🤖 AI Regression Zones — The cBot that knows WHEN to trade and WHEN to stop

📊 What is AI Regression Zones?

AI Regression Zones is a next-generation automated trading system. It combines the power of linear regression with built-in artificial intelligence that classifies the market in real time. This bot doesn't just look for entries: it first analyzes whether market conditions are favorable, then decides if and how to trade.

The core strategy is mean-reversion: the bot calculates an equilibrium line on daily closes and builds dynamic entry zones around it. When price moves too far from the mean, it enters a position expecting a return. But unlike classic mean-reversion bots, AI Regression Zones has a brain: it recognizes when the market is trending and does NOT open against it.

🧠 The intelligence that makes the difference

Most mean-reversion bots lose money because they trade blindly in any condition. AI Regression Zones solves this with 5 layers of intelligence working together.

🔍 AI REGIME FILTER — The most important module. Automatically classifies the market as Ranging, Trending or Choppy by analyzing ADX and regression slope. In trending markets it blocks entries that would lose. In choppy markets it stops completely. It only operates when conditions favor the strategy. Can optionally connect to an external AI server for advanced machine learning classification.

💰 INTELLIGENT SIZING — Forget fixed lots. AI Regression Zones automatically calculates position size based on your equity and current volatility. You always risk the same percentage per trade, whether the market is calm or explosive. Drawdown becomes predictable and controllable.

📈 ADAPTIVE VOLATILITY — Entry zones, stop loss and take profit automatically adapt to real market volatility using average or median ATR over 14 days. No more stops that are too tight in volatile markets or too wide in calm ones. Everything calibrates itself.

🔎 MULTI-TIMEFRAME CONFIRMATION — Before every entry, the bot checks RSI on H4 to confirm timing. A BUY is only executed if H4 confirms oversold, a SELL only if it confirms overbought. Drastically reduces premature entries.

🤖 EXTERNAL AI INTEGRATION — For advanced traders: connect a machine learning regime classification server. The bot sends market data, receives the classification and uses it to filter trades. Fully optional: the bot works perfectly without it.

🛡️ 360° Capital Protection

AI Regression Zones isn't just smart at entering. It's relentless at protecting.

⛔ Max daily loss percentage — once the threshold is hit, the bot stops until the next day. ⛔ Max total loss from initial equity — protection against cumulative drawdown. ⛔ Max drawdown from peak equity — if equity drops too far from its high, everything locks. ⛔ Consecutive stop loss lock — after 3 losing SLs in a row (configurable), the bot stops. No tilt. 🔒 Automatic break-even — when the position goes into profit, SL moves to entry price. 📉 Trailing stop — follows price protecting profits while letting the trade run. 📶 Protected pyramiding — adds positions only when the first one is in profit with SL already at break-even.

⚙️ Full Parameter Guide

🏷️ 01) BASE — Core Settings

📌 Label — Unique tag assigned to every position opened by this bot. If you run multiple instances on the same symbol, use different labels to keep them separate. Default: RegOscZones.

📌 Trading Enabled — Global on/off switch. When set to false the bot keeps calculating levels, managing trailing and monitoring risk, but opens no new positions. Useful during high-impact news events. Default: true.

📌 Entry Mode — Defines how the bot enters trades. Three options available. ZoneEntry opens a position as soon as price touches the buy or sell zone. RegressionCrossClose requires a full daily cross-close on the regression line. CrossCloseWithZoneFilter combines both: cross-close signal plus price must be inside the zone. CrossCloseWithZoneFilter is the most conservative and recommended. Default: RegressionCrossClose.

📌 Sizing Mode — Chooses how position volume is calculated. FixedLots uses the static lot value you set. RiskPercent automatically calculates lots as Equity × Risk% divided by SL pips × pip value per lot. RiskPercent is strongly recommended for professional risk management. Default: RiskPercent.

📌 Volume in Lots — Fixed lot size used only when Sizing Mode is set to FixedLots. One standard lot equals 100,000 units. 0.01 is a micro lot of 1,000 units. Default: 0.01.

