Pipsgrowth EX18113 TrendFollow

Pipsgrowth EX18113 is a USDJPY M5 multi-indicator scalper built around EMA(9/21) crossovers, with six separate confirmations stacked on top of every entry. The .mq5 file declares version 2.00, magic 22218113, and a 12-layer architecture spanning regime, signal, entry, confirm, no-trade, capital cap, risk, sizing, manage, exit, scaling, and OnTester — though the actual code is closer to ten functional layers (the regime filter is the True Hull MA slope via IsEMASlopeFlat, and there is no OnTester pass-through in this build; the trailing and pyramid blocks carry the manage and scaling roles).

The signal fires on closed-bar EMA crossings. CheckEntrySignals reads emaFast at bar 1 and bar 2 plus emaSlow at bar 1 and bar 2; a bullish cross is registered when bar-2 fast is at-or-below slow AND bar-1 fast is strictly above slow (and vice versa for shorts). Once a cross is detected, the EA does not just enter — it routes the trade through ValidateBuyFilters or ValidateSellFilters, which is a six-check consensus gate. First, the True Hull MA on the current timeframe must agree with direction (g_hullFast > g_hullSlow for longs). Second, the same relationship must hold on the higher timeframe (g_hullFastHTF > g_hullSlowHTF), where the higher timeframe is selected by the InpHTF dropdown (1–7 mapping to M1/D1, default H1). Third, the fast EMA slope over the last two bars must clear InpEMASlopeMin = 0.05 points/bar to avoid flat / chop regimes. Fourth, the daily pivot S/R filter rejects entries within InpSRProximity = 50 points of R1/R2 for buys and S1/S2 for sells. Fifth, the volume filter reads the 20-bar tick volume average and rejects when the last bar's volume is below 0.8x that average. Sixth, RSI(14) and MACD(12,26,9) histogram must agree on direction, ADX(14) must be at or above InpADX_Min = 25, and ATR(14) must be at or above InpATR_MinVol = 0.0005 to guarantee a live market.

A standout design choice is the True Hull MA. MT5 does not ship a built-in Hull indicator, so the .mq5 builds one from two LWMA handles per period: WMA(N) and WMA(N/2). CalculateHullMA computes rawHull = 2 * WMA(N/2) - WMA(N), then instead of the textbook WMA(rawHull, sqrt(N)) (which can't be expressed in nested LWMA handles), it smooths by averaging sqrt(N) recent rawHull values across the same bar shifts. This approximation is applied to four periods — fast and slow on the current TF, fast and slow on the higher TF — so the file allocates eight LWMA handles (hHull_WMA_N_Fast, hHull_WMA_N2_Fast, hHull_WMA_N_Slow, hHull_WMA_N2_Slow, plus the HTF mirror set) before any of the standard filter handles. The two EMA handles, the four filter handles (RSI, ADX, ATR, MACD), and the eight Hull handles total fourteen indicator handles, which is heavier than the typical TrendFollow sibling.

Risk is sized by CalculateLotSize, which defaults to 1% of balance via the formula lots = balance * 0.01 / ((slPoints * point / tickSize) * tickValue), floored to the lot step, clamped between the symbol's min and max, and snapped to two decimals. With InpFixedLot = 0 the EA is auto-risk; setting InpFixedLot = 0.01 switches to a fixed 0.01 lot per leg. The EA caps total positions at 3 (InpMaxPositions) and per-direction at 2 (InpMaxPosPerDir), but the pyramid block relaxes the per-direction cap to InpMaxPyramid = 5 legs on each side provided every existing leg on that side is in profit AND the latest leg is at least InpPyramidMinProfit = 30 points in the money. The pyramid uses a multiplicative size formula: lots = baseLot * MathPow(InpPyramidMultiplier, currentCount), so a 1.0 multiplier keeps legs equal, while 1.2 would scale up. SL and TP are static at InpStopLossPoints = 200 and InpTakeProfitPoints = 400 — a 1:2 risk-reward in JPY terms (20 / 40 pips).

Management is two-stage. ManagePositions runs every tick: once a position is at least InpBreakevenPoints = 50 points in the money, the stop is moved to entry + InpBreakevenOffset = 10 points (entry − 10 for shorts). Once profit reaches InpTrailingPoints = 100, the stop is ratcheted to currentPrice − 100 points (or + 100 for shorts) provided the new stop is tighter than the existing one. All modifications go through TryModify_EX18113, a 3-attempt retry wrapper that backs off 100 ms on requote or timeout. Exits are also driven by CheckExitSignals, which runs on each new bar: a reverse EMA cross flattens the opposite side via ClosePositionsByType, and an RSI extreme with realized profit (ClosePositionsAtExtreme) closes winners in that direction. There is no time-based exit.

The safety stack adds three more guards. IsDrawdownExceeded pauses new entries when equity drawdown from the all-time peak (g_peakEquity, updated on every tick) exceeds InpMaxDDPercent = 20%. IsWithinSession only allows trades when the server hour falls in the InpSessionStart–InpSessionEnd window, with the default 2-21 wrapping midnight for an Asia-through-NY book. The news filter is gated by InpUseNewsFilter = false; when enabled, LoadNewsEvents parses news_events.csv (date, time, currency, impact) and IsNearNews vetoes entries whose time falls within InpNewsBefore = 30 minutes before or InpNewsAfter = 15 minutes after a USD or JPY event. If the file is missing, CreateSampleNewsFile writes a header plus 2026 NFP, FOMC, and BOJ dates as a starting point.

The dashboard sits in the top-left corner as a 200x230 dark-slate-gray rectangle with thirteen labels — title, signal, EMA, Hull, RSI, ADX, pivot PP, position count, realized P/L, spread, volume, session status, news status. It refreshes every tick via ChartRedraw, which is unusual for an M5 strategy; expect CPU usage to track spread update frequency. Chart visuals add a dashed cyan line at the Hull Fast value and a dashed magenta line at the Hull Slow value (UpdateChartVisuals re-draws them on every new bar). When InpShowPivots is on, DrawPivotLines renders the daily PP in solid gold with R1/R2 in dotted red and S1/S2 in dotted green, recalculated at the day rollover from yesterday's D1 high, low, and close.

In backtesting on USDJPY M5, expect trade frequency of 5–15 entries per London–NY day when the trend regime is active and the consensus gate clears; flat Asia sessions produce very few signals because of the 2-21 session window and the ADX floor. The pyramid can stack five legs on a single impulsive USDJPY move, so the largest winning streaks will outsize the largest losing streaks 1.5–2x — but the 5-leg cap on each side plus the 20% drawdown pause should keep worst-case equity drift within bounds. As always, run the EA in Strategy Tester visual mode for at least a month before going live, and confirm the broker allows ORDER_FILLING_IOC for USDJPY (the EA hardcodes IOC in trade.SetTypeFilling).