Pipsgrowth EX16051 Trend

Pipsgrowth EX16051 Trend is a role-assignable multi-indicator scalping EA, hard-locked to USDJPY at OnInit (the function returns INIT_FAILED outright if the chart symbol is anything else and OnlyTradeUSDJPY stays true), built around a single architectural idea: every indicator on the chart can be promoted to one of seven roles — Primary Signal, Confirmation, Filter, Exit Signal, Exit Filter, Trend Filter, or Disabled — and the EA then runs the same logical chain regardless of which indicator you assigned. The shipped role map is EMA(8) crossed with EMA(21) as Primary Signal, RSI(14) as Confirmation, Stochastic(14,3,3) as a second Confirmation, MACD(12,26,9) as Filter, Bollinger Bands(20,2) as Exit Filter, and ATR(14) as Filter, with an extra ATR(14) on the higher timeframe pulled in for multi-timeframe alignment. That means eight native handles are created in InitializeIndicators() — fastEMA, slowEMA, RSI, MACD, BB, Stochastic, ATR, and ATR_HTF — and the same buffers are read three different ways depending on role: GetEMASignal() returns 1 / 0 / -1, GetRSISignal() returns 1 / 0 / -1, GetMACDSignal() returns 1 / 0 / -1, GetStochasticSignal() returns 1 / 0 / -1, and the same numbers are then consumed by GetPrimarySignal(), GetTradeSignal(), CalculateSignalStrength(), PassesFilters(), and CheckExitSignals() in turn.

GetEMASignal() is the only one configured as Primary by default and it gates everything. A buy needs the fast EMA crossing above the slow EMA on the current bar (fast1 ≤ slow1 and fast0 > slow0), the slope of the fast EMA on the latest bar (fast0 − fast1) being at least EMA_MinSlope = 0.00020, and the absolute separation between fast and slow being at least EMA_MinSeparation = 0.0003. A sell is the mirror. If EMA_RequireCrossover is flipped off, the function degenerates into a directional check: positive separation plus positive slope = buy, negative separation plus negative slope = sell. Every signal-reading function has the same three-bar structure (current, previous, two-bars-back) and the same slope-momentum logic — RSI only counts a bounce if r0 was oversold (≤30) one bar ago AND r0 > r1 > r2 AND r0 is still inside the neutral band 40–60, so RSI fires on a controlled recovery, not a raw oversold read; MACD requires a histogram zero-cross (hist1 ≤ 0 and hist0 > 0) with the main line above the signal line and, when MACD_RequireHistogramGrowing is on, hist0 > hist1; Stochastic requires %K to cross %D (k1 ≤ d1 and k0 > d0) with both lines moving in the same direction and %K sitting between the 20 and 80 levels.

GetTradeSignal() reads the primary, then counts how many of the four indicators currently sit in ROLE_CONFIRMATION agree with it, and demands the count reaches RequiredConfirmations (default 3). If three or four of them line up with the EMA direction, the function then calls PassesFilters() which walks every indicator still in ROLE_FILTER — in the shipped setup that means MACD and ATR — and also Bollinger Bands when its role is set to filter, with each filter hard-failing the signal if its individual condition is not met. The BB filter rejects entries where the current close sits closer to either band than BB_MiddleBandFilter (0.5 means within 50% of half the band width) and rejects during a squeeze (band width < ATR × 1.5). The ATR filter rejects when atr[0] in pips falls outside MinATR_Pips..MaxATR_Pips, default 8 to 20 pips on a 5-digit USDJPY chart. Bollinger Bands is wired as ROLE_EXIT_FILTER by default, so BB itself doesn't veto entries but contributes to exit decisions via PassesBBFilter().

CalculateSignalStrength() takes the base (confirmations / total enabled confirmations) and stacks four bonus terms on top: +0.15 if IsTrendAligned (close > fast EMA > slow EMA for buys, the mirror for sells), +0.10 if IsHTFAligned (current HTF ATR not more than 5% below previous HTF ATR — a deliberately permissive check that essentially means 'HTF ATR is stable'), +0.05 if current ATR is within 20% of the midpoint of the volatility band, and a session bonus of 0.03 for Tokyo, 0.05 for London, 0.05 for New York, 0.02 for Sydney. The total is clamped to 0..1 and the entry chain only fires when the value is at least MinSignalStrength = 0.85 AND MinSignalConfidence = 0.90 — those two thresholds are gated separately, so a setup with a 0.87 base strength gets rejected by the quality filter while a 0.91 setup that fails the bad-trade filter is also rejected. With three or four confirmations plus a trend-aligned bonus, the typical passing setup lands in the 0.85–0.95 range, which is exactly the narrow window the EA is designed to fire in.

