Neutral Momentum family — side-by-side

All 7 neutral_mom_* variants paper-tracking forward. The extended-window backtest metrics (Sh / CAGR / DD / β) are the HONEST forward expectations from the 2022-26 walk-forward, not the in-sample dashboard numbers. Equity curves overlaid below are normalized to 100 at each variant's first paper-tracked day so they can be compared on the same scale regardless of position sizing.

Live paper-track + extended-window stats

VariantStatusExt. ShExt. CAGRExt. DDExt. βLive retOpen posTracked
Neutral Mom (K=10, original)
Original automated K=10 3-factor L/S. Longest paper-tracking history. Regime-resilient (K=10 dilution helps in 2022-23 bears).
active0.87+7%−11%≈0-0.01%38021d
Neutral Mom 5F (K=3, 5-factor)
K=3 with vol_z + stealth_accumulation added to the 3-factor base. Higher in-sample Sharpe; weaker extended-window.
superseded0.75+8%−30%+1.67-0.01%11421d
Neutral Mom E (MEGA — breakthrough)
6-sleeve equal-weight (equity + TLT + GLD + USO + EFA + sector-rot) + cond hedge + 12% vol-target. THE breakthrough.
champion3.00+45%−5%≈0+0.09%27221d
Neutral Mom ETF (5-sleeve simplified)
Drops the equity stock-picking sleeve (contributes least per leave-one-out). 5-sleeve ETF-only. ~7 positions vs ~20.
champion2.78+41%−5%≈0-0.00%621d

Live equity curves overlaid (normalized to 100 at each variant's first day)

Y-axis = relative equity (start = 100). Curves diverge as variants accumulate paper-tracking data. Early on, gaps between variants mostly reflect different start dates and small sample noise — wait ~30+ trading days before drawing forward-conclusions.

What to watch as data accumulates

  • Path E vs Path ETF: do they track closely (confirming the equity sleeve adds little) or diverge (revealing equity's contribution in live conditions)?
  • Champions vs superseded: do E/ETF maintain their backtest advantage over A/B/5F/K2 in live data, or does the gap shrink?
  • Original K=10 neutral_mom: does its regime-resilient profile show in stress periods (compare DDs when SPY drops)?
  • Path A (max-Sharpe): does the high backtest Sharpe (2.11) hold up, or does the +1.3 β profile cause meaningful drag?
  • Forward Sh vs backtest: expect a 10-20% haircut from backtest to forward due to in-sample optimism. Path E should land near Sh 2.5-2.9 forward if real.