Pipsgrowth EX12036 MultiIndicatorConfluence

Pipsgrowth EX12036 sits in the MultiIndicatorConfluence family and bills itself as a No-Lag LWMA + CCI + Zero-Lag RSI confluence system, but a careful read of the .mq5 shows a more interesting story: every indicator in the stack is actually a standard MT5 handle. The header still says 'Absolutely No-Lag LWMA Period = 14', the CCI block still says 'CCI-T3 Smoothing Period = 5' with a T3 coefficient of 0.618, and the RSI block still says 'Zero-Lag RSI Period = 14', but the OnInit() body says the quiet part out loud — the author calls standard iMA with MODE_LWMA, standard iCCI, and standard iRSI, and leaves a comment that reads 'Using standard indicators as approximations since custom indicators need to be compiled'. So the 'no-lag' label is descriptive, not a guarantee. What the EA really does is run a three-line mean-reversion gate on top of a linear-weighted trend filter, then optionally lets a fourth indicator (ADX) veto trades when the market is ranging.

The signal engine is in GenerateSignals(). On every new bar, after IsNewBar() confirms a fresh candle and UpdateIndicatorValues() copies three buffers from each handle, the code defines a long condition as: (1) the LWMA is rising — current LWMA > previous LWMA — and the close is above the LWMA, (2) at least one of three sub-conditions is true — CCI crossed above -100 from below, Zero-Lag RSI crossed above 30, or the CCI is already positive — and (3) when UseADXFilter is on, ADX > 25 and +DI is above -DI. The sell side is a mirror. This 'one-of-three' structure is the practical meaning of the 'relaxed conditions for testing' comment in the source. It is deliberately permissive: a single bar with the right confluence is enough to trigger, but in the absence of a CCI/RSI cross the EA will still fire a buy if the LWMA is up and CCI is already in positive territory.

What makes EX12036 different from its EX12 siblings is the second switch. The header's 12-layer claim is marketing copy; the actual behaviour is exposed by the next five lines: 'UseSimpleStrategy = true' is the default. When neither the confluence check nor its mirror fires, the code falls through to a plain moving-average crossover — close above LWMA while previous close was below previous LWMA = buy, the opposite = sell. With the default inputs, that fallback path is the one most new bars will actually take, because the strict confluence of 'LWMA up + CCI crossed -100 + RSI crossed 30' on a single bar is rare. The fallback turns the EA from a strict-confluence system into a permissive LWMA-trend-follower with a triple-confirmation upgrade path. Flip UseSimpleStrategy off and the EA stops trading in ranges; flip the optional UseADXFilter on and the same confluence has to clear ADX > 25 with the right DI alignment.

Open and close paths are simple and a bit brutal. OpenBuyPosition() and OpenSellPosition() both call trade.Buy() / trade.Sell() with explicit sl=0 and tp=0 — no stop, no target on the order itself. The only position-management is ManagePositions(), which only acts when UseTrailingStop is on, only touches profitable positions, and ratchets the stop on every new bar where the price has moved at least TrailingStep (default 10 points) beyond the prior stop. TrailingStop defaults to 50 points, so the trail keeps roughly 50 points of distance. There is no break-even step, no dynamic-lot scaling, no per-tick ratchet — the EA operates strictly on bar-close events because OnTick() returns early on every tick that is not the first tick of a new bar. If a position goes against the EA and never recovers, it stays open indefinitely; the trailing logic only tightens a stop that already exists. That is why the simple-fallback default is risky without an additional protective layer at the broker side.

The retry layer is the most carefully built part of the code. Both close helpers (TryClose_EX12036 with a ulong ticket and a string-symbol overload, plus TryClosePartial_EX12036) loop three times on the standard transient retcodes — TRADE_RETCODE_REQUOTE, TRADE_RETCODE_TIMEOUT, TRADE_RETCODE_PRICE_OFF, TRADE_RETCODE_PRICE_CHANGED — sleeping 200ms between attempts. TryModify_EX12036 (both overloads) accepts a narrower set: only REQUOTE and TIMEOUT get the 100ms retry; PRICE_OFF and PRICE_CHANGED are treated as terminal for modifications. This split makes sense — a stale price on a close is worth re-trying, but a stale price on a stop-loss modify is dangerous because the broker may have moved the SL/TP to where you asked while the market was elsewhere, and re-issuing the same SL is a real risk. The retry helpers are defined in the source but ManagePositions() never calls TryClose_EX12036 in this EA, so the close path only fires from the opposite-signal block in OpenBuyPosition / OpenSellPosition. That is the only way a losing position is exited: when the EA flips bias, the opposite order is opened and the prior position is closed via TryClose_EX12036 with its 3-attempt 200ms cadence.

Capital protection is light by design. There is no per-day loss cap, no max consecutive-loss cooldown, no equity-floor stop, and no position-sizing formula beyond a single LotSize input. Magic number is 22212036 and the trade comment is the literal string 'Psgrowth.com Expert_12036'. A new position on the same symbol in the same direction is rejected — PositionSelect() returns early if a same-direction position is already on. There is no SetDeviationInPoints call, no SetTypeFillingBySymbol, no GetBestFillingMode probe. On a broker that requires a non-zero deviation, every order will reject with TRADE_RETCODE_PRICE_CHANGED and the EA will log the failure but never trade. That is the single most common cause of 'EA does nothing' reports on this build.

What to expect in a backtest: with UseSimpleStrategy = true and UseADXFilter = false, the EA behaves like a classic LWMA-trend-follower on M5. It will catch a meaningful fraction of trending bars, take small losses on the range-bound ones, and rely on the 50-point trail to convert trends into runners. With UseSimpleStrategy = false and UseADXFilter = true, it becomes a strict triple-confluence swing trader with a low trade frequency — expect fewer than five entries per week on XAUUSD M5 in either direction. The minimum recommended balance is $100; at the default LotSize of 0.10 on XAUUSD M5, point value is roughly $1 per 10-point move per 0.01 lot, so 0.10 lot against a 50-point stop is about $50 of risk per trade, which means a $100 account cannot survive more than two consecutive full-stop losses. Raise LotSize only after raising the account, or test with 0.01 first to confirm the broker accepts the order format. The session filter is off by default; if you turn it on, the StartHour/EndHour inputs are server hours, not your local hours, so check your broker's server time offset before relying on the gate.