📌 Risk Per Trade % — Percentage of equity risked on each trade. Used only with RiskPercent sizing. The bot automatically adjusts lot size based on your stop loss distance and account equity. 0.5-1% is conservative, 1-2% moderate, above 2% aggressive. Default: 1.0%.

📌 Max Lots Cap — Maximum lot size per single trade regardless of the risk calculation. Protection against oversized positions on large accounts or very tight stops. Default: 5.0.

📐 02) REGRESSION — Linear Regression Settings

📌 Regression Period — Number of daily candles (D1) used to compute the linear regression. Longer periods produce a smoother line that captures the underlying trend. Shorter periods react faster but introduce more noise. 15-25 recommended for forex majors, 30-50 for indices. Default: 20 days.

📌 Forecast Bars — Number of days to project the regression line into the future. This shifts the reference point (LR Forecast) used to build zones. 0 means use the last calculated value, 1 projects one day ahead. Values above 3 become speculative. Default: 1.

✋ 03) ENTRY LOGIC — Cross and Touch Settings

📌 Touch Tolerance (pips) — Margin around the regression line that counts as a "touch". If price gets within this distance of the LR, it is treated as if it touched it. Prevents false negatives from candles that miss the line by a few pips. 3-7 pips for forex majors, 10-20 for indices. Default: 5.

📌 Confirm Close (pips) — Minimum distance beyond the LR that the daily close must reach in the opposite direction to confirm the cross. For a BUY the close must be at or above LR plus this value. Filters out indecisive candles. Higher values mean fewer but more reliable signals. Default: 10.

📌 Require Open Opposite Side — When enabled, the daily candle open must be on the opposite side of the LR from the trade direction. For a BUY the open must be below the LR. Adds an extra layer of confirmation for a complete cross. Default: false.

🟩🟥 04) ZONES — Oscillation and Trading Zones

📌 Oscillation Lookback — Number of daily candles used to calculate the oscillation value (ATR, range, etc.). Longer lookbacks produce more stable values but react slower to recent volatility changes. 10-20 for stability, 5-7 for reactivity. Default: 14 days.

📌 Oscillation Mode — How the oscillation (base volatility measure) is calculated. Five options available. RangeHighLow takes the maximum High-Low range in the lookback period. TrueRange takes the maximum True Range. MaxDeviationFromRegression measures the maximum deviation of closes from the regression line. ATR_Average computes the average True Range over the lookback and is the recommended default for stability. ATR_Median uses the median True Range and is the most robust against outlier days. Default: ATR_Average.

📌 Outer Mult — Multiplier for the outer edge of the zones in symmetric mode. The BUY zone extends from LR minus Osc×OuterMult to LR minus Osc×InnerMult. Higher values create wider zones requiring bigger moves before entry. 0.8-1.2 is typical. Default: 1.0.

📌 Inner Mult — Multiplier for the inner edge of the zones in symmetric mode. Defines where the zone ends closest to the LR. Lower values create a wider zone, higher values create a thinner band near the outer edge. Default: 0.25.

📌 Asymmetric Zones Enabled — When enabled, allows separate multipliers for the BUY zone and the SELL zone. Useful on instruments with directional skew, for example equity indices that tend to drift upward. Only activate after statistical analysis. Default: false.

📌 BUY Outer/Inner Mult — Specific multipliers for the BUY zone only. Active when Asymmetric Zones is enabled. Tighten the BUY zone on bullish instruments. Default: 1.0 / 0.25.

📌 SELL Outer/Inner Mult — Specific multipliers for the SELL zone only. Active when Asymmetric Zones is enabled. A wider SELL zone compensates for bullish drift. Default: 1.1 / 0.25.

📌 Allow Beyond Outer — When enabled, the bot accepts entries even if price is beyond the outer edge of the zone, further away from the LR. When disabled, price must be strictly inside the Inner-Outer band. Set to false for pure mean-reversion, true to catch extreme spikes. Default: false.

🌡️ 04B) REGIME — Market Regime Filter

📌 Regime Filter Enabled — Master switch for the regime classification system. When disabled the bot operates in any market condition like a basic mean-reversion bot. Keep this enabled at all times in production. It is the primary protection against trend losses. Default: true.

