A large, multi‑strategy backtest using RSI and momentum signals on leveraged ETFs to rotate among a broad set of assets, with cash‑equal weighting within groups and hedges for downturns, all aimed at capturing varying market regimes without automatic top‑level rebalancing.

- The system runs many sub‑strategies (groups) in parallel, each testing a wide list of leveraged ETFs across stocks, tech, bonds, and even inverse/ Bear bets. - Within each sub‑strategy, it uses momentum signals (like RSI over a window, and momentum measures such as cumulative return or moving‑average comparisons) to decide which assets look strongest or weakest. - It then ranks candidates and picks a small number (often 1 or 2) to hold, giving those picks a cash allocation (typically cash equal weighting within that group). - The process is repeated across many groups, so the overall portfolio ends up as a mosaic of different styles and lookbacks. - Many components are levered ETFs (e.g., TECL 3x tech, TQQQ 3x QQQ, UPRO 3x S&P 500) or bear/inverse bets (SQQQ, PSQ, SH, LABD, LABD, HIBS/HIBL) to address different market regimes. - There is no single scheduled rebalance at the top level; decisions to enter/exit are driven by the outcome of each group’s rules, effectively rotating exposure as signals change. - The overall aim is to improve performance by capturing upside in rising regimes while providing hedges during downturns, using a wide array of scenarios and lookbacks. - The language and structure show explicit lookbacks (e.g., 4d, 11d, 20d, 15d windows) and multiple nested decision rules, indicating a zoo of conditional tests rather than a simple buy/hold rule.

Out-of-sample annualized return ~59% vs ~25% for the S&P, with regime-rotating hedges across leveraged ETFs. Calmar ~1.38 and Sharpe ~1.09 indicate strong risk-adjusted upside across markets, not a single-market bet.