Quantitative · Portfolio Construction

Why 20 Uncorrelated Alphas Beat One Great Signal Every Time

January 2026 · 7 min read · Atman Capital Research

There is a persistent myth in quantitative trading: that success comes from finding "the signal" — the single great alpha that consistently beats the market. The reality is more mathematical, and ultimately more reliable.

Information ratio scales with the square root of the number of independent signals. If you have N uncorrelated alphas each with IR₀, the combined information ratio is approximately IR₀ × √N. Twenty signals with IR 0.5 each combine to an IR of 2.24. One signal with IR 0.5 stays at 0.5.

Twenty signals with IR 0.5 each combine to an IR of 2.24. One signal at 0.5 stays at 0.5. The mathematics are unambiguous.

The Diversification Imperative

Our Statistical Arbitrage Strategy is built on this foundation. We run 20+ uncorrelated alphas, each derived from independent data sources: order book imbalances, funding rate dynamics, cross-exchange price differentials, options skew signals, and momentum factors across dozens of altcoin pairs. No single alpha is dominant. This is intentional.

Cross-Venue Diversification

Beyond signal diversification, we add a second layer: venue diversification. Binance, OKX and Bybit have distinct liquidity profiles and market microstructures. We measured the correlation between our Binance and OKX strategy returns: approximately 0.4. Combining both smooths the equity curve significantly — delivering the mixed portfolio's IR of 3.83.

Signal Decay and Renewal

Alpha decays. Every edge that becomes widely known gradually erodes as more participants exploit it. This is why the pipeline of new signal research is as important as the current alpha stack. Atman Capital maintains ongoing research into new signals, continuously adding promising alphas and retiring those showing decay. The 20+ signal count today reflects over two years of live operation and iteration.

Explore Our Strategies ← All Insights