Measuring Single-Candlestick Pattern Frequencies Across Markets

Summary

This article turns individual candles into symbols and counts their occurrences in samples of GBPUSD and gold across hourly, 15-minute, and 5-minute charts. It describes a workflow that encodes historical bars, groups Marubozu variants together, assigns unmatched candles to an unclassified category, and reports counts, percentages, and simple distribution summaries. An MQL5 script and text report are presented as tools for producing these profiles. The reported samples show Marubozu symbols as the most frequent classified types, with bullish and bearish totals generally close. Around a third of observations are unclassified, and the article suggests this may reflect strict classification thresholds. It also proposes exploring whether unclassified candles divide directionally in a similar way, but acknowledges this cannot be established from the current encoding. The analysis covers single candles only and does not demonstrate predictive power or a trading edge; sequence analysis and out-of-sample evaluation are identified as necessary next steps.

Key ideas

• Individual candlesticks can be mapped to a finite alphabet and analyzed as countable events.
• The study compares symbol frequencies in 1,500-bar samples of GBPUSD and gold across three timeframes.
• Marubozu categories are the most common classified symbols, while bullish and bearish classified totals are broadly similar.
• A substantial share of candles falls outside the defined categories, limiting conclusions about those observations.
• Single-candle frequency profiles describe distributions but do not establish predictability or profitability.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.