📌 Regime Source — Where the regime classification comes from. Internal uses ADX plus regression slope calculated inside the bot. AI_API uses the result from an external machine learning server. Both uses the AI server when available and confidence is sufficient, otherwise falls back to Internal. Default: Internal.

📌 ADX Period — Period for Wilder's Average Directional Index calculated on D1 data. ADX measures trend strength regardless of direction. 14 is the industry standard. 10 for more reactivity. Default: 14.

📌 ADX Trending Threshold — Above this value the market is classified as Trending. Classic values range from 20 to 30. An ADX at 25 or higher indicates a well-defined trend where mean-reversion is dangerous. 20-25 for forex, 25-30 for indices. Default: 25.

📌 ADX Choppy Threshold — Below this value the market is classified as Choppy, meaning noise with no direction. Low ADX combined with low slope indicates no tradable structure. 12-18 depending on instrument. Default: 15.

📌 Slope Trending Threshold (pips/day) — Regression slope threshold in pips per day for classifying the market as trending. Even with low ADX, a strong slope indicates a trend. This is an OR criterion together with ADX. 2-4 for forex majors, 5-10 for indices or crypto. Default: 3.0.

📌 Block MeanRev in Trend — When enabled, blocks all mean-reversion entries when the regime is Trending. This is the key protection: it prevents opening positions against a strong trend. Never disable in production. Default: true.

📌 Block Trading in Choppy — When enabled, blocks all trades when the regime is Choppy. In directionless markets mean-reversion has low edge and high spread costs. Recommended true. Disable only on very low spread instruments. Default: true.

🔎 04C) MTF — Multi-Timeframe Confirmation

📌 MTF Confirmation Enabled — Activates the multi-timeframe filter. When disabled, entries depend solely on D1 logic plus regime. Keep enabled to reduce false signals. Default: true.

📌 MTF Timeframe — Timeframe for the confirmation RSI calculation. Accepted values: m1, m5, m15, m30, h1, h4, h8, h12, d1, w1. H4 offers a good balance between noise and timing for D1 swing trading. H1 for more aggressive scalping. Default: Hour4.

📌 MTF RSI Period — RSI period calculated on the MTF timeframe. 14 is the Wilder standard. 9-10 for more reactivity. Default: 14.

📌 MTF RSI Oversold — RSI threshold below which a BUY is allowed. With mean-reversion we want to buy when the MTF RSI confirms an oversold condition. Higher value is more permissive and allows more trades. 30 for maximum selectivity, 45 for more trades. Default: 40.

📌 MTF RSI Overbought — RSI threshold above which a SELL is allowed. We want to sell when RSI confirms overbought. Lower value is more permissive. 60-70 recommended, 70 for maximum selectivity. Default: 60.

📌 MTF Structure Lookback — Number of MTF bars for market structure analysis of higher-highs and higher-lows. The lookback is split in two halves and pivot points are compared. 15-25 captures meaningful structure. Default: 20.

📌 MTF Require Structure Confirm — When enabled, besides RSI also requires structure confirmation. For a BUY the structure must not be bearish (lower-highs and lower-lows). For a SELL the structure must not be bullish. Can significantly reduce the number of trades. Default: false.

🤖 04D) AI — Artificial Intelligence Integration

📌 AI Enabled — Activates HTTP calls to the AI endpoint. When false, no network requests are made and the regime is determined only by the internal ADX plus slope calculation. The bot is fully functional without AI. Default: false.

📌 AI Endpoint URL — Full URL of the endpoint that receives market data via POST and returns a JSON response with regime, confidence and bias. Uses cTrader's native Http API with AccessRights.None. Default: http://127.0.0.1:8000/regime.

📌 AI Timeout (ms) — Maximum time to wait for the HTTP response. If the server does not respond within this time, the bot ignores the AI for that cycle and uses the internal fallback. 2000-5000 typical, 1000 on co-located VPS. Default: 3000.

📌 AI Refresh Interval (min) — How often the bot queries the AI endpoint. Since the regime changes slowly on a D1 basis, there is no need to update too frequently. 30-120 minutes recommended. Default: 60.

