Technical pattern definitions and reader-added rules
Make price geometry, equality, context and information timing explicit before attributing or testing a chart premise.
A name does not identify one numerical event
A pattern description may leave body ratios, shadows, equality, lookback and context implicit. Two readers can therefore produce different labels without either implementation being uniquely attributable to the source. Match the actual edition and inspected location before claiming a precise author threshold. The retained chart-reference companions keep original worked geometry and source data; this route adds a translation checklist rather than a new predictive result. A reference can define useful objects while not proveing positive expectancy for every possible numerical interpretation.
Work the geometry before the market hypothesis
Our original candle has open 100, high 104, low 98 and close 103. Its body is three, upper shadow one and lower shadow two, with total range six. A named label still requires your declared ratio and context. Zero-range and equality cases should remain explicit rather than be silently dropped. These are authored values, not actual exchange observations or an inspected book's exercise. Correct geometry is an intermediate learning output, not a causal entry, appropriate quantity or independently validated financial system.
The window and timing belong in the definition
A moving reference must state which earlier observations are included and when they became available. A developing bar cannot inherit its final close before completion. Revised price history or present membership may introduce another source mismatch. Preserve all candidate definitions and outcome-informed changes rather than choose whichever pattern looks strongest afterward. A source hash, clean chart or current runtime version cannot independently certify those choices. The reader's operational additions need separate identity before real market evaluation or attribution.
Testing requires more than one attractive figure
A genuine hypothesis needs complete eligibility, source causality, baseline and independent design. Actual instrument completion retains the canonical full factory/master and native entry/exit and fill rules. A synthetic image and a passing software fixture cannot replace the absent financial population. The useful reading output is a definition agreement table, one numerical example and an unresolved observation record. A rejected ambiguous definition can be progress without making the entire book poor or declaring that every chart technique lacks or has an edge.
Original reading and evidence case
Two learners agree on body three and shadows one/two but choose different thresholds for a named label. The geometry is reproducible while label agreement remains unresolved. Both translations stay in the register, with no source claim that the author prescribed either. Selecting the version with a favourable later outcome would be additional research selection, not independent validation of the original teaching figure.
| Object | Record | Bounded finding |
|---|---|---|
| Geometry | Body 3, shadows 1/2 | Authored inputs |
| Label | Reader-added threshold | Not author attribution |
| Clock | Earlier available input | Not final-bar hindsight |
| Evaluation | Actual independent design | Not one chart |

Primary references and scope
Source review: 14 October 2026. Primary records identify their own scope, not a full-book verdict, measured reader outcome or paid-account test.