence relative to the formation or recent average |

| Trend context | 20% | The preceding trend on the chart being analyzed |

| Time symmetry | 10% | The balance of left and right formation lengths, when supplied |

| Pattern completeness | 10% | Whether the detection is forming or confirmed |

These are base weights, not a promise that every detection has five measurements. Components without a measurement are excluded and the remaining weights are renormalized. Pattern families supply different geometry: harmonic ratios, candle bodies and shadows, or the shape of a multi-swing formation.

The detector's geometry score is the main shape-fit input. Available metadata can adjust it; the score is not simply a distance from one universal textbook curve. Trend context examines the supplied chart, rather than automatically awarding a bonus for agreement with a separate higher timeframe.

Interpreting Higher and Lower Scores

A higher score means the available inputs fit the scoring criteria more closely. A lower score suggests inspecting the chart for imperfect proportions, weaker supporting evidence, or incomplete formation. Neither score proves that a breakout will follow through.

Missing measurements and different pattern rules limit comparisons. Start by comparing detections of the same pattern type and timeframe. Do not interpret a 70-score cup and handle and a 70-score bull flag as having identical odds.

Confidence Scores and Historical Outcomes

To establish a relationship between scores and win rates, a study needs a dated sample, score bands, timeframe, outcome rules, and sample sizes. This article does not provide such a calibration study, so it does not assign completion probabilities to score bands.

The track record methodology explains the platform's outcome metric. Its win rate includes positive timeout returns as well as target hits; it is not a measured-move completion rate. A visually strong pattern can still stop out or expire.

Using Scores During Chart Review

1. Filter to a manageable set. Choose a review threshold appropriate to your research, then inspect the actual formation and levels.
2. Check confirmation. A forming detection and a confirmed breakout represent different stages, even when their scores are close.
3. Inspect the risk geometry. Calculate the distance to the stop and target from the entry you could actually obtain.
4. Compare outcomes separately. Read sample sizes and resolution rules in pattern histories. A high score alone is not a reason to increase position size.

Scores can change when a detection is recalculated or its inputs change. A revision does not by itself demonstrate improving future returns; compare the chart and detection status as well as the number.

Find Scored Detections

The scanner displays confidence scores alongside detected patterns. Use them to organize research, and check the pricing page for current access to patterns and tools.