📌 AI Override Confidence Threshold — Minimum confidence returned by the AI to accept its classification. Below this threshold the bot uses the internally calculated regime. Protects against uncertain predictions. 0.65-0.80 recommended. Default: 0.7.

🚫 05) LIMITS — Filters and Operating Limits

📌 Trade BUY — Enables or disables opening long positions. Useful to force the bot to operate in only one direction on instruments with strong directional bias. Default: true.

📌 Trade SELL — Enables or disables opening short positions. Complementary to Trade BUY. Disable on equity indices during stable bull markets. Default: true.

📌 Max Positions TOTAL — Maximum number of positions open at the same time (long plus short) for this bot instance on this symbol. 1-3 for conservative, 4-5 with pyramiding active. Default: 2.

📌 Max Positions LONG — Maximum simultaneous long positions. Includes the base position plus any pyramid additions. 1 without pyramiding, 2-3 with pyramiding. Default: 1.

📌 Max Positions SHORT — Maximum simultaneous short positions. Same logic as long. Default: 1.

📌 Max Trades Per Day — Maximum number of new trades (openings) per day. Resets at midnight UTC. Protects against overtrading during volatile sessions. 2-4 for swing, 5-10 for aggressive intraday. Default: 2.

📌 Max Spread (pips) — Maximum acceptable spread for opening a position. If the current spread exceeds this value the bot will not trade. Protects against slippage and illiquid hours. 1.5-2.5 for EUR/USD, 3-5 for exotic crosses. Default: 2.1.

📶 06) PYRAMIDING — Adding Positions in Profit

📌 Pyramiding Enabled — Activates the pyramid adding system. When disabled the bot only opens the base position. Activate only on instruments with good intraday follow-through. Default: true.

📌 Max Pyramid Adds LONG — Maximum additional long positions. With 2 adds and one base you can have up to 3 long positions. 1-2 for conservative, 3 or more for aggressive. Default: 2.

📌 Max Pyramid Adds SHORT — Same as above for short positions. Default: 2.

📌 Pyramid Breakout (pips) — Price must move at least this many pips beyond the last entry price in the favorable direction to "arm" the pyramid. Confirms the trade is working before adding. 20-40 for forex, 50-80 for indices. Default: 30.

📌 Pyramid Retrace (pips) — After the breakout, price must retrace this many pips from the peak or trough to trigger the addition. The retrace is the optimal entry on continuation. Half of the breakout value is a good starting point. Default: 15.

📌 Min Profit for Pyramid (pips) — The base position must be in profit by at least this many pips before pyramiding activates. Protection against premature additions. Never set to 0 in production. Default: 10.

📌 Require SL for Pyramid — The base position must have a stop loss set to allow additions. Without a defined SL the risk is undefined and adding capital is imprudent. Always keep true. Default: true.

📌 Require SL at BE or Better — The base position stop loss must be at break-even (entry price) or better. Ensures the base position is essentially "free" before adding risk. Always keep true. Default: true.

📌 Pyramid Min Locked (pips) — Minimum pips of profit locked by the base position stop loss. For example with 5, the SL must be at least at entry plus 5 pips. More restrictive than simple break-even. 0-10 recommended. Default: 0.

🎯 07) STOPS — Stop Loss and Take Profit

📌 Stop Mode — Defines how SL and TP are calculated. Three options. None means no SL or TP which is very risky. FixedPips uses static pip values configured separately for long and short. OscMultiplier calculates SL and TP as multiples of the current oscillation value, adapting automatically to volatility. OscMultiplier is strongly recommended. Default: OscMultiplier.

📌 LONG SL / TP (pips) — Fixed stop loss and take profit for long positions in pips. Used only with FixedPips mode. Setting 0 disables that specific level. Always maintain a TP/SL ratio of at least 1.5. Default: 200 / 300.

📌 SHORT SL / TP (pips) — Fixed stop loss and take profit for short positions. Allows asymmetry between long and short if the instrument behaves differently in each direction. Default: 200 / 300.

