AI Trading Strategy Research with Backtesting

You have 2 hours. Research trading concepts and build strategies for BTCUSDT, backtesting every one with fees and slippage on. Before you start, hide the most recent third of the data from yourself. Build and optimise only on the older two thirds. Rules for keeping a strategy: 100+ trades, max drawdown under 20%, profit factor above 1.1, and it has to beat buy and hold on the same period. No multiple timeframes, no trailing stops, every exit fixed at entry. When a strategy passes, validate it on the third of the data you hid. Anything that falls apart there gets binned with the reason. Log every strategy you test, pass or fail, and show me only the survivors at the end.