TSMOM (Moskowitz, Ooi, Pedersen 2012) + regime + breadth. Refreshes every 15 min. Panels marked ENDPOINT NOT YET WIRED require a future /api/v1/live-signals endpoint.
NOW
REGIME CLASSIFIERSRC P5 engine
ENDPOINT NOT YET WIREDEndpoint not yet wired: /api/v1/live-signals. This panel requires a dedicated endpoint serving regime classifier / TSMOM signal / factor evidence output from the Python engine layer.
Regime classification is rule-based and point-in-time (no persistence filter). Composite signals: TREND direction, annualized vol, SPY drawdown depth, HY OAS spread. RISK_ON / NEUTRAL / RISK_OFF with a confidence score. Backed by a P5 vol-targeting backtest (CAGR 10.9%, Sharpe 1.03, Max DD -21.9% vs SPY -33.7%).
MARKET BREADTH | SECTOR ETF TRENDSRC FMP
TSMOM SIGNALS | 13-ETF UNIVERSESRC P6 live
ENDPOINT NOT YET WIREDEndpoint not yet wired: /api/v1/live-signals. This panel requires a dedicated endpoint serving regime classifier / TSMOM signal / factor evidence output from the Python engine layer.
Universe: 13 ETFs across Equity, Bond, Commodity, FX. Signal: 12-month return skipping the most recent month (Moskowitz 2012). LONG if 12-1 ret greater than 1%, SHORT if less than -1%, FLAT otherwise. Vol targeting: EWMA halflife=60, target_vol=15%. CAGR 5.87%, Sharpe 0.19 on the full 1993-2026 backtest. FX sleeve has historically negative cumulative contribution.
EVIDENCE
FACTOR EVIDENCE (BACKTEST)SRC P3 snapshot
ENDPOINT NOT YET WIREDEndpoint not yet wired: /api/v1/live-signals. This panel requires a dedicated endpoint serving regime classifier / TSMOM signal / factor evidence output from the Python engine layer.
Factors: Value, Momentum, Quality, Size, LowVol. Headline metric: ICIR (IC Information Ratio). Grinold-Kahn threshold: ICIR greater than 0.30 = a useful factor. Source: P3 factor-backtest-engine committed snapshot (503 tickers, 10-year backtest 2016-2026, 72 combinations). Best result: Quality factor Sharpe 1.55, IC 0.084. Provenance tag shown with data.
VOL REGIME GATESSRC FRED
TSMOM Signal: 12-month cumulative return, skipping the most recent month (12-1 momentum). LONG if return greater than 1%, SHORT if less than -1%, FLAT otherwise. Regime: composite classification from trend direction, volatility stress, drawdown depth, and credit conditions (HY OAS spread). Rule-based, not predictive. Breadth: percentage of S&P 500 sector ETFs showing positive momentum (full 200d MA requires /api/v1/live-signals). Limitations: signals computed from daily close prices, not intraday. TSMOM uses ETF proxies, not futures (no leverage cost or roll yield). FLAT threshold (1%) is configurable but not optimized. Regime classification is backward-looking.