Pipsgrowth EX12045 MultiIndicatorConfluence

EX12045 is the v1 baseline cut of the MultiIndicatorConfluence family — the cleanest input panel in the lineup at six groups and thirty-five parameters, with no separate entry-mode toggles, no crossover/divergence sub-banks, and no breakout/mean-reversion switches bolted on top. The whole trade decision collapses to one line: a four-way AND-stack on each tick of a new bar, and that line is the same one the live path has always used. Later siblings in the family add per-mode flags and a broader input bank; this file ships without them, which makes it the most direct expression of the underlying confluence idea in the .mq5 tree.

The signal path is short. On every new bar, OnTick() calls UpdateAdaptiveIndicators(), which in turn calls four code-level calculators — CalculateAdaptiveVolatility(), CalculateAdaptiveMA(), CalculateRSI(), and CalculateHeikenAshi() — that fill seven manually-populated 100-element arrays (adaptiveMA, volatility, rsiBuffer, haOpen, haClose, haHigh, haLow) without ever allocating a native MT5 indicator handle. The "adaptive moving average" is in fact a plain SMA whose period is recomputed from the current bar's high-low range relative to a twenty-bar average: adaptivePeriod = round(AdaptiveMA_Period / (1 + (volRatio - 1) * AdaptiveMA_Sensitivity)), then clamped to the [5, 50] integer band. Higher recent volatility shortens the average; quieter markets lengthen it back toward the base period of 14. The Wilder-style RSI is a textbook implementation summing closes over ConfirmationRSI_Period (default 14) bars, with a neutral seed of 50 in the buffer and the up-sum-over-down-sum RS formula. The Heiken Ashi candles are constructed in code (haClose = (O+H+L+C)/4, recursive haOpen[i] = (haOpen[i-1] + haClose[i-1]) / 2), not pulled from a chart indicator — so disabling UseHeikenAshi swaps the comparison over to raw candle close, but the buffer is still built every tick regardless.

GetBuySignal() and GetSellSignal() are mirror images and each return a single boolean from a small AND chain. Buy fires when the working price is above the adaptive MA, AND either a fresh crossover happened on the most recent bar OR a strong bullish candle appeared with body larger than fifty percent of its full range, AND the latest volatility reading clears the VolatilityThreshold * pointValue floor (0.3 points by default). Sell reverses every comparison. After the signal, GetConfirmationSignal() gates the trade on the RSI sitting inside the 30-70 corridor, with a volumeConfirm flag hard-wired true because retail EURUSD rarely has trustworthy tick-volume data. Three further gates sit between the signal and the order ticket: IsMarketConditionsOK() blocks entries when the spread exceeds MaxSpread (5 points) or, more strictly, when it exceeds 2x MaxSpread as a hard extreme-spike cut-off, or when the equity drawdown breaches MaxDrawdownPercent (10%), or when total open positions on this magic have already hit MaxOpenTrades (5). IsPositionManagementOK() walks every open position for the magic and demands that all of them be at or above MinProfitInPointsPerTradeToAdd (20 points) before any new entry is allowed — a one-strike-and-you're-out rule that prevents the EA from doubling into a losing basket. IsOptimalTradingSession() short-circuits to true when the tester flag is set, which means strategy-tester curves will look smoother than live forward runs once the session filter is engaged.

Order construction is plain CTrade: a market order at the current ask (or bid for sells) with a stop of StopLossPoints * pointValue (50 points) below and a take profit of TakeProfitPoints * pointValue (100 points) above — a 1:2 reward-to-risk on the initial ticket — using the fixed LotSize input (1.0 by default) regardless of MaxRiskPerTrade. Slippage tolerance is the Slippage input (3 points). When the same-direction position count is already above zero and AllowOnlyProfitableAdditions is true, the code re-checks IsPositionDirectionProfitable() for that specific type before sending a pyramid entry; this is a tighter check than the global position-management gate, but it uses the same 20-point threshold. RequireToCheckForConfirmationBeforeAdding forces the RSI corridor to be alive on the current bar before any addition, even on the first pyramid.

Position management runs every tick, not just on new bars, through ManagePositions() → ManageDynamicStopLoss(). The watermark ratchet tracks lastHighestProfit[positionIndex] per ticket, computes lockSteps = floor(peak / LockProfitEvery_X_Points) (every 30 points of unrealized gain), and proposes a new stop at PriceOpen + lockSteps * 30 - LockMinusBuffer (the 15-point buffer keeps the lock a hair below the round number). The modification only fires when the proposed stop is strictly better than the existing one, so the ratchet can only tighten — it never loosens. The TryModify_EX12045 helper retries the modify up to three times with a 100ms sleep on REQUOTE/TIMEOUT, and TryClose_EX12045 / TryClosePartial_EX12045 do the same for full and partial closes with a 200ms sleep on REQUOTE/TIMEOUT/PRICE_OFF/PRICE_CHANGED.

What this v1 file deliberately does not do is just as important as what it does. The .mq5 allocates zero native indicator handles — no iRSI, no iATR, no iBands, no iMACD calls anywhere in OnInit — so the visible inputs in the panel that read like crossover, divergence, breakout, mean-reversion, and HA-confirmation toggles on the later siblings are absent here. The single UseHeikenAshi boolean chooses between HA-built and raw-candle price comparison; that is the only signal-mode switch on the file. There is no ManagePositions body doing partial closes or basket TP, no working IsOptimalTradingSession curve for live trading, and no per-mode divergence routine. The 10% MaxDrawdownPercent global cap is the load-bearing safety in the absence of those features: if a streak of losers accumulates, the equity-floor gate freezes new entries and the per-ticket ratchet tries to salvage whatever profit is in the basket.

Deployment is EURUSD on M5 or H1 with a $100 account at 0.01-lot scaling (the LotSize input is a literal fixed lot, not a percentage of balance, so 1.0 means 1.0 standard lot — the user must scale it down to the broker's micro-lot step before live). The EA expects a low-spread ECN or no-dealing-desk feed (the 5-point ceiling and the 2x extreme-spike hard stop both assume sub-1.5-point typical spreads) and clean tick data (the iBars / iTime new-bar detection at the top of OnTick() will skip ticks until the bar count stabilises, which is correct for backtests and noisy for live under high-latency brokers). The London/NY session filter is off by default for forward-compatibility with 24-hour servers, but flipping UseSessionFilter to true and the three session toggles to taste confines entries to 08-21 GMT, with the 13-16 overlap the most active window. Above H1 the adaptive MA stops reacting to intraday volatility, so the strategy thins out into a slower trend-follower and the ratchet becomes the dominant profit mechanism — keep the timeframe at M5 for the design as written.