Technical Deep Dive: How We Validate Macro Factors

[michaeljiangmingfeng-debug]

Every strategy starts with a macro factor question — not "what pattern maximizes returns?" but "what is the real macro factor, and is it persistent?"

Factor Validation Approach

CNH-CHAU (Offshore RMB Long/Short): Does offshore RMB trading reflect real cross-border capital flows?

• Validation: Track CNH vs. CNY spreads against actual cross-border transaction data, FDI flows, capital account data
• • The factor is constant: global capital votes on China's economy through offshore RMB pricing
    DIP-A (Dip Signal A-shares): Is limit-down an accurate panic gauge?
• Validation: Compare limit-down frequencies against retail positioning data, margin liquidation events
• • China's limit-down mechanism is globally unique — it creates a real-time panic thermometer that doesn't exist elsewhere
    IF-IC (Large/Small Cap Rotation): Can capital controls really create two independent fund channels?
• Validation: Measure correlation between IF (large-cap) and IC (small-cap) during regime shifts — they decorrelate when capital controls tighten
• • This is a structural feature of China's financial system, not a statistical artifact
    DIP-US (Dip Signal US): Does VIX spiking to extremes create mean-reversion edges?
• Validation: Measure realized volatility post-spike, compare to implied volatility expectations
• • 21 trades, 100% win rate since 2008 — the edge comes from behavioral panic, not curve fitting

Our Philosophy

We don't chase dead statistical benchmarks. We find macro factors that are:

1. Grounded in economics — not data mining artifacts
2. 2. Persistent — the underlying mechanism doesn't expire
3. 3. Transparent — fully auditable at AI Hall
4. 4. Simple — fewer parameters means less overfitting
      The beauty of a strategy lies in its honest expression of a market truth. Overfitting destroys that beauty.

Try the Strategy Combiner: https://www.quanttogo.com/playground (30-day free trial)