New Strategy Nimbus: The N-Series Active Diversification System
Nimbus is the newest core pillar in our N-Series. It’s a late-generation implementation of our methodology, built around active diversification and advanced exposure management. What sets it apart is a transparent subsystem design: four fully independent, uncorrelated trading models with high-fidelity backtests (tick data since 2007) and 2+ years of live results that closely mirror the model.
At the core, Nimbus optimizes allocations using advanced correlation analysis, NSGA-II genetic optimization, and ensemble ML filtering. The system dynamically reweights across its four models to balance return, drawdown, and diversification—staying within a tight risk budget while remaining “always on.”
Why Nimbus matters now
Markets have been unforgiving for traditional beta. Nimbus is engineered to deliver equity-independent return streams with disciplined drawdown control, making it a compelling diversifier for portfolios that need consistent, liquid alt exposure in FX and commodities.
How Nimbus works (at a glance)
- 4 uncorrelated subsystems, each independently vetted and auditable
- Dynamic allocation via NSGA-II across return, drawdown, and diversification objectives
- Ensemble ML filters and a proprietary scoring framework that continually reweights for stability
- Operational profile: ~48h average hold, ~300 trades/month, 100% time in market
- Universe: Major FX pairs + liquid commodities
Live performance (Jul 2023 → Oct 2025)
- Total return: 151% (SPY 47% over the same window)
- CAGR (period): 49.6% vs SPY 18.3%
- Sharpe: 2.19 | Sortino: 3.75
- Max DD: –10.97% | Avg DD: –2.7% | 1-day VaR: ~2.2%
- Win months: ~82% | Worst month: –4.28%
- Correlation to SPY: ≈ –0.05 (near zero)
Why R² = 0.98 matters: We measure R² between live returns and the tick-data backtest over the overlapping period. 0.98 means 98% of live P&L variance is explained by the model—indicating high model fidelity, realistic cost/slippage assumptions, and low model drift. That’s foundational for trustable forecasting and compliant marketing.
Governance & assurance
- Backtests: Tick data since 2007 with explicit execution, costs, and slippage modeling
- Third-party audit: Alpha Performance Verification Services (APVS)
- Transparency: Independent subsystems with auditable rules and allocations
Capacity, funding & fit (important)
- Not compatible with the current x10 funding offering
- Technical capacity per license: ~USD 6M (scale via additional licenses)
- Typical account size: Viable from $20,000 (≈90% efficiency by $10k)
- Pricing: $12,500 per license