Pipsgrowth EX10011 Momentum-Scalper

Pipsgrowth EX10011 is a closed-bar momentum-scalper that fires trades only when at least four of six independent momentum voters agree on direction and a five-state regime classifier gives the green light. The six voters are computed inline from raw price data on the last closed bar (shift=1): a fast Hull moving average of period 21 against a slow Hull of period 55 (sloping up = long vote), an EMA-ribbon alignment check across periods 8/21/50 (all stacked in the trade direction), a Heikin-Ashi candle colour built from the OHLC of bars 1 and 2 (green = long vote), a MACD histogram direction on the (12,26,9) classic pair (main minus signal expanding in the trade direction), an RSI(14) side check that defaults to RSI ≥ 55 for long votes and RSI ≤ 45 for short votes (the asymmetric 55/45 split is what makes the EA prefer trades with confirmed trend bias rather than neutral-momentum setups), and a Bollinger Band squeeze-breakout check using the (20, 2.0) bands on close — a long vote when the current bar's band width is at least 1.4× the minimum width of the prior 100 bars, indicating that price has just broken out of a compression. Each voter adds one tally; the result is a confidence score in 0–100 (votes / 6 * 100) and the entry fires only when votes >= InpMinVotes (default 4). The default threshold of four means the EA needs roughly two-thirds of its voters to align before risking capital, so a single noisy indicator cannot drive a position on its own.

The five-state regime classifier runs every tick and is independent of the signal stack. It reads the same three handles — ADX(14), ATR(14), and Bollinger Bands(20, 2.0) — and walks a 100-bar lookback to compute two percentile features: where the current ATR sits in its own recent range, and where the current Bollinger width sits in its own recent range. StrongTrend is the state where ADX ≥ 28 and the ATR percentile is at or above the 70th percentile; WeakTrend covers the ADX-≥-20 band without the high-ATR confirmation; Range is the compressed state when Bollinger width is in the bottom 30% of its own distribution; Breakout fires only when both the ATR and the Bollinger width percentiles are at or above 85%; and everything else lands in Choppy. The classifier blocks entries in Choppy and Range, allows them in WeakTrend, StrongTrend, and Breakout, and uses the live state to force-close positions the moment it rolls back into Choppy or Range. That asymmetry — letting the regime deteriorate open positions even when the signal is still valid — is the EA's built-in defence against a momentum setup quietly running into chop.

The HTF confirmation layer reads a single EMA(50) on a higher timeframe (default M15) and refuses to fire a long unless price is on or above the line, and a short unless price is on or below it. The default InpRequireHTF = true means that, out of the box, the EA will not chase a long when the higher timeframe trend is clearly down. Spread is gated twice — once in ConfirmFilters and once in NoTradeGate — at InpMaxSpreadPts of 30 points, which is intentionally tight for an XAUUSD scalper and pushes the operator towards a low-spread ECN or RAW broker. The R:R floor at InpMinRR of 1.2 is checked against the projected SL/TP distance, so any setup whose implied reward-to-risk would fall below 1.2 is rejected before PlaceEntry is reached. Risk is sized at InpRiskPercent of 0.5% of effective capital per trade, the lowest of the Momentum-Scalper family, with EffectiveCapital() returning min(InpCapitalCapAmount, equity) when a capital cap is set and raw equity otherwise. The 0.01 fixed-lot fallback never appears in the code — InpAutoLot is true by default and risk sizing is the only path. A capital-floor of $50 (InpCapitalCapFloor) blocks entries below that equity threshold.

Position management is a three-stage cascade that runs on every tick through ManageOpenPositions. When unrealised profit reaches 1.0R (where R is the initial ATR-anchored stop distance), the stop is ratcheted to entry + 1 point for longs (mirror for shorts) — a one-shot break-even that the function will not repeat on the same ticket. At the same 1.0R threshold, but only if remaining volume is greater than 1.9× the broker's minimum lot, half the position is closed via TryClosePartial_EX10011 and the comment tag becomes the EA's standard trade comment. After that, an ATR trailing stop at InpATRMulTrail of 2.0× ATR(14) tightens the stop every tick in the favourable direction, gated by the broker's SYMBOL_TRADE_STOPS_LEVEL so the broker never rejects a stop move for minimum-distance violation. Exits other than the cascade are: an opposite-momentum signal (the same 6-vote check, evaluated for the trade's opposite direction, on a closed bar), the regime-collapse force-close into Choppy or Range, a maximum-hold of 240 bars (four hours on M1, twenty hours on M5, sixty hours on M15), and the live no-trade gate for daily/weekly loss caps and the cooldown timer.

