Statistical arbitrage · Pairs spread reversion
JPM/GS Pair Spread z2 · 1H
Pairs spread reversion on JPM, GS CFDs, evaluated on completed 1H bars and executed at live bid/ask.
Annual return+0.3%Backtest
Max drawdown−0.4%Volatility 2.1%
Sharpe1.13Trades 6
Since go-live—market closed
Backtest on real market data
Annual return+0.3%
Max drawdown−0.4%
Sharpe1.13
Trades6
Winning trades50%
Time in market27%
Backtest on real historical prices from Aug 10, 2026 to Oct 2, 2026: signals on completed bars, filled at the next bar's open, net of typical spread, commission and overnight financing. Past and backtested results are not a reliable indicator of future results. (Alpaca)
Methodology
- Trades the log-price spread of the first two instruments. When its 60-bar z-score exceeds +2, short the first and buy the second; below −2 the opposite.
- Both legs are closed when |z| falls below 0.5. The book is close to market-neutral.
- Trades both directions (long and short).
- Position size targets 8% annualised volatility (risk level 2/5), capped at 2:1 leverage per instrument. Costs: live bid/ask spread, commission and overnight financing.
Monthly returns · Backtest
| Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec | Σ | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2026 | 0.1 | 0.2 | 0.0 | 0.3% |
