Pipsgrowth EX14020 Scalper

Pipsgrowth EX14020 AIXAUUSDScalper is a single-bar vote-counter for XAUUSD. Each closed M5 bar is fed through four independent filters — a 14-period RSI, a 12/26/9 MACD line-vs-signal comparison, a 50-period EMA, and a 200-period EMA — and each filter that agrees with the bullish direction adds twenty-five points to a buy score. The mirror set of conditions feeds a parallel sell score. When either score reaches the configurable confidence threshold (default InpAIConfidence = 75, i.e. three out of four filters in agreement), the EA fires a market order in that direction.

The decision math is the entire strategy. There is no position-management subroutine. Once a trade is on, nothing else happens to it: no trailing stop, no break-even migration, no partial close, no reverse-signal exit, no time-based exit. The fixed 30-point stop and 50-point target attached at order entry are the only way a position can leave the book, and the only thing the EA does to manage exposure is count open positions through CountTrades() and refuse to open a fourth if three are already live under magic 22214020.

The 200-period EMA is the structural filter that gives the system its directional bias. The buy branch is mathematically reachable only when bid > ema200[0], and the sell branch only when bid < ema200[0]. That single condition forces every entry to be on the same side as the long-term trend, which is the only thing keeping the score from routinely firing both ways on the same bar. In practice the two branches are still evaluated independently — there is no veto that would explicitly cancel a buy when sellScore has also crossed threshold — so during the bar where price crosses the 200-EMA, both branches can resolve true. The OnTick logic does not guard against that; the broker may simply reject the second order, or the spread gate may catch it on the next tick.

The ATR handle (hATR, 14-period) is created in OnInit and released in OnDeinit, but its value is fetched inside OnTick and never used. The indicator exists in the source purely as a placeholder in the '5 indicator handles' stack claimed by the source header; in the live code path the ATR buffer copy is read into a local variable and discarded. The 12-layer architecture claim in the header comment ('REGIME, SIGNAL, ENTRY, CONFIRM, NO-TRADE, CAPITAL CAP, RISK, SIZING, MANAGE, EXIT, SCALING, OnTester') is marketing scaffolding — the runtime implementation is a single-pass score, a single guard pair (spread and trade count), a fixed lot input, fixed stops, and OnInit/OnDeinit housekeeping. There is no OnTester routine, no regime detector, no news filter, no session gate, no day-of-week guard, and no drawdown cap.

The three retry helpers TryClose_EX14020, TryClosePartial_EX14020, and TryModify_EX14020 are declared in the source but never invoked from OnTick. They follow the standard 3-attempt pattern with 200ms close / 100ms modify backoff and treat TRADE_RETCODE_REQUOTE, TRADE_RETCODE_TIMEOUT, TRADE_RETCODE_PRICE_OFF, and TRADE_RETCODE_PRICE_CHANGED as retryable conditions, but the live entry path uses the m_trade object's Buy() and Sell() methods directly, so the only failures a real user will see are raw trade-result codes from a single attempt. OnInit contains an OnInit_20260007_placeholder() stub that exists solely to satisfy the unused-function warning — it returns 0 and is never called.

The RSI thresholds sit at 40 and 60 — closer to the midline than the textbook 30/70 oversold/overbought levels. That choice lets the system act on mild momentum shifts rather than extreme reversals, but it also widens the rate at which the score reaches threshold, which is the entire point of the configurable confidence input: tightening InpAIConfidence up to 100 effectively requires all four filters to agree (a much stricter bar), and dropping it to 50 forces only two filters to align. With the default 75, the system typically needs three of four to agree, which in real XAUUSD M5 data means entries cluster around bars where the 200-EMA bias, the 50-EMA proximity, the MACD direction, and the RSI position all line up within a single five-minute window.

The fixed 30/50 economics in pips (the EA operates in raw _Point units, so the user must calibrate the lot against their broker's point value) make this a high-frequency, low-per-trade-risk design: 0.01 lot on XAUUSD with a 30-point stop is roughly three dollars of risk per trade on a 5-decimal gold quote, and three concurrent positions max out the per-trade risk at around nine dollars per closed-bar burst. The cap on concurrent positions is enforced by CountTrades() on every tick, so the EA will keep stacking entries bar after bar up to the cap, but each new entry is independently scored — there is no averaging, no grid, no martingale, and no recovery multiplier. If all three positions hit their stops, the loss is three times the per-trade stop, full stop.

There is no protection against the classic 200-EMA whip-saw. When gold chops around the daily 200-EMA, the score will keep flipping between buy and sell branches on consecutive bars, generating paired entries in opposite directions that both get stopped. The M5 timeframe and the 3-position cap make this relatively bounded, but it is a structural property of the design: the system follows the trend it sees on each bar, with no memory of what the previous bar said, no concept of trend strength beyond the four indicators, and no mechanism to step aside during a regime where the trend signal is unreliable.

Run it on a low-spread ECN or RAW-spread account where the typical XAUUSD spread is comfortably under the 30-point cap (the InpMaxSpread = 30 default is a generous ceiling for gold — most retail brokers sit at 20-30 points normally and widen to 50+ around news). Test in the strategy tester on real-tick data for at least two years, walk it through a known high-volatility period (March 2020, March 2023 banking stress, August 2024 carry unwind) to verify the score behaves under regime change, and be aware that the fixed lot and fixed stop inputs leave no room for position-sizing adaptation — if the account equity is small, the same 0.01 / 30-point exposure is the only option the EA offers.