Direct answer
To assess a Double Top Bottom concept, you need data that lets you (1) define the two swing points, (2) measure the distance and relationship between them using consistent rules, (3) specify the timeframe and price type you used, and (4) verify that the conclusion is not an artifact of data handling or selective observation.
Because outcomes vary, treat this as an evaluation of structure, not a prediction. Also document what cannot be verified from charts alone, such as future behavior, trading costs, and execution conditions.
Mechanism and definition: what data lets you “see” it
A Double Top Bottom pattern is usually described as market structure with two related turning points (often two tops for a “double top” or two bottoms for a “double bottom”), plus a connecting reference level that separates the first move from the next.
To assess it with clear inputs, collect:
- Price history: a time-ordered series of open/high/low/close (OHLC) or another consistent price definition (for example, using highs for the top and lows for the bottom). Stable input means you know exactly which fields you used.
- Swing-point identification evidence: the timestamps and prices of the two candidate turning points, and the connecting “intermediate” level you treat as the key reference (for example, the level between the two swings).
- Timeframe definition: the bar interval (such as 1H vs 4H) and whether you analyze candles, fractals, or another method. Timeframe changes what counts as a “swing.”
- Context boundaries: a defined lookback window (how far back you examine) and a rule for where the pattern search begins. Without this, assessment risks “lookback bias” (seeing the pattern only after it appears).
How it “works” in practice (conceptually): you take the first turning point, identify the second one within your timeframe rules, and then check whether the relationship to the intermediate reference level matches your stated criteria. The key point is that every decision must trace back to the data fields above.
Evidence or example: what checks make the assessment reproducible
To make your assessment independently verifiable, you need quality checks and measurement discipline.
Use repeatable measurement rules and record them:
- Equality or similarity rule: define how you decide the two tops/bottoms are “similar” (for example, within a tolerable range measured in price units or percentages). You must state the assumption.
- Reference level rule: define which level matters for the pattern’s internal logic (for example, the intermediate trough/peak level connecting the two swings). Again, state which OHLC component you use.
- Event ordering rule: record that the first swing occurred, then the second swing occurred later, and then the intermediate level relationship was tested according to your definition.
Add data provenance and timeliness documentation:
- Source provenance: note where the price series came from (market data provider or platform export) and how it was obtained (manual charting vs downloaded historical data). Even without naming the provider, the process must be describable.
- Timeliness: confirm that the data snapshot you used matches the timeframe you claim (charts updated later can change the most recent swing boundaries).
Finally, include red-flag checks:
- Lookback bias: if you “found” the pattern only after seeing the later move, you should redo the search using a pre-defined window and criteria.
- Ambiguous swing selection: small differences in swing identification methods can change the “two points.” If your two peaks/bottoms are not stable under a reasonable alternate swing rule, the assessment is fragile.
- Overfitting: using many adjustable criteria until it “matches” a specific past segment reduces external validity.
Limitations and risks: material failure modes
At least one important limitation is structural: a chart resemblance can be incomplete or misleading.
Common failure modes include:
- Timeframe sensitivity: the pattern may appear on one timeframe and not another. This does not mean it is “wrong,” but it affects how confident you can be in any conclusion.
- Measurement subjectivity: if the criteria for “two similar tops/bottoms” are not stated, different analysts can reach different results from the same data.
- Non-pattern variables: historical relationships do not establish future behavior. Costs, liquidity conditions, and execution quality can change what you observe versus what you would experience.