Pipsgrowth EX12033 MultiIndicatorConfluence

Pipsgrowth EX12033 EA3 is the smallest, cleanest member of the MultiIndicatorConfluence family built directly on MetaTrader's standard CExpert base class. Where most EAs in this family hand-roll their own signal engine in OnTick, EX12033 leans entirely on the wizard-generated CExpertSignal/CExpertSignal filter pattern: two indicator filters contribute a strength score on every bar, the platform sums the weighted scores, and a position opens when the combined direction crosses the configured threshold. There is no custom multi-timeframe matrix, no Kelly-style position scaling, no news filter. What the .mq5 actually contains is 335 lines organised around four input groups and three retry helpers — a deliberately lean implementation meant to be a portable starting point rather than a fully opinionated system.

The signal stack is two filters, both running on the chart timeframe. The first is a Double Exponential Moving Average with period 12, applied to the close, weighted 1.0; the second is an 8-period Relative Strength Index on the close, also weighted 1.0. Each filter returns a per-bar direction vote in the -100 to +100 range that the standard MQL5 SignalDEMA and SignalRSI classes are designed to produce. CExpertSignal aggregates those votes: a long is opened when the aggregate strength reaches or exceeds Signal_ThresholdOpen (default 10), and a short is opened when the aggregate strength is ≤-10. The same threshold gate is mirrored for exits in Signal_ThresholdClose (default 10). Because both filters carry the same weight, the threshold of 10 in practice requires at least one filter to be confidently directional — a setup in which both filters disagree tends to fall short of 10 in absolute value and produces no entry. The 8-period RSI is unusually fast; in combination with a 12-period DEMA this EA reacts to short-term momentum shifts, not the slower regime changes a 70/290 EMA pair would highlight.

Risk on the open is governed by Signal_StopLevel and Signal_TakeLevel, both defaulted to 50 points. With XAUUSD on a 2-decimal broker those 50 points are 5 pips; on a 3-decimal broker they are 0.5 pip. The 1:1 reward-to-risk ratio is intentionally narrow — it reflects the EA's short-horizon view. A wider stop would not help here because the position is expected to either work quickly (the 1:1 TP fires) or to be ratcheted out by the trailing stop before the SL is ever tested. The pending-order mechanism (Signal_PriceLevel, Signal_Expiration=4 bars) is wired through CExpert but disabled at the default PriceLevel=0.0; if a trader sets PriceLevel to a non-zero value, the EA will route signals through limit/stop pending orders with a 4-bar expiry instead of firing market orders directly.

Position management uses CTrailingPSAR, the standard Parabolic SAR trailing class. Step defaults to 0.02 and Maximum to 0.2 — the textbook values for a SAR applied to a 4–5 digit broker, fast enough to lock profits within a few bars of an SAR flip. Because Parabolic SAR trailing operates on the SAR dot, the stop only ever moves in the trade's favour, never backwards. On XAUUSD M5 this means a tight ratchet: as soon as price extends by roughly 0.02 * ATR beyond the most recent SAR dot, the new SAR dot becomes the new stop. Combined with the 1:1 fixed TP, most profitable trades close either on the TP fill or the PSAR trail, with very few hitting a true SL. Money management is delegated to CMoneyFixedLot: the trade size is 0.1 lot by default, and Money_FixLot_Percent (10.0) is the platform's margin-percent clamp that caps the notional relative to free margin — it does not size the lot dynamically. To change position size the trader changes Money_FixLot_Lots directly.

The header describes a 12-layer stack (REGIME, SIGNAL, ENTRY, CONFIRM, NO-TRADE, CAPITAL CAP, RISK, SIZING, MANAGE, EXIT, SCALING, OnTester) but in the source code the actual wired layers are SIGNAL, SIZING, MANAGE and EXIT — six of the twelve are placeholders. There is no regime classifier, no news filter, no scaling-in, no OnTester formula, no equity cap. This is worth knowing up front: a backtest will look like a pure DEMA+RSI crossover system with PSAR trailing, and that is exactly what the strategy is. Anyone who wants regime detection, news avoidance, or scaling-in should look at the heavier MultiIndicatorConfluence variants in this family (e.g. EX12010's Ehlers Gaussian or EX12020's directional grid).

Symbol and timeframe portability is one of EX12033's strengths. The CExpert framework abstracts the symbol away: the only symbol-specific code is Symbol() in OnInit. The DEMA(12) and RSI(8) parameters that work on a 2-decimal gold feed will need different point interpretations on EURUSD (50 points = 0.5 pip) or USDJPY (50 points ≈ 5 pips on a 3-digit broker), so the trader is responsible for scaling Signal_StopLevel and Signal_TakeLevel when moving the EA across symbol classes. Timeframes M5 through H1 are the suggested range — lower than M5 and the 8-period RSI starts to lose its meaning; higher than H1 and the 12-period DEMA becomes too smoothed for the threshold-10 logic to fire often enough.

The three helper functions at the bottom of the file — TryClose_EX12033, TryClosePartial_EX12033, TryModify_EX12033 — wrap a CTrade object with up to 3 retries and 100–200 ms sleeps on the standard transient retcodes (REQUOTE, TIMEOUT, PRICE_OFF, PRICE_CHANGED). TryModify uses a tighter retcode set (REQUOTE/TIMEOUT only) than TryClose (which includes PRICE_OFF and PRICE_CHANGED), reflecting that a price-change rejection on a modify is usually a broker-quote issue, not a transient one. The trade wrappers are not invoked from the EA's main path (the CExpert framework calls the trade object directly); they exist as drop-in helpers for any custom modification a developer wants to add later.

In backtest this EA is a fast-cycle DEMA/RSI system: many small trades, frequent PSAR flips, a 1:1 RR that wins roughly as often as the underlying signal quality. The lack of regime filter means it will trade in both trending and ranging markets; on a tight range it will whipsaw around SAR flips. Traders who see a flat or sideways backtest curve should not be surprised — that is the EA's natural behaviour with these parameters. To reduce whipsaw, raise Signal_ThresholdOpen to 20 or 30 (forcing stronger filter agreement), increase the DEMA period from 12 to 18 or 21, or push the SAR step from 0.02 to 0.03 for a slower ratchet.