Mean-Reversion & Z-Score Stat-Arb: Ornstein-Uhlenbeck & Time Stops

Capturing short-horizon snapbacks with Z-Score extremes, RSI(2), and mandatory N-bar time stops. Within the Algorithmic Trading curriculum (Systematic Strategy Design → Signal Architecture), this topic establishes the concrete quantitative rules and execution mechanics required for live market deployment.

Professional desks never rely on ambiguous visual guesses for mean reversion. Instead, every entry, invalidation level, and risk parameter is defined mathematically before a single order reaches the exchange gateway.

Hover over key terms below for instant definitions: Mean Reversion, Z-Score, and Rsi 2. For the complete mathematical derivation and edge-case analysis, open worked example & quantitative lookup matrix for mean reversion stat arb.

• Mean Reversion — Core structural driver governing mean-reversion & z-score stat-arb: ornstein-uhlenbeck & time stops.
• Z-Score — Prevents premature entries and enforces mechanical discipline across Signal Architecture.
• Rsi 2 — Ensures worst-case scenario losses stay strictly within predetermined R-multiple budgets.

1. 1
   1. Pre-Session Calibration & Regime Check

   Measure baseline volatility and liquidity parameters for mean reversion prior to market open.
2. 2
   2. Mechanical Signal Confirmation

   Wait for a completed bar close confirming the z-score threshold—never pre-empt mid-candle.
3. 3
   3. Risk-Normalized Sizing & Order Placement

   Compute exact lot quantity from the invalidation distance and dispatch via a Marketable Limit order.
4. 4
   4. Post-Trade Telemetry & Audit

   Log arrival price, fill slippage, and MAE/MFE metrics against the Signal Architecture benchmark.

Implementation & Quantitative Model

1export function zScoreSignal(price: number, mean20: number, std20: number, barsHeld: number) {2  const z = (price - mean20) / Math.max(1e-6, std20);3  if (barsHeld >= 8 || z >= 0) return "EXIT_LONG";4  if (z <= -2.2) return "ENTER_LONG";5  return "HOLD";6}

Cross-reference standard risk definitions such as drawdown, slippage, and Sharpe ratio when auditing your logs.