Methodology
Chart Pattern Backtesting Concepts
Pattern Repeater compares completed historical chart windows. It does not automate a complete trading-strategy backtest.
What Pattern Repeater does
Choose a ticker, date range, time step, and match threshold. The app compares that chart shape with eligible completed historical windows and shows the later outcomes that followed those comparable cases. This is useful historical context for a question such as: what happened after charts that looked like this?
What a full backtest would require
A trading-strategy backtest needs explicit entry and exit rules, position sizing, execution timing, spreads, slippage, fees, taxes, corporate actions, delisted securities, and the information actually available at each point in time. Pattern Repeater does not simulate those rules or produce a strategy performance report.
Define the research question before results
Write down the chart window, timeframe, minimum similarity, follow-up period, and outcome metric before reviewing matches. Changing settings until an appealing result appears is optimisation for the past, not a useful test.
Use the full eligible sample
Report all matching completed windows that pass the chosen cutoff. Useful descriptive statistics include sample count, median outcome, range, and the share of positive and negative outcomes. They describe the historical sample; they do not prove a trading edge.
Methodology FAQ
Can recent matches belong in a completed outcome test? No. Their future period is pending.
What would weaken a result? A tiny sample, one ticker, one date range, or results that disappear under reasonable settings.
Can this prove a trading strategy works? No. A strategy needs independently specified trading rules and its own complete backtest.