Pipsgrowth EX18112 TrendFollow

The most stripped-down entry in the EX18xxx trend family, EX18112 trades a single Hull-style moving-average proxy on XAUUSD's 5-minute chart, confirmed by a one-sided RSI bound and a single Bollinger Band filter. The header file is Pipsgrowth_com_EX18112_PipsGrowth_XAUUSD_5M.mq5, version 2.00, magic 22218112, intended for gold on M5 only.

What the source actually does at every tick is read four indicator buffers and decide. CheckSignal() pulls bar-1 close along with the current fast LWMA, slow LWMA, RSI, and the upper/lower Bollinger Band from the handles created in OnInit(). The trend gate is binary: fast > slow is treated as up, anything else is down. There is no slope filter, no angle, no multi-timeframe confirmation — just the crossover state read on the previous bar. RSI then plays a one-sided ceiling/floor: longs require RSI below 70, shorts require RSI above 30. The Bollinger Band adds a "room to grow" condition — a long only fires if price is below the upper band, a short only fires if price is above the lower band. The result is a tight three-condition consensus that enters almost immediately when the cross direction is unambiguous and the band is not at an extreme.

The Hull implementation deserves a callout. The .mq5 header comments explicitly admit that Hull is not a built-in MT5 indicator, so the EA approximates Hull behaviour with two LWMAs: hLWMA_Fast = iMA(_Symbol, _Period, InpHullPeriod/2, 0, MODE_LWMA, PRICE_CLOSE) and hLWMA_Slow = iMA(_Period, InpHullPeriod, 0, MODE_LWMA, PRICE_CLOSE). With the default InpHullPeriod = 20, that is a 10-period fast LWMA and a 20-period slow LWMA, both on close. The user's choice of InpHullPeriod is a single knob — half of it drives the fast leg. RSI is iRSI(_Symbol, _Period, InpRSIPeriod=14, PRICE_CLOSE), and Bollinger Bands use iBands(_Symbol, _Period, InpBBPeriod=20, 0, 2.0, PRICE_CLOSE) with a 2.0 standard deviation, so the band sits on the wider side. All four handles are released in OnDeinit().

Risk management is the most surprising section of the code. The InpRiskPercent = 1.0 input exists in the Money Management group, but the EA never reads it. OpenBuy() and OpenSell() both call trade.Buy(InpFixedLot, _Symbol, mysymbol.Ask(), sl, tp, InpTradeComment) and the mirror for sell, with the fixed lot input default 0.01. Stop and take-profit are point-based, also from the Money Management group: InpStopLoss = 400 and InpTakeProfit = 800, with the actual price computed as mysymbol.Ask() - InpStopLoss * mysymbol.Point() (and the bid-side mirror for shorts). On XAUUSD that translates to roughly a $4 stop and $8 target against a 0.01 lot — a 1:2 reward-to-risk bracket that depends on the symbol's point definition (XAUUSD typically has two decimal places on retail feeds, so 400 points is $4 of price distance). The 0.01 / $100 minimum deposit the broker card advertises is the floor the EA was sized for; anything smaller risks margin rejection on the bracket itself.

The pyramiding code is the second surprise. ManagePyramid() exists with the right signature and is called every tick from OnTick() after mysymbol.RefreshRates(). It checks InpEnablePyramid (default true) and gates on count == 0 || count >= InpMaxPositions with InpMaxPositions = 3 and InpStep = 200, but the body of the function does not actually add positions. The comment "Standard Step Logic (Same direction) ... (Simplified for brevity: check last pos, if > Step profit, add)" is followed by an empty block. With default inputs, a single entry per signal cycle is what actually happens; a second position is never opened by the EA itself. The header's "12 layers: REGIME, SIGNAL, ENTRY, CONFIRM, NO-TRADE, CAPITAL CAP, RISK, SIZING, MANAGE, EXIT, SCALING, OnTester" claim is aspirational — the implemented layers visible in the code are signal, entry, confirm (RSI + BB), bracket risk, and a partial exit attempt via the three retry helpers.

OnTick() is short. It refreshes the symbol, runs ManagePyramid() (which is a no-op for actual position building), and only calls CheckSignal() if PositionsTotal() == 0. That means signals are not generated while a position is open, so the EA waits for the bracket to resolve before re-entering. The filling mode is hardcoded: trade.SetTypeFilling(ORDER_FILLING_FOK). On brokers that do not support FOK — most ECN feeds default to IOC or RETURN — this single constant can cause outright order rejection. A trader intending to use EX18112 on a real account should either confirm FOK support with the broker or expect to switch the constant before deployment. The trade comment is set to InpTradeComment = "Psgrowth.com Expert_18112", which is the only on-platform identification beyond the magic number.

The three retry helpers — TryClose_EX18112, TryClosePartial_EX18112, TryModify_EX18112 — follow the same 3-attempt pattern as the rest of the family: each loops up to three times, sleeps 200 ms on a requote/timeout/price-off/price-changed return, and sleeps 100 ms for modify on requote/timeout, returning true only when trade.ResultRetcode() == TRADE_RETCODE_DONE or TRADE_RETCODE_DONE_PARTIAL. They are declared and visible in the function table, but no entry in the source calls them. Like the ManagePyramid body, the retry layer is in the file but not wired to the live order path; the trade.Buy / trade.Sell calls in OpenBuy() and OpenSell() go directly through CTrade::Buy / Sell and accept whatever single-shot return the broker sends.

That minimalism defines who should consider this EA. EX18112 is for traders who want a single signal, a fixed lot, a 1:2 R:R bracket, and no active position management beyond the bracket. It is not a "set and forget" grid, not a multi-timeframe system, and not an AI-augmented engine. It is, in effect, a hand-coded crossover with one filter from each of two well-known indicators, intended to teach the structure of a working but bare-bones MT5 expert advisor. A trader who wants to see what a single-condition crossover looks like in production code can read this file end-to-end in under five minutes and understand every line that touches an order.

Practical expectations on a 5-minute XAUUSD chart: the LWMA(10)/LWMA(20) cross fires on most M5 candles, so the gating factor is the RSI + BB room filter, not the cross itself. Expect entries to cluster around pullbacks inside the bands — the long signal requires price to be inside the upper band but not at it, which translates to late entries on retracements rather than breakouts. The fixed 0.01 lot with a 400/800 point bracket produces small absolute P&L per trade; scaling position size upward is on the trader, not the EA. The 250-point spread filter (InpMaxSpread = 250) is loose enough to clear on most retail brokers during London and New York hours, but a widened spread during the low-liquidity rollover can still pass through. Anyone deploying this in production should plan to test on a demo account for at least a full trading week before risking real capital, both to confirm the FOK filling path on their specific broker and to gauge how often the LWMA cross + BB filter combination actually fires versus how often it sits idle waiting for a band pullback.