A two-branch tactical strategy: if SPLV RSI > 75, lean to a bearish S&P bet via SPXS; else build an inverse-volatility weighted basket of six large-cap stocks (COST, NVO, LLY, PGR, NVDA, AMZN). Equal cash between branches; no fixed rebalancing schedule, with a small tolerance for switching.

Here’s how it works in plain terms: 1) The system watches SPLV (a low-volatility version of the S&P 500) and looks at its momentum over the last 10 days. Momentum is measured with RSI, a simple gauge of how strong recent price moves have been. 2) If SPLV’s RSI is above 75, the system places a bet that the market will fall by allocating funds to SPXS, a leveraged inverse ETF that moves opposite to the S&P 500 (about 3x daily). In plain terms: if you expect stocks to drop, SPXS goes up, and this branch aims to profit from that drop. 3) If SPLV’s RSI is 75 or below, the system ignores the SPXS bet and instead builds a long portfolio of six large-cap stocks (COST, NVO, LLY, PGR, NVDA, AMZN). Weights are assigned using an inverse-volatility rule based on a 20-day lookback, meaning more weight goes to steadier, less volatile stocks and less to more volatile ones. 4) The two branches are treated as equal-cash options (wt-cash-equal), so capital is split roughly evenly between the short bear bet and the long stock basket when both are active. The model does not use a fixed calendar for rebalancing (rebalance: none) but uses a small tolerance (rebalance corridor width of 0.1) to decide when to adjust if conditions move. This setup is meant to switch exposure mainly on the SPLV signal rather than on a fixed schedule. 5) The strategy sits in equities with a short-term horizon emphasis and carries risks from leverage (SPXS) and concentration in a limited stock set on the long side. It seeks to combine a bearish hedge when momentum on SPLV is extreme with a diversified, lower-volatility-weighted stock portfolio otherwise.

Out-of-sample edge: lower max drawdown (17.0% vs 18.8%), lower volatility, and Calmar ~0.84 from a hedged two-branch design. Annualized return ~14.3% vs SPY’s 17.7%, but with tighter risk control and more resilience.