Pipsgrowth EX12003 MultiIndicatorConfluence

Pipsgrowth EX12003 is a four-of-four agreement engine. Every entry decision has to clear an AND gate of four independent indicators on the same closed bar: EMA(12) above EMA(50), MACD main above its signal, RSI(14) sitting on the right side of the midline with a five-point buffer, and the closing price on the right side of the Bollinger Band(20, 2.0) midline. Any condition that fails to agree downgrades the vote; if even one of the four disagrees the EA returns 0 and does nothing. This is a deliberate trade for signal quality over signal frequency — the EA is not trying to find trades, it is trying to find bars where momentum, trend, oscillator, and price location all line up in the same direction at the same time. The four conditions are computed in the CoreSignal() function and counted as bullVotes/bearVotes out of a total of 4; the function only returns +1 or -1 when all four votes align, with confidence hard-pinned at 100.

The four signals are all inlined from native MQL primitives — iMA, iRSI, iMACD, iBands — plus iADX and iATR for the regime and risk layers. There is no iCustom call anywhere in the source, which means the EA reproduces identically across brokers and indicator vendors. The eight indicator handles (fast EMA, slow EMA, RSI, MACD, ADX, ATR, Bollinger Bands, and a higher-timeframe EMA) are created in OnInit and released in OnDeinit, and the OnInit block also pulls the broker's point, tick value, tick size, stops-level, and volume step/min/max so all the helpers (NormPrice, ClampVol, MinStopsDist) work portably. g_trade.SetTypeFillingBySymbol(_Symbol) lets the EA pick FOK, IOC, or RETURN automatically based on what the broker actually supports.

The regime classifier in DetectRegime() returns one of seven enum states — STRONG_TREND, WEAK_TREND, EXPAND, COMPRESS, BREAKOUT, RANGE, CHOPPY — by reading ADX(14) and an ATR percentile rank over the last InpATRPctLookback (60) bars plus the current versus prior Bollinger width. ADX at or above 30 is a strong trend; ADX at or above the InpADXMin of 20 is a weak trend; ATR-percentile above 70 with current BB-width at least 1.15 times the prior BB-width is expansion; ATR-percentile below 25 is compression; a very tight band with rising ADX is treated as a breakout setup; otherwise the bar is range. CHOPPY is the catch-all that the function returns when there is not enough indicator data on the bar — and CHOPPY is the one state the no-trade block rejects outright. A regime-change into CHOPPY, RANGE, or COMPRESS during an open trade also triggers an exit (see OppositeAndRegimeExit).

Three further filters run after the signal. ConfirmEntry() requires the close on the current timeframe to be on the right side of the higher-timeframe EMA(50) — the HTF defaults to H1 via MapTimeframeInt(InpHTF) where InpHTF=8 falls through to H1 — and the spread to be at or below InpMaxSpreadPts (250 points, set with XAUUSD in mind), and the server clock to fall inside the 6-22 trading window. NoTradeBlock() adds the kill switch check, the kill-switch-adjacent abnormal-spread and out-of-session checks, the CHOPPY regime check, a news/big-candle check that blocks entries when the previous bar's range is more than 3x the ATR (BIG_CANDLE_ATR_MULT), the cooldown-after-losses timer, the effective-capital floor (default $50), the 3% daily loss limit, the max-open-trades cap of one, and a Friday-after-21 plus weekend gate. Only when all of those clear does the EA try to send.

Position sizing is straightforward risk-percent. CalcLots(slDistancePrice) reads the effective capital — which is the smaller of InpCapitalCapAmount and account equity when the cap is enabled, or the full equity when the cap is zero (the default) — multiplies it by InpRiskPercent (0.5%), divides by the per-lot loss at the proposed stop, and floors the result to the broker's volume step. The stop distance is ATR(14) * InpATRSLMult (2.0), lifted to the broker's SYMBOL_TRADE_STOPS_LEVEL if necessary. The take-profit is slDist * InpRRTP (2.0), giving a 1:2 reward-to-risk ratio. Margin is pre-checked with OrderCalcMargin before any send, and InpDryRun is true by default — the EA logs the intended trade but does not call g_trade.Buy/g_trade.Sell until you flip it off. A single retry with a 150ms sleep handles a transient REQUOTE.

ManageExits() runs every tick, not just on a new bar, and layers five exit paths. The first three fire at +1R from entry: the stop ratchets to break-even at the open price (one-shot, never loosens), 50% of the position is closed as a partial TP (PARTIAL_TP_PCT=50), and the ATR(14)*InpATRTrailMult (2.5) trail starts ratcheting forward-only. The fourth is a time exit at MAX_BARS_IN_TRADE=120 bars (10 hours on M5, scaled to whatever the timeframe is). The fifth, in OppositeAndRegimeExit(), closes the trade when CoreSignal flips to the opposite side, or when the regime drops into CHOPPY, RANGE, or COMPRESS. The close and modify helpers (TryClose_EX12003, TryClosePartial_EX12003, TryModify_EX12003) each retry three times at 200ms or 100ms on REQUOTE, TIMEOUT, PRICE_OFF, or PRICE_CHANGED responses, and give up cleanly on anything else.

Loss tracking and the cooldown use the OnTradeTransaction event — not the EA's own bookkeeping — and read the DEAL_PROFIT, DEAL_SWAP, and DEAL_COMMISSION of every DEAL_ENTRY_OUT and DEAL_ENTRY_INOUT deal whose magic and symbol match. Three consecutive losses (COOLDOWN_AFTER_LOSS=3) set g_cooldown_until to now + 3 periods, after which the counter resets. The OnTester custom criterion is (net * profitFactor) / (1 + |balance DD|), gated on at least 30 trades, which keeps the EA's reported backtest score from being skewed by tiny sample sizes. The InpKillSwitch input is a panic button — set it to true and the EA will close every position on the next tick and not open new ones. Scaling into winners (pyramiding) is intentionally disabled: MAX_OPEN_TRADES is hardcoded to 1, the comment in the scaling section reads "Pyramid path intentionally not enabled — gating by max open trades."

What to expect on a chart. EX12003 is a session-filtered, regime-aware, all-of-confluence EA. On XAUUSD M5 in a clean London or New York trend, the four conditions align often enough to keep trade frequency in the comfortable single-digit-per-day range; in a quiet Asian session or a choppy re-entry pattern, you will see hours of zero trades and then one clean bar. The 1:2 reward-to-risk and the 1R BE+partial+trail stack mean a typical winner is 1.5-2x the stop, and the EA gives back very little on the losers because the stop moves to break-even before the partial fires. The dry-run default is the right place to start — let it log a few weeks of intended trades, check the magic and comment in the journal, then disable InpDryRun when you trust the regime/HTF settings on your broker's hours and gold-spread profile.