We combine validated factor research, cross-market liquidity analysis, and disciplined risk management to pursue asymmetric equity opportunities before they become consensus.
The market prices equities on five orthogonal forces. Most funds bet on one. We measure all five and let the data tell us what matters at each moment.
Weights are adaptive — the composite regenerates every 42 days. No single factor dominates across regimes.
Out-of-sample returns per 42-day rebalance window. Q5 = top quintile (buy). Q1 = bottom quintile (avoid). 2020–2026.
We do not backtest. Every result is out-of-sample. The model trains on the past, scores the present blind, and lives through the future. Every 42 days, it proves itself again.
252-day rolling window of factor exposures and forward returns. Ridge regression with L2 regularization.
Apply learned weights to current factor values. Rank every stock. The model has never seen this data.
Survive 42 trading days. Record actual returns. Feed results into the next training window. Repeat forever.
| Parameter | Configuration |
|---|---|
| Regression | Ridge (L2-regularized), λ = 0.05 |
| Training window | 252 trading days, rolling |
| Rebalance cycle | 42 trading days |
| Factor normalization | Cross-sectional Z-scores per step |
| Return winsorization | ±50% cap |
| Minimum breadth | 30 stocks per cross-section |
S&P 500 quintile spreads average 3–5%. In China, we observe 15.8%. Fewer analysts per stock, more retail flow, more policy-driven mispricing. The opportunity set is structurally wider.
Most global funds apply US-calibrated factor models to Asia. Valuation is inverted here — expensive beats cheap. Flow is contrarian in China. Liquidity dominates both markets. These signals persist because consensus ignores them.
Stock Connect, A-share inclusion, and HKEX reforms are adding hundreds of investable names. The opportunity set is growing faster than the capital chasing it.
Every strategy faces regimes where its edge disappears. We design for those moments explicitly.
| Risk | Mitigation |
|---|---|
| Factor regime shift | Five-factor composite. When one factor decays, others compensate. Weight autocorrelation is positive — regimes are sticky enough to trade. |
| Liquidity crisis | F5 filter eliminates illiquid names. Minimum market cap $1B. VIX overlay reduces exposure during stress events. |
| Single-name blowup | Equal-weighted within Q5. No position exceeds 3% at entry. Maximum 39 positions (Japan), ~114 (China). |
| Turnover costs | 42-day rebalance, ~25–40% turnover per cycle. At 20bps per trade, cost drag under 1% annually. |
| Capacity constraints | Hard cap at $200M AUM. Beyond this limit, top-quintile entry and exit would move prices in mid-cap A-shares. |
| Governance failure | Rolling 12-month IC below 0.05 for two consecutive quarters triggers an automatic strategy halt and review. No exceptions. |
This strategy is designed for allocators seeking systematic, transparent, and uncorrelated Asia equity exposure.
Asian factor correlations to US equities are structurally lower. The same 5-factor framework produces different dominant signals in each market — fundamentals lead in China, flow and liquidity in Japan.
Every factor is defined. Every weight is traceable. Every trade is explainable. No black boxes. No proprietary secret sauce hidden behind a curtain.
We're seeking seed capital to launch the strategy live. Managers co-invest alongside LPs. 12–18 months to build institutional track record.