Price action explained.
Candles, trends & levels.

One candle. Four prices.

A candle represents an interval: for example, one minute, one hour or one day. Its body connects the opening and closing prices; the wick extends to the interval’s high and low. An up candle closes above its open, and a down candle closes below it. Colours depend on chart settings.

An unfinished candle can change shape before the interval ends. Also, open–high–low–close data do not tell you which extreme occurred first. A backtest using only candles must make an execution assumption if both a stop and target were touched inside one bar.

Describe the swings before naming the trend.

A rising sequence is commonly described through higher swing highs and higher swing lows. A falling sequence uses lower highs and lower lows. A range revisits an area without sustained progress in either direction.

The timeframe matters. A pullback on a daily chart can contain several falling swings on a five-minute chart. Write down the timeframe and the rule used to identify a swing before labelling a change of structure.

This is an illustrative definition, not a recommended strategy. Using the later confirmation as though it were known at the earlier high introduces look-ahead bias.

Treat support and resistance as areas to investigate.

Support describes an area where a decline has previously met buying interest. Resistance describes an area where an advance has previously met selling interest. Prior highs, lows and repeated reactions are common reference points.

A thin line is a drawing choice, not proof that every participant values the same exact price. Use a consistent zone rule rather than widening the area after a trade loses. A former resistance area may later behave as support, but this role reversal can fail.

For a reproducible exercise, mark a previous session’s high and low before the next session begins. Record how price behaves on the first revisit. Keep every observation, including clean breaks and days with no revisit. This avoids selecting only the attractive examples after the outcome is known.

A break is an event, not confirmation of profit.

A wick beyond a level, a close beyond it and a later retest are different conditions. If your rule says “breakout”, define which one you mean. Also state the timeframe, distance threshold, order type and time window.

Imagine a prior high of 1.1050. A bar trades to 1.1054 and closes at 1.1048. A touch-based rule and a close-based rule reach different decisions. Neither becomes correct merely because the next candle moved in the desired direction.

A false break is a description applied when price returns after breaking a reference area. It is not necessarily knowable at the first crossing. Avoid explaining a losing entry with information that arrived only later.

Two candles, one explicit breakout rule.

Use a rule fixed before the outcome: a completed candle must close strictly above a prior high of 1.1050. The original diagrams below show two invented candles on the same price scale. They explain what satisfies that definition, not which trade would make money.

A touch-based rule would count both candles as crossings; the close-based rule counts only A. Specify the timeframe and quote side as well: a bid chart and an ask chart can cross a level at different times. A close exactly at 1.1050 would not satisfy the strictly-above condition.

Neither diagram includes the next candle. Using later price movement to label an earlier entry as “obviously valid” adds information that was not available then. Also, OHLC data alone cannot tell you whether the high or low occurred first within either candle.

Pattern names need measurable rules.

The purpose of a precise definition is to make observations comparable. A collection of labelled screenshots is not the same as a strategy evaluated on an unseen sample.

Turn a chart idea into an honest test.

1. Freeze the rules. Define the market, timeframe, session, setup, entry, stop, exit and cancellation conditions.
2. Use available information only. Account for swing-confirmation delays and incomplete candles.
3. Model execution. Include bid/ask differences, commission, slippage and intrabar uncertainty.
4. Separate development from evaluation. Keep data you did not use to tune the idea and record any later changes.
5. Inspect the distribution. Review losses, drawdown, trade count and average net outcome—not just win rate.

Use position sizing to understand exposure. It can constrain a modelled trade loss; it does not turn an unprofitable entry rule into an edge.

Common questions.