A daily, signal-driven rotation among defensive bond/sector ETFs and AI/Infra semis, using short- and medium-term momentum (RSI) and moving-average trend checks to pick 2–3 assets per sleeve from defined groups.

- Step 1: Gauge the market regime using a bond proxy (AGG) by comparing its short-term and longer-term price trends. If the short-term price is above the longer-term price, the model interprets this as a more favorable regime for selecting specific equities; otherwise, it biases toward defensive plays. - Step 2: Check semiconductors’ momentum signals (via RSI and moving averages on SMH and SOXX). Extremely high RSI readings (around 79–80) on SMH or SOXX trigger certain sleeves, while other related momentum and trend checks can steer toward AI/Infra or defensive groups. - Step 3: Choose a sleeve for allocation based on signals. The main sleeves are: - Safe BND/Sector: a cluster of defensive ETFs and bond-related funds (short-term bonds, long-term Treasuries, investment-grade bonds, and defensive equity sectors like Consumer Staples, Utilities, Health Care). Within this sleeve, the strategy screens the group by RSI over a short window (e.g., 10 days) and then selects the bottom 2 assets (relative to RSI) to hold. This can be viewed as a contrarian tilt among defensive assets or a rotation into underperformers that may revert. - AI Infra: a basket of AI-infrastructure and defense-forward equities (e.g., Innodata, Arista, Marvell, Lumentum, Coherent, Astera Labs, Corning, Pure Storage, Seagate, Ciena, Vertiv, Archer Aviation, Joby, etc.). The strategy first picks the top 3 by RSI over a 14-day window, then from that subset picks the top 2 by RSI over a 7-day window. This is a momentum-driven tilt toward AI hardware and related tech if conditions allow. - Additional filtering steps (e.g., ITA-related aerospace/defense ETFs and related equities) provide supplementary exposure within the AI/Infra theme. - Step 4: Within each active sleeve, weights are set (often 100% allocated per path in the given branch) and the selected assets are the ones you would buy or hold until the next daily rebalance. - Step 5: Rebalance daily, recomputing signals and re-selecting the top/bottom assets per the current regime. - Step 6: Asset class and exposure: all selections are equities or equity-ETFs (defensive equities and AI/Infra equities), but the structure groups include bond-related ETFs to enable a defensive tilt. - Step 7: Risk checks and guards: multiple decision nodes enforce trend/price checks (e.g., current price vs. a moving-average benchmark like 200 days) to avoid buying into downtrends. - Step 8: Output: a daily, rule-based roster of assets to hold in the portfolio under the current regime, with explicit groupings and a rotating selection of 2–3 assets per sleeve depending on signals.

Daily, rule-based rotation among AI/Infra semis and defensive ETFs targets superior risk-adjusted gains vs. the S&P. OOS return ~161.6%, Calmar ~6.53; note higher drawdowns (~24.7% vs 5.1%) in downturns.