📌 LONG SL Mult (x Osc) — Oscillation multiplier for long stop loss. SL in pips equals Osc times Mult divided by PipSize. With Osc of 100 pips and Mult of 1.0, the SL is 100 pips. 0.8-1.2 for standard, below 0.5 is tight, above 1.5 is wide. Default: 1.0.

📌 LONG TP Mult (x Osc) — Multiplier for long take profit. The TP/SL ratio equals TP Mult divided by SL Mult. With defaults of 1.5 and 1.0 you get a reward-to-risk of 1.5 to 1. Keep TP Mult higher than SL Mult. Default: 1.5.

📌 SHORT SL Mult (x Osc) — Oscillation multiplier for short stop loss. Can differ from the long side to compensate for market asymmetries. Default: 1.0.

📌 SHORT TP Mult (x Osc) — Multiplier for short take profit. Setting 0.0 means no TP and the trailing stop will manage the exit. 1.5-2.0 recommended, 0 only with trailing active. Default: 1.5.

🔒 08) BREAK EVEN — Moving SL to Entry Price

📌 BE LONG Enabled — Activates automatic break-even for long positions. Always keep true in production. Default: true.

📌 BE LONG Trigger (pips) — Profit in pips the long position must reach before the SL is moved to break-even. Too low triggers premature stops on noise, too high delays protection. 20-40 for forex, 40-80 for indices. Default: 30.

📌 BE LONG Offset (pips) — Pips added to entry price when moving SL to break-even. Covers spread and commissions. With offset 2 the SL goes to entry price plus 2 pips, guaranteeing a small profit if hit. 1-3 for forex, 3-5 for indices. Default: 2.

📌 BE SHORT Enabled — Same as BE LONG but for short positions. Always keep true. Default: true.

📌 BE SHORT Trigger (pips) — Trigger for short break-even. Same criteria as long. Default: 30.

📌 BE SHORT Offset (pips) — Offset for short BE. SL goes to entry price minus offset. Default: 2.

📉 09) TRAILING — Trailing Stop

📌 TRAIL LONG Enabled — Activates trailing stop for long positions. Always enable if you want to capture extended moves. Default: true.

📌 TRAIL LONG Trigger (pips) — Profit in pips that activates the trailing. Should be equal to or greater than the BE Trigger to avoid conflicts. Default: 40.

📌 TRAIL LONG Distance (pips) — Distance between current price and the trailing SL. The SL is positioned at Bid minus Distance. Tighter distances protect more profit but risk premature stops. 15-30 for forex, 30-60 for indices. Default: 25.

📌 TRAIL LONG Step (pips) — Minimum SL movement per update. Prevents continuous micro-adjustments. The SL only moves if the new level is at least Step pips above the current one. 1-3 typical, 0 for continuous updates. Default: 2.

📌 TRAIL SHORT Enabled — Trailing stop for short positions. Default: true.

📌 TRAIL SHORT Trigger (pips) — Trigger for short trailing. Default: 40.

📌 TRAIL SHORT Distance (pips) — Trailing distance for short. SL at Ask plus Distance. Default: 25.

📌 TRAIL SHORT Step (pips) — Minimum step for short trailing. Default: 2.

⛔ 10) RISK — Risk Control

📌 Risk Control Enabled — Enables all risk protections. When disabled no automatic limits are applied, though individual SLs remain active. Never disable in production. Default: true.

📌 Max Daily Loss % — Maximum daily loss as a percentage of equity at the start of the day. If the loss reaches this threshold the bot stops trading until the next day. 3-5% for conservative, 5-8% for moderate. Default: 5%.

📌 Max Overall Loss % — Maximum total loss as a percentage of the initial equity when the bot started. Protection against cumulative drawdown. 8-15% recommended, beyond 20% recovery becomes very difficult. Default: 10%.

📌 Max Equity DD % (from Peak) — Maximum drawdown from the peak equity. If equity drops by this percentage from its all-time high during the bot session, trading locks. Catches losses even if equity had grown before. 8-15% recommended. Default: 10%.

📌 Risk Breach Action — What happens when a risk limit is hit. StopNewTrades blocks new openings but keeps existing positions. CloseAllAndStop closes all positions and blocks new openings. CloseAllAndStop provides maximum protection. Default: CloseAllAndStop.

