Produce a chart interpretation that keeps every visual layer attached to its evidence basis.
By the end, you should be able to name the data source, instrument, interval, displayed range, layer units, and the time at which each layer became knowable. If any of those are missing, narrow the claim.
Identify the source before looking for a pattern.
- Data source: who supplied the bars, events, indicators, or profile bins?
- Instrument and interval: which market and aggregation rule does the picture represent?
- Range and session: which dates, trading hours, and time zone are included?
- Units: price, volume, time-at-price, indicator units, or another basis?
Use OHLC bars for ordered range evidence, not hidden intrabar certainty.
Open, high, low, and close summarize one interval. Read gaps, range expansion, clustering, and sequence across bars, but do not infer whether the high or low occurred first unless finer-grained evidence establishes it.
OHLC interpretation limits →Read Profile and TPO as descriptions of concentration under a stated construction rule.
Confirm bin size, session bounds, source units, and displayed range before comparing nodes or sparse regions. A prominent profile node describes where measured activity concentrated; it does not become support or resistance merely because it is visually prominent.
Profile & TPO interpretation limits →Historical replay is useful only when future information stays hidden.
At each replay cutoff, ask which bars, events, profiles, and derived layers were actually available. A final-series indicator can leak information even when every plotted timestamp appears historical.
Historical replay rules →Keep event and trade identity exact instead of snapping it to the nearest visual story.
When an event or trade marker is present, keep its source timestamp, price, direction or semantic label, and source identity. A marker can explain where a recorded event occurred; it does not prove causation or validate the surrounding pattern.
When price, profile, events, and replay tell different stories, investigate the mismatch.
Visual layers summarize different dimensions. Disagreement can come from different time windows, units, sessions, data availability, or simply from the fact that a distribution and a path answer different questions. Do not average them into a verdict.
A compelling picture can still be weak evidence.
- Inferring intrabar order from one OHLC candle.
- Comparing profiles built from different sessions, units, or binning rules.
- Calling a high-volume node automatic support or resistance.
- Showing final-series indicators during replay and treating them as point-in-time evidence.
- Snapping events to bars when exact timestamps are available.
- Ignoring missing data, stale data from the source, or a changed chart range.
Before using a chart as evidence, carry these fields with the screenshot or conclusion.
- The data source and a fixed reference to the exact data used, where available.
- Instrument, timeframe, range, session, and time zone.
- Layer names, units, and construction rules.
- Event/profile availability timing for replayed views.
- Any missing, withheld, or unavailable layer.
- The narrower question the chart actually answers.