Define Pennants before you test anything
A pennant is a short-term chart formation described as a small continuation pattern: after a relatively strong move, price consolidates into a narrowing range (often drawn with two converging trend lines) before breaking out. Verification starts with a shared definition, because different sources may use different tolerances for “narrowing,” “converging,” and what counts as the prior move.
Stable mechanics to lock down:
- What “the prior move” means (direction, approximate length, and whether it must be substantial).
- How you draw the two converging boundaries (line-touch rules, which swing points you select).
- What “breakout” means (a close beyond the boundary, an intrabar touch, and whether volume is required).
Build a source hierarchy for claims you want to verify
Use information sources in a hierarchy so you know what is solid and what is commentary:
- Charting and pattern-definition material (stable): educational explanations that define geometry and rules.
- Method documentation (reproducible): how authors measure the pattern (selection rules, thresholds, and breakout conditions).
- Empirical discussions (variable): examples showing outcomes; treat these as observations, not guarantees.
- Provider- or platform-specific claims (variable): any statements that depend on execution, fees, or data processing. Verify by looking for the exact method, not marketing.
When a claim depends on current conditions (for example, how a broker charges or how a platform computes prices), you cannot verify it from pattern geometry alone. In that case, require the underlying documentation and confirm the exact definitions used.
Reproducible verification steps you can run yourself
1) Confirm you can reproduce the drawing
Pick one historical window you already have on your chart (no real-time data needed). Then:
- Apply your definition: identify the prior move, then select the swing points that anchor the converging lines.
- Use the same rule each time (for example: “lines must touch at least two swing points” or “use the first and last visible touches”).
- Measure the narrowing: does the range clearly contract, or is it a wide sideways drift?
If two independent viewers using the same rules disagree, that is a sign the definition is too subjective for reliable verification.
2) Test sensitivity to measurement choices
Repeat the same window with small changes in your inputs:
- Use a slightly different pair of swing points for the boundaries.
- Change whether breakout is defined by “close beyond” versus “touch of the line.”
A pattern that remains consistent under reasonable drawing differences is more verifiable than one that appears or disappears based on tiny adjustments.
3) Check outcomes without treating history as proof
If you record what happened after breakouts, do it as a descriptive check:
- Document your assumptions: timeframe, asset, and exact breakout rule.
- Separate descriptive results from causal claims.
Historical relationships do not establish future results, so verification here is about understanding how often your rules produce the formation—not about predicting.
4) Look for limitation and failure modes early
Material limitations commonly include:
- Ambiguous boundaries: converging lines can be drawn in multiple ways, producing different “pennants.”
- Overfitting: using very specific thresholds that match past examples but fail elsewhere.
- Selection bias: only reviewing cases where the breakout “worked.”
- Execution and costs: even if a breakout is visible on a chart, real fills depend on liquidity, spreads, slippage, and order type.
Because outcomes vary with market conditions, costs, execution, and jurisdiction, your verification must include what is and is not captured by the chart description.
Verification checklist and next question to ask
Use this checklist whenever you evaluate “Pennants” information from an article, video, or forum post:
- Definition clarity: Are the rules for the prior move, convergence, and breakout explicit?
- Reproducibility: Can you redraw the same pennant using the same selection rules?
- Scope: Is the claim stable (geometry) or variable (current execution conditions)?
- Assumptions: Does any example state timeframe and measurement method?
- Failure modes: Does the author address ambiguity, costs, or selection bias?
Next question to ask: Are you verifying the formation rules, or are you verifying a claimed outcome? Pennants are easier to verify as a chart-geometry concept than as a standalone predictor of results.