The execution chain is straightforward. ExecuteBuyTrade() pulls the ask, calls CalculateLotSize() (which returns the fixed LotSize = 1.0 unless UseAutoLotSize is on, in which case it sizes to RiskPercent = 1.0% of balance divided by the SL distance in pips times the tick value), CalculateStopLoss() (15 pips or ATR × ATR_StopLossMultiplier = 2.0, whichever is larger and broker-stops-level-safe), and CalculateTakeProfit() (25 pips or ATR × ATR_TakeProfitMultiplier = 3.0), then submits a market order via trade.Buy() with ORDER_FILLING_FOK and a slippage allowance of SlippagePoints = 5. The trade comment embeds the signal strength as a decimal for later forensics. The same path runs in reverse for sells. CalculateStopLoss() and CalculateTakeProfit() both honour SYMBOL_TRADE_STOPS_LEVEL — if the broker enforces a 10-point minimum stop distance, the EA widens the SL to that minimum rather than sending a rejected order — and both cap their final distance at MaxATR_Pips = 20 pips, so the EA will not place a 60-pip SL on a single 60-pip-Atr bar even if StopLossPips is set to 50.

Management runs every tick once the entry throttle (EntryTickDelay = 2 ticks) has elapsed. ManagePositions() loops the open positions and, for each, calls CheckQuickExit() first, then CheckMaxHoldingTime(), then UpdateTrailingStop(), then CheckExitSignals(). CheckQuickExit() re-evaluates GetTradeSignal() and closes the position if the current signal direction is opposite to the position type — so a buy that flips to a confirmed sell closes immediately. CheckMaxHoldingTime() counts closed bars since the open time via iBarShift and force-closes when the count reaches MaxHoldingBars = 10. UpdateTrailingStop() ratchets the SL by TrailingStopPips = 10 pips every TrailingStepPips = 5 pips of favourable move, and only moves the SL toward the open price and beyond it (no initial break-even if the EA opens a trade at the broker minimum stop distance and price hasn't moved 10 pips yet, the trail stays put). CheckExitSignals() walks the indicators in ROLE_EXIT_SIGNAL — in the shipped map that includes EMA, RSI, and MACD — and closes the position if any of them fires a direction-flipping signal. The trailing modification goes through TryModify_EX16051 (3 attempts, 100 ms sleep on REQUOTE or TIMEOUT) and the close goes through TryClose_EX16051 (3 attempts, 200 ms sleep, also handles PRICE_OFF and PRICE_CHANGED). TryClosePartial_EX16051 is defined at the file footer but never called from any live path — it's dead code carried over from the template.

Two additional safety rails run before entries. CheckDailyLimits() reads the broker-side account balance, tracks daily P/L deltas in UpdateDailyPL(), and force-closes every open USDJPY position if dailyLoss ≥ MaxDailyLoss = 50.0 or dailyProfit ≥ MaxDailyProfit = 150.0. The new-day reset is keyed off a StringToTime of the YYYY.MM.DD date string, so the rollover fires at server midnight. IsBasicTradingAllowed() is the per-tick guard: it blocks when the spread exceeds MaxSpreadPips = 2.0, when the GMT hour isn't inside any of the four allowed sessions (Tokyo 23–07, London 08–16, NY 13–21, Sydney 21–05), when AvoidNews is on and the current minute is within NewsAvoidanceMinutes = 30 of the next hour mark, when ATR isn't in MinVolatilityPips..MaxVolatilityPips, when the time since lastTradeTime is less than MinBarsSinceLastTrade × PeriodSeconds, or when the open position count has hit MaxConcurrentTrades = 3. ValidateBuySignal() and ValidateSellSignal() add a second pass that rejects entries near recent 10-bar highs (resistance) or lows (support) within 10 points, rejects whipsaw setups (CheckWhipsaw() requires current ATR ≥ midpoint of recent ATR range × WhipsawThreshold = 0.8), and rejects entries when the current bar's tick volume is below MinBarVolume = 10.

What you actually get in live trading is a slow but selective system. With M5 USDJPY and the shipped session filter on, the EA only fires inside four windows totalling roughly 20 hours per day but the actual entries land in the 0.85–0.95 strength band, the SL is 15 pips with a 25-pip TP for a 1:1.67 reward-to-risk, the trail activates at +10 pips and steps every +5 pips, and the daily caps cut the EA off at +$150 or −$50 whichever comes first. OnInit, OnTick, and OnDeinit are the only event handlers — there is no OnTrade, no OnTimer, no OnChartEvent, no OnTester. The HTF ATR pull is the only piece of multi-timeframe work the EA does, and IsHTFAligned is a deliberately permissive check (current HTF ATR within 5% of previous), so it doesn't actually filter out a meaningful number of setups. The two thresholds MinSignalStrength and MinSignalConfidence are checked sequentially with different minimums, which means a setup that crosses one but not the other still fails — that is the design, not a bug.