Crypto Strategy Optimization: Choose the Market Before the Parameters

Start with an instrument specification: exchange, symbol, quote currency, spot or derivative, chart interval and test dates. Two BTC charts can have different prices, volume and available history. Do not treat a spot chart as a complete simulation of a leveraged perpetual contract. Document any omitted funding, borrowing, liquidation or execution costs before comparing results. Continuous trading does not imply constant liquidity or uninterrupted venue access. Inspect the selected venue and period before assuming weekend volume, spread or slippage follows a fixed pattern. If you test time-of-day filters, fix the timezone and define the rule before evaluating the holdout. A maintenance interruption or missing bar is a data condition to investigate, not an entry signal. A useful first experiment changes only a few inputs while keeping the dataset and costs fixed. For example, 5 lookback values, 4 exit distances and 2 filter states make 40 candidate configurations. This is a hypothetical test design, not a performance result. Retain the original baseline, inspect nearby configurations and evaluate the chosen rules on dates excluded from selection. Record every extra search: refining a winning region still uses the same evidence and does not create a new independent test.