Pipsgrowth EX14001 Scalper

The Pipsgrowth EX14001 Scalper is a closed-bar trend-following EA built around an inline Hull Moving Average consensus on a configurable signal timeframe (H4 by default). The position-management code is conservative: 0.5% risk per trade, a 1.5:1 minimum reward-to-risk on the take-profit, an ATR(14) trailing stop that ratchets behind price, and a one-shot break-even that locks risk at +100 points of profit. The system reads as a deliberate, slow scalper rather than a tick-grabber: the order book is held to a single position per symbol, and every new entry must clear a 9-gate no-trade check before a market order is sent.

The Hull Moving Average is computed inline rather than as an iCustom call. The function HullMAAt(period, tf, shift) builds the classical Allan Hull construction — Hull(n) = LWMA( 2 * LWMA(n/2) − LWMA(n), sqrt(n) ) — by pulling the close series from the signal timeframe with CopyClose and running the lightweight linear-weighted moving average across the requested period. Two Hulls are sampled per evaluation: a fast Hull at InpHullFastPeriod = 65 bars, and a slow Hull at InpHullSlowPeriod = 90 bars, each measured at two closed-bar shifts (1 and 2). The signal direction is derived from the slope of each: the fast direction is sign(hf1 − hf2) and the slow direction is sign(hs1 − hs2), where the sign function is bounded by a single _Point of dead-zone. A signal is fired only if both slopes are non-zero and both point the same way; disagreement between the two Hulls — or a flat slope on either — returns a no-trade. A confidence value (20-100) is also returned, scaled by the magnitude of the slow slope relative to the fast.

Layered over the signal is a 7-state market regime classifier. ClassifyRegime() pulls 50 bars of ATR(14) to compute an ATR percentile (how the current ATR ranks against the last 49 readings) and 50 bars of Bollinger Band width (period 20, deviation 2.0) to compute a BB-width percentile. It also reads ADX(14). The output is an ENUM_REGIME label: ADX ≥ 25 → REG_STRONG_TREND; ADX ≥ 18 → REG_WEAK_TREND; ATR-percentile ≥ 80 or BBW-percentile ≥ 75 → REG_EXPAND; BBW-percentile ≥ 60 → REG_BREAKOUT; ATR-percentile ≤ 25 with BBW-percentile ≤ 20 → REG_COMPRESS; ADX < 15 → REG_CHOPPY; the rest falls into REG_RANGE. Two of those states — CHOPPY and COMPRESS — block entries entirely; the rest are permitted. This means the EA will deliberately pass on setups during congestion or when volatility is drying up, and only engages when ADX is at least 18 or when the Bollinger envelope is opening up.

The confirmation layer requires the HTF EMA(50) to agree. ConfirmHTF() reads the closed bar of the higher timeframe (D1 by default via InpHTFtf = 7) and checks that the candle close is on the correct side of the EMA — above for long signals, below for short signals. A long signal that fires on the H4 Hull consensus but where the daily candle is sitting below its 50-EMA gets dropped on the floor. The HTF EMA is only sampled on the closed bar (shift 1), never on the still-forming bar, to avoid acting on noise.

The order flow itself runs through a 9-gate NoTradeReason() check before any ticket is sent. The gates in evaluation order: InpKillSwitch on, equity below InpCapitalCapFloor (50 by default), open position count at InpMaxConcurrentTrades (1), 100-trades-per-day hit, g_consec_losses ≥ InpCooldownAfterLosses (3), daily realized loss ≥ InpMaxDailyLossPct (3%) of effective capital, weekly realized loss ≥ InpMaxWeeklyLossPct (6%) of effective capital, spread > max(30 points, broker stops-level × 2), and either CHOPPY or COMPRESS regime. After that, the session filter requires server time to fall between 07:00 and 21:30, with an additional cut-off at Friday 21:00 to avoid the weekend closeout. Any single failing gate returns a string reason that is printed to the log and the entry is abandoned.

Sizing is risk-based. CalcLot() takes the SL distance (ATR(14) × InpATRStopMult = 2.5) and converts it into a lot size such that the per-trade risk equals InpRiskPercent (0.5% of effective capital) of the equity. The lot is floored to the broker's volume step, clamped to the symbol's min/max volume, and re-validated by OrderCalcMargin() against the free margin before the order is sent. If InpCapitalCapAmount is set to a positive value, effective capital becomes min(cap, equity) — a way to size from a fixed notional rather than from the full balance.

Once a position is open, the exit logic runs on every closed bar of the signal timeframe regardless of whether a new entry is being considered. The break-even is a one-shot: once profit ≥ InpBreakevenPoints (100 points), the SL is moved to open + 5 points (a small lock) and never reset. The ATR trail, controlled by InpTrailATRMult (2.0), pulls the stop behind price only once the position is in profit by at least the trail distance; if the ATR sample is invalid, the trail is skipped for that bar. The opposite-signal exit is the third lever — if the Hull consensus flips direction on a new bar, all live positions in the opposite direction are closed via TryClose_EX14001(), which attempts the close three times with a 200ms sleep on REQUOTE / TIMEOUT / PRICE_OFF / PRICE_CHANGED, then gives up. TryModify_EX14001() does the same retry pattern for the SL/TP modifications, but only on REQUOTE / TIMEOUT (100ms sleep).

The order layer itself is conservative. trade.SetDeviationInPoints(50) caps the slippage at 50 points on a market order, and the open attempts a single retry on REQUOTE or TIMEOUT before returning false. Filling mode is selected automatically via SetTypeFillingBySymbol. The capital cap, kill switch, and dry-run flags are all operator inputs — InpCapitalCapAmount defaults to 0.0 (no cap), InpKillSwitch defaults to false, and InpDryRun defaults to false. InpDryRun is worth knowing about: when flipped to true, the EA logs the order it would have sent (with full price, SL, TP, ATR, lots) without actually placing a ticket — useful for forward-testing the new-bar timing on a live account before risking capital.

A backtest run with the EX14001 defaults — H4 signal timeframe, D1 HTF, XAUUSD, M15 chart, 0.5% risk, 2.5 ATR stop multiplier, 1.5 RR — will produce a low trade count compared to a tick-scalper. The shape of the equity curve depends more on whether H4 Hull consensus plus D1 EMA agreement fires on the historical bar than on micro-scalping alpha, so most of the time the EA sits flat. That is by design: this is a regime-conditional trend engine, not a maker-style market-neutral grid. Run it on a sub-30-point gold-spread ECN with at least $500 to give the 0.5% sizing room to breathe through the 3-loss cooldowns, and expect long flat patches punctuated by short bursts when H4 and D1 finally line up.