The no-trade gate is a stack of independent checks, all of which must pass for PlaceEntry to be reached: kill switch off, session open (London 8–12 server, New York 13–17 server, weekends closed, Fridays at or after 18:00 closed), trades-today under InpMaxTradesPerDay of 10, no active 60-minute cooldown after three consecutive losing closes, equity above the $50 floor, daily realised PnL not more negative than 3% of effective capital, weekly realised PnL not more negative than 9% of effective capital, regime not in Choppy or Range, spread at or under 30 points, and the symbol's SYMBOL_TRADE_MODE allowing trades. Daily and weekly realised PnL are walked directly from the deal history filtered by magic number, so the counters are accurate even after a terminal restart. The session window of 4h London plus 4h New York with the weekend and Friday-evening blocks is a deliberately conservative frame — the EA expects to find both flow and tight spreads in those eight hours, and the rest of the trading day is reserved for management of positions opened during the window.

InpDryRun ships as true, which means that on a fresh install every entry and every exit is logged but no order is sent to the broker — the operator must explicitly flip the flag to false before any live trade fires. InpKillSwitch is false by default and acts as a panic-button: when toggled to true the next OnTick call forces NoTradeGate to return true, the no-trade code path then exits, and any open position is closed by ManageOpenPositions on the same tick. Sends are hardened with a single retry on TRADE_RETCODE_REQUOTE, TRADE_RETCODE_PRICE_CHANGED, and TRADE_RETCODE_TIMEOUT after a fresh sym.RefreshRates(). Closes and partial closes are retried up to three times at 200ms intervals on the same four transient retcodes, and modifies are retried up to three times at 100ms intervals on requote and timeout. The filling mode is auto-picked from SYMBOL_FILLING_MODE in FOK > IOC > RETURN order, with the chosen mode then written via trade.SetTypeFillingBySymbol(_Symbol) to make the CTrade object aware of the choice. Slippage is hard-coded at DEV_POINTS of 20 for the trade object and ULONG_MAX for the closing helpers (the helper falls back to the symbol's default deviation). The OnTester custom criterion is (net * PF) / (1 + balance_dd) with a 30-trade hard floor and a dd <= 0 guard, which means the optimiser ranks parameter sets by composite quality rather than raw net profit, and zero-drawdown runs are penalised to discourage curve-fitted flatlines.

The single most consequential input is InpMinVotes. Tightening it to 5 produces a low-frequency, high-conviction run; loosening it to 3 produces a higher trade count with weaker consensus. InpRSILevel is the asymmetric tuning knob: 55 (the default) biases entries towards the higher-RSI side of momentum confirmation, lowering it to 50 produces more symmetric long/short behaviour, raising it to 60 reduces long entries to only the strongest RSI readings. InpRequireHTF is the safety switch — when set to false the EA will fire in both directions regardless of the higher timeframe, and that mode is only sensible for traders who actively want counter-trend scalps. InpMaxOpenTrades defaults to 1, which means the EA never pyramids and never hedges — a second entry attempt on the same symbol is blocked at the position-count check, not by the signal. InpATRMulTP of 2.5 vs InpATRMulSL of 1.5 produces a default TP:SL of 1.67:1, comfortably above the 1.2 R:R floor in InpMinRR, and the 2.0× ATR trailing multiplier in InpATRMulTrail is what converts the partial-close stage into a longer-run trade without giving back the early profit.

When backtesting, expect the EA to behave like a momentum-then-manage system: a small number of high-conviction entries during the eight server-hour trading window, a 50% partial close at 1.0R on the majority of winners, and the trailing stop doing most of the work on the remainder. The 240-bar max-hold exit means that on the M1 chart a trade cannot sit for more than four hours; on the M5 chart, twenty hours; on the M15 chart, sixty hours. The deal-history-walked daily and weekly loss caps mean that the optimiser will see real drawdown gating, not phantom blocking, and the 30-trade OnTester floor combined with the dd <= 0 guard means a parameter set with fewer than 30 trades or zero drawdown scores exactly 0. The strategy is rated HIGH risk in the corpus because the 0.5% per-trade risk, the 3% daily cap, the 9% weekly cap, and the 1.67:1 R:R still allow significant account-level drawdown on a streak of failed momentum breaks, especially on a broker that widens spreads during the news windows the EA does not explicitly block.