Pipsgrowth EX12017 MultiIndicatorConfluence

Pipsgrowth EX12017 reads market microstructure through a single composite index it constructs on the fly — the Adaptive Market Liquidity Index, or AMLI. The indicator does not live in a chart subwindow; it is computed in the EA's own CalculateAMLI() function by combining a smoothed tick-volume ratio with an ATR ratio into a 0-100 oscillator. The reading sits at 50.0 (neutral) by default, drifts above 50 when both volume and volatility expand in lockstep, and falls below 50 when the market thins out. That sub-50 zone is the liquidity trap the EA is designed to trade: when order flow dries up and price stops moving meaningfully, AMLI collapses, and a subsequent burst of activity is the breakout the EA wants to capture.

The trap logic is event-driven, not bar-driven. OnTick() first refreshes the symbol and rejects any tick where the spread has blown out past the InpMaxSpread input (200 points on XAUUSD M5, i.e. a 20-pip tolerance — loose enough to survive most intraday widening but tight enough to filter illiquid prints). It then runs ManageProfitLock() against the open position list before the new-bar check, so the trailing engine is always one tick ahead of the signal engine. Once a new M5 bar opens, the EA pulls AMLI from the current closed bar (shift 1) and the bar before that (shift 2). The trap state is the boolean (amli_curr < InpLiquidityTrap), and the EA tracks a counter g_trap_duration that increments only while AMLI stays sub-trap. The signal fires on the transition: when the previous bar was in the trap and the current bar is not — provided the trap lasted at least two bars (g_trap_duration >= 2) and the absolute change in AMLI exceeds the InpMinAMLIChange floor (3.0 index points). When the trap-breakout gate is missed, a second path is left open: a strong-momentum entry is allowed whenever |amli_change| >= InpMinAMLIChange * 1.5, regardless of trap state. The direction follows the sign of the AMLI delta — positive change is a buy, negative change is a sell.

Position management is where EX12017 quietly diverges from a vanilla breakout EA. It does not pyramid by adding to a position on a new signal at the same level; it only adds when the existing position has already banked secured profit. CheckPyramidSafety() walks the open list and computes secured = (sl - open) / _Point for buys (and the mirror for sells). If that secured distance is below InpPyramidMinProfit (100 points, 10 pips), the new entry is rejected. This means the second and third pyramid adds (the InpMaxPyramidTrades cap is 3) only happen after the trailing stop has been pulled above entry — the EA is forced to be patient. Adding against the open direction (a buy signal while a sell is live, or vice versa) is also blocked at the safety layer, so the EA cannot net itself into a hedged mess when the trap-breakout direction flips on consecutive bars. The total open-position ceiling is 3, which keeps the EA from compounding into a runaway basket when the trap re-arms.

The trailing engine is a step-ratchet, not a continuous curve. ManageProfitLock() only acts once the trade is at least InpLockTrigger (300 points, 30 pips) into profit, and even then it only moves the stop if the new target beats the current stop by at least InpLockStep (50 points, 5 pips). The new stop is set at current price minus InpLockDistance (100 points, 10 pips), which means once a long trade is past the 30-pip trigger, the SL follows price in 5-pip increments with a 10-pip cushion. That cushion is the trade's max drawdown from current price during the trail — it never widens. The original fixed SL (500 points) and TP (750 points) remain on the position; the lock is an overlay, not a replacement. TryModify_EX12017 wraps the modify in a 3-retry loop with 100ms backoff on TRADE_RETCODE_REQUOTE and TRADE_RETCODE_TIMEOUT, and TryClose_EX12017 / TryClosePartial_EX12017 do the same for closes with 200ms backoff on the broader requote/timeout/price-off/price-changed set.

Risk and money management follow a clean separation. GetLotSize() sizes from balance when InpRiskPercent is positive (1.0% of account balance by default), with the per-point value computed against InpStopLoss so that a wider stop produces a smaller lot — the standard risk-normalized sizing math. When InpRiskPercent is set to zero, the EA falls back to InpFixedLot (0.01) and a hard cap at InpMaxLotSize (10.0 lots per trade). The stop distance is InpStopLoss = 500 points (50 pips on 5-digit XAUUSD), the take is InpTakeProfit = 750 points (75 pips), giving a 1:1.5 risk:reward on the fixed branch. The EA is wired to m_trade.SetTypeFilling(ORDER_FILLING_FOK) and SetDeviationInPoints(InpSlippage=30), and the InpMagicNumber defaults to 22212017 so multiple instances on the same account can be partitioned cleanly.

The time filter is plain-vanilla but worth noting. IsTradingTime() parses InpStartTime ("08:00") and InpEndTime ("20:00") into minutes and gates the new-bar branch on that window. Outside the window, the EA neither opens new trades nor modifies existing ones (ManageProfitLock still runs because it sits above the time gate), unless InpCloseAtEnd is true — in which case CloseAllPositions() is invoked to flatten the book. The 08:00-20:00 server window covers the full London session and the first two-thirds of the New York session, which is where XAUUSD spot volume and the trap-vs-breakout behavior the AMLI is measuring actually occur. The fill mode is FOK, so any trade that cannot be filled at the requested price is rejected outright rather than partially filled.

What to expect in a backtest versus live. Because AMLI is a derived indicator (volume ratio + ATR ratio normalized to 0-100), the EA's behavior is highly regime-dependent: in trending gold years (2020, 2024) the strong-momentum branch produces a steady stream of entries; in range-bound years (2022 Q3) the trap-breakout branch fires less often and most signals come from the relaxed 1.5x threshold. The 500/750 fixed point SL/TP pair means the EA does not breathe with volatility — on a 2x ATR day, the 50-pip stop is tighter than ideal, and on a 0.5x ATR day it is generous. The pyramid cap of 3 is conservative; aggressive users will want to stress-test InpMaxPyramidTrades up to 5 only after they have seen the secured-profit gate work the way they expect. A live deployment should run with InpRiskPercent = 0.5 (not the default 1.0) for the first two weeks, sized to the worst-case 3-position basket at 3% of balance per slot, and InpUsePyramid should be flipped to false for the first week of paper trading to isolate the single-entry branch.