Methodology

Overview

YouPattern is a visual encyclopedia of technical analysis patterns. Every diagram, chart example, and cheat sheet on this site is generated programmatically using Python with historical OHLCV (Open, High, Low, Close, Volume) market data. This page explains our methodology in full — from data sourcing to pattern validation to visual rendering.

1. Ideal Pattern Diagrams

Each pattern page includes an Ideal Pattern Diagram — a clean, annotated chart that shows the textbook structure of the pattern without noise from real market data.

How ideal diagrams are generated

• Tool: Python 3.11 with matplotlib and mplfinance
• Data: Synthetic OHLCV data constructed to match the exact mathematical criteria of each pattern (e.g., for a Hammer: lower shadow ≥ 2× body length, body in upper third of range)
• Annotations: Key measurements, ratios, and identification criteria are labelled directly on the chart
• Style: Dark background, consistent colour scheme (green for bullish candles, red for bearish), clear gridlines
• Resolution: 900×630 pixels at 150 DPI for crisp display on all screen sizes

Ideal diagrams are intentionally simplified — they show the pattern in its most recognisable form. Real-world examples will always look slightly different.

2. Real Chart Examples

Each pattern page includes 3 real chart examples showing the pattern as it appeared on actual market data.

Asset selection

We primarily use BTC/USDT, ETH/USDT, SOL/USDT, BNB/USDT and selected high-liquidity cryptocurrency pairs sourced from the Binance public API. Asset selection per page is based on where the pattern is most clearly and representatively visible — not a fixed rotation.

• BTC/USDT — Bitcoin, highest liquidity, most widely followed cryptocurrency
• ETH/USDT — Ethereum, second-largest cryptocurrency, high liquidity and pattern clarity
• SOL/USDT — Solana, high-growth asset with strong pattern formation history
• BNB/USDT — Binance Coin, selected for specific pattern examples with high clarity

Timeframe selection

Timeframes vary by page and include Daily (1D), 4-Hour (4H), and Weekly (1W) depending on the pattern type and example clarity. The timeframe used is labelled on each chart image. Backtest statistics specify the exact timeframe(s) used in the backtest for that pattern.

Pattern detection algorithm

Real examples are identified using a rule-based algorithm that checks the mathematical criteria for each pattern against historical OHLCV data:

• Each pattern has a defined set of quantitative criteria (e.g., shadow-to-body ratio, relative position of peaks, RSI thresholds)
• The algorithm scans historical data and flags candles or sequences that meet all criteria
• Flagged examples are then visually reviewed to confirm they represent clear, unambiguous instances of the pattern
• Only examples where the pattern is clearly visible to a human analyst are included

Chart rendering

• Tool: Python with mplfinance
• Context window: 30–60 candles surrounding the pattern, with the pattern candle(s) highlighted
• Annotations: Pattern candle(s) marked with arrows or boxes; key price levels shown
• Resolution: 600×300 pixels at 150 DPI

3. Cheat Sheets

Each pattern page includes a Quick Reference Cheat Sheet — a compact visual summary of the pattern's key identification criteria, common mistakes, and confirmation signals.

• Generated programmatically using Python matplotlib
• Contains: pattern diagram thumbnail, identification criteria checklist, signal type (bullish/bearish/neutral), difficulty level, and key rules
• Designed for quick reference while scanning live charts
• Available for free download as PNG
• Resolution: 900×540 pixels at 150 DPI

4. Pattern Criteria and Definitions

Pattern definitions on YouPattern are based on the following primary sources:

• Steve Nison — Japanese Candlestick Charting Techniques (1991) — the foundational reference for candlestick patterns
• Thomas Bulkowski — Encyclopedia of Chart Patterns (3rd edition) — statistical analysis of chart pattern performance
• John J. Murphy — Technical Analysis of the Financial Markets — comprehensive reference for technical indicators
• CMT Association curriculum materials for standardised definitions

Where sources disagree on specific criteria (e.g., minimum shadow-to-body ratio for a Hammer), we use the most widely accepted definition and note any significant variations in the pattern description.

5. Quality Standards

Accuracy

• All pattern criteria are cross-referenced against at least two primary sources before publication
• Real chart examples are visually reviewed by a human analyst before inclusion
• Ideal diagrams are checked against the mathematical criteria to ensure they accurately represent the pattern

Freshness

• Real chart examples are generated from historical data up to the most recent available date at time of publication
• Each page displays a "Last updated" date
• We review and update examples periodically as new clear instances of patterns appear in market data

Transparency

• All charts are clearly labelled with the asset, timeframe, and approximate date range
• We distinguish between ideal (synthetic) diagrams and real (historical) examples
• Limitations of pattern analysis are noted in the disclaimer on every page

6. Limitations and Disclaimer

Technical analysis patterns are statistical tools, not predictive guarantees. Past pattern performance does not guarantee future results. All content on YouPattern is for educational purposes only and does not constitute financial advice, investment recommendations, or trading signals.

Pattern recognition is inherently subjective — two analysts may identify the same pattern differently, or disagree on whether a specific chart instance qualifies. Our methodology aims to be as objective and consistent as possible, but users should apply their own judgment when identifying patterns on live charts.

See our full Disclaimer for complete terms.

Reporting Errors

If you notice an error in a pattern definition, an incorrectly labelled chart example, or any other inaccuracy, please contact us. We are committed to maintaining the highest possible accuracy and will review and correct errors promptly.