Mean Reversion Trading Strategy for ES and NQ Futures: Setup, Entry, and Management

Mean reversion describes the statistical tendency of price to return toward its mean after an extended move away from it. In the context of ES and NQ futures, the mean is defined by reference levels — VWAP, moving averages, or statistical bands — and an extended move is characterized by price being significantly above or below that mean on above-average volume. When price is two or more standard deviations from VWAP, or when it is far outside a Bollinger Band, or when it has moved 150%+ of the prior day's ATR in a single session, mean reversion probability is elevated. Mean reversion is not universal — it is regime-dependent. The most important rule in mean reversion trading is that it fails systematically in trending markets. A market that is in a strong institutional uptrend (daily chart showing consistent higher highs and higher lows over 20+ sessions) will extend far from its mean before reverting, or may not revert at all during the active trend. Applying mean reversion logic in a strong trend produces repeated losses at new highs as the trader fades the trend expecting a return to the mean that never comes on the current timeframe. The Marty strategy, which uses mean reversion logic refined over 6 years of live trading, has survived because its regime filter classifies market conditions before entry and only activates in the range-bound, grinding conditions where mean reversion has documented edge. Building a mean reversion setup requires four components. Component one: the mean reference level. VWAP is the primary intraday mean reference — it represents the volume-weighted average price for the session and is the level at which position-neutral participants can execute without directional bias. Additional mean references include the session midpoint (midpoint of the day's range so far), the 20-period EMA on the 5-minute chart, and the opening price. When price is extended from multiple mean references simultaneously, mean reversion probability is highest. Component two: the extension measurement. Standard deviation bands on VWAP (1SD, 2SD, 3SD) provide statistical quantification of how far price has moved from the mean. Price at the +2SD band above VWAP has moved approximately 2 standard deviations from the session's volume-weighted center — a statistically unusual extension that, in non-trending conditions, has high historical probability of returning toward VWAP. The 2SD band is the primary mean reversion entry zone. The 3SD band is a secondary, more extreme entry zone used for larger extensions. Component three: the regime filter. The most important filter is whether the session is trending or rotating. Pre-market session type classification (described in the trend day vs. range day framework) provides the primary filter. Real-time confirmation comes from VWAP behavior: is price holding above VWAP and making higher lows on 5-minute timeframe (uptrend)? Or is price oscillating above and below VWAP multiple times (range)? Mean reversion entries are only taken in confirmed range conditions. In trending sessions, extending moves away from VWAP represent continuation opportunities, not reversion opportunities. The Marty strategy's regime classification is the automated version of this filter — it reads market structure conditions and only places mean reversion entries when the regime is confirmed as range-bound. Component four: the confirmation trigger. Price reaching the +2SD band is a zone to watch, not an automatic entry signal. The confirmation trigger — the signal that the extension has peaked and price is beginning to revert — must also be present before entry. Three confirmation triggers are acceptable: a candlestick reversal pattern (shooting star, bearish engulfing) on the timeframe being traded, a structural break on a lower timeframe (first lower high on the 2-minute chart for a short entry at the +2SD band), or a volume divergence (price makes a new extension high but volume is decreasing, indicating exhaustion of the directional move). The confirmation trigger transforms the mean reversion level from a zone into a specific, time-defined entry point. Trade management for mean reversion setups follows a different structure than trend-following trades. The target is the mean (VWAP or session midpoint), not an extended continuation. Position sizing must account for the fact that mean reversion entries are often made against short-term momentum — the initial period after entry frequently shows the trade moving further against before reversing. This characteristic means that mean reversion trades require wider stops than trend-following trades of equivalent risk-reward. The stop placement sits beyond the extension level — for a short entry at the +2SD band, the stop goes above the +3SD band, allowing room for the extension to push slightly further before the reversion begins. The YMI Marty strategy automates the entire mean reversion framework: regime classification, extension detection, confirmation trigger identification, entry, and stop/target management. The 6-year live trading history with zero losing days reflects the combination of a sound mean reversion edge with a robust regime filter — Marty does not trade in trending conditions where mean reversion fails. For traders who want to understand and implement mean reversion manually, the framework above provides the structural foundation. For traders who want institutional-quality mean reversion execution without managing each element manually, the Marty bot implements these principles algorithmically across the sessions where the edge is proven.