📌 Risk Lock Reset Mode — How the risk lock is released. Never means it stays locked until manual restart. NextDayDailyLossOnly unlocks the next day only if the cause was the daily loss limit. NextDayAnyBreach unlocks the next day regardless of cause. NextDayDailyLossOnly is the best compromise. Default: NextDayDailyLossOnly.

📌 Reset Peak Equity On Unlock — When the risk lock is removed, the peak equity is reset to current equity. Prevents the old peak from causing an immediate re-lock. Default: true.

📌 SL Streak Lock Enabled — After a configurable number of consecutive losing stop losses, trading locks until the next day. Protects against losing streaks and emotional tilt. Always keep true. Default: true.

📌 Max Consecutive SL — Number of consecutive losing stop losses before the lock triggers. The counter resets with any winning trade. 2 for maximum caution, 4 for more tolerance. Default: 3.

📌 Count StopOut as SL — When enabled, StopOuts (forced closures by the broker due to insufficient margin) also count as stop losses in the consecutive streak. Always keep true. Default: true.

📌 Close All on SL Lock — When the SL streak lock triggers, also closes all remaining open positions. When false, existing positions stay open with their SLs and trailing. Default: false.

📊 11) CHART — Visual Display

📌 Draw Zone Rectangles — Draws semi-transparent colored rectangles for the BUY zone (green) and SELL zone (red) on the chart. Helps visualize where the bot is looking for entries. Default: true.

📌 Zone Rect Alpha — Transparency of zone rectangles. 0 is fully transparent, 255 is opaque. 40-80 works well for most setups. Default: 60.

📌 Draw Regime Label — Shows a label in the top-left corner of the chart with the current regime, ADX value, slope, MTF RSI. When AI is active it also shows the AI regime and confidence. Default: true.

🐛 12) DEBUG — Diagnostics and Logging

📌 Debug Enabled — Enables debug messages in the cTrader log. When false only critical messages such as executions and errors are printed. Use true during testing, false or Level 1 in production. Default: true.

📌 Debug Level — Verbosity level. Level 1 shows only important events like start, stop, trades, risk locks. Level 2 adds filters, zones, blocked signals. Level 3 adds tick-by-tick evaluation, every zone check, every trailing adjustment. Default: 2.

📌 Debug Throttle (sec) — Minimum interval between debug messages of the same type. Prevents flooding the log with repetitive messages every tick. 0 disables throttling and prints everything. 10-30 during testing, 60 or more in production. Default: 10.

📌 Print Zones On Recalc — When enabled, prints all calculated levels (LR, OSC, BUY/SELL zones, ADX, slope, regime, RSI) every time daily data is recalculated. True for verifying calculations, false in production. Default: true.

🤖 Custom AI Server — Available on Request

Want to unlock the full power of AI Regression Zones? We develop a custom Python AI server tailored to your needs. The server uses a Random Forest machine learning model trained on 12 market features including ADX, Hurst exponent, autocorrelation, volatility ratio and price structure. It classifies the market regime in real time and sends the result directly to the cBot via HTTP.

What you get: a ready-to-run FastAPI server with regime classification endpoint, a training endpoint to retrain the model with your own labeled data, and full documentation. The server runs on any VPS or local machine and communicates with the cBot using cTrader's native Http API — no FullAccess required.

📩 Contact us to discuss your setup and get a custom AI server built for your trading style and instruments.

✅ Technical Specifications

AccessRights.None — 100% cTrader Store compatible, zero security risks. Zero dependencies — ADX, RSI, ATR, regression, market structure all calculated internally. Multi-instance — separate labels for parallel execution on multiple symbols. Native networking — AI calls via cTrader's Http.Send(), no workarounds. Full logging — every decision documented in the log with emoji for quick identification.

⚠️ Disclaimer

AI Regression Zones is an automated trading tool. Past performance does not guarantee future results. We recommend backtesting on at least 2-3 years of data and starting with a demo account. Always configure Risk Control before going live. Trading involves significant risk of capital loss.