Asian session (23:00–07:00 UTC) on AUD / NZD / JPY crosses
Range-bound (low-ADX) regimes on H1
Quiet mid-week days between data clusters
sections.commonPitfalls
✗ Fading a genuine breakoutsections.pitfallFix Gate every entry behind a regime filter (ADX, band-width, session logic). Reversion signals inside an expansion are countertrend gambling, not mean reversion.
✗ Oversizing because the win rate is highsections.pitfallFix A 65% win rate with 1:1 R:R leaves thin margin — a size increase turns the normal losing cluster into real damage. Keep per-trade risk at scalping levels.
✗ Running it 24 hourssections.pitfallFix The statistical edge concentrates in specific sessions and regimes. Around-the-clock reversion trades the strategy's own noise.
✗ No stop because 'price always comes back'sections.pitfallFix It doesn't. Every reversion system needs a hard stop for the range-break case — that single assumption has ended more accounts than any indicator choice.
Mean reversion is the strategy class that pays for discipline most directly: a real statistical edge inside well-defined range regimes, and a reliable account-shredder the moment you trade it outside them. The craft is entirely in the regime filter — knowing when not to trade matters more than the entry logic. Our implementation is NightOwl AI, launched publicly in July 2026: Asian-session range trading on AUD/NZD/JPY crosses with session-scoped logic, single-position exposure and deterministic exits, running on the same engine discipline as the rest of the EASY line. It is our newest niche system and we present its forward window honestly rather than backfilling a legend. If you run reversion manually: pick range-prone crosses, define the regime gate mechanically, and let the strategy sit out trends without apology.
均值回归 — 常见问题
Which mean-reversion EA should I look at for MT5?
Our NightOwl AI trades Asian-session mean reversion on AUDJPY / NZDUSD / USDJPY-class crosses — session-scoped entries, hard stops, single position per symbol. Launched July 2026, so evaluate it the way we recommend evaluating anything new: on its accumulating forward results, not promises. Third-party alternatives: demand a verified 12-month track and check the trade list for hidden averaging.
What is the best session for mean-reversion trading?
The Asian session, by a wide margin. Between the NY close and London open, AUD / NZD / JPY crosses spend most nights oscillating in defined ranges without directional drivers — the statistical home ground of reversion entries. London and NY hours belong to breakout and trend classes.
RSI or Bollinger Bands for reversion signals?
Either works as the trigger; neither works as the system. The profitable component is the regime filter that decides whether reversion logic applies right now. An RSI-30 bounce in a range is an edge; the same signal in a trend is the top of a slide. Spend your optimisation budget on the filter, not the oscillator.
Why did my reversion EA blow up in a trending month?
Because that is the strategy's known failure mode: it shorts strength and buys weakness, which a persistent trend punishes repeatedly. Survivable implementations lose small and stand aside via regime detection; implementations that average into the trend convert the losing season into account loss.
Is mean reversion good for small accounts?
Reasonably — stops are modest (15–40 pips typical), trade frequency is decent, and cent accounts host it well. The high win rate also makes it psychologically sustainable for beginners, provided position sizing stays boring and the averaging temptation stays off.
Mean reversion vs scalping — aren't they the same?
They overlap on hold time but differ in thesis. Scalping monetises microstructure and spread dynamics at any hour with heavy infrastructure demands. Mean reversion monetises a statistical property of specific sessions and pairs, is far less latency-sensitive, and lives or dies by regime selection instead of execution speed.
Do reversion strategies pass prop firm challenges?
The profile fits well: high win rate, controlled per-trade risk, no exposure-stacking. The risk is the trending-week cluster of losses against a daily-loss cap — conservative sizing and the regime gate handle it. Verify the firm permits overnight Asian-session holds; most do.
How do I know a range is about to break?
You don't — you detect it fast instead of predicting it. Band-width expansion, ADX rising through ~25, and session-boundary volume shifts are the standard tells that flip a reversion system to stand-aside. The honest answer is exits and filters, not foresight.
自主研发 AI 交易系统
经过编辑审核、获得许可的 AI 智能交易系统,内部构建并支持。选择适合您资金和交易窗口的策略。
Scalperology AI
精选
Highest-frequency system in the line — trades directly on the AI signal across majors, crosses, metals and crypto, with cloud-optimised parameters and per-account adaptation. Raw-spread ECN required.
The same AI engine as Scalperology plus one hard rule: every signal must clear an institutional pivot level before it becomes a trade — structure confirms, the engine executes.
The AI engine gated by a rolling 200-tick momentum buffer — rides H4/D1 trends and holds positions while the move lasts. Lower-frequency, calmer volatility profile.
A trained XGBoost model exported to ONNX (16 engineered features) picks the active sub-strategy for the current gold regime, with an FOMC/CPI/NFP event guard on top.
Fades stretched overnight moves back to the mean in the window every London/NY bot sleeps through — uncorrelated flow that smooths a portfolio's equity curve.
A research-validated structural edge on 6 FX majors — timed holds, one position per pair, strict 1:2 reward-to-risk minimum. Order Blocks and FVGs are drawn as live context, honestly labelled telemetry.
Trades the market's weekly institutional rhythm with timed Monday-long / Friday-short entries and full-day holds. 108 'smarter' management variants were tested — every one reduced performance.
Trades only the minutes around Tier-1 macro releases, entering after the print once spreads re-normalize and the measured impulse clears cost-calibrated gates built from 13 years of event history.
Volume Profile (POC/VAH/VAL), anchored VWAP and a non-repainting CVD divergence detector, honestly built on tick activity — no footprint fantasy for spot FX. Analysis tool; places no trades.