Pipsgrowth EX04004 HAEA

Pipsgrowth EX04004 HAEA is a bar-by-bar trend-following EA for XAUUSD on the 5-minute chart. It looks for two things to line up on the same closed bar: a clean Heiken Ashi candle color and price sitting on the correct side of a slow exponential moving average. When both confirm in the same direction, the EA fires one market order and steps out of the market until that position is closed. There is no scaling in, no martingale, no re-entry on the next tick — the next decision waits for the next bar.

The signal is built from raw OHLC, not from a custom indicator. On every new bar the EA pulls the open, high, low, and close of the just-closed bar (index 1) plus the open and close of the bar before that (index 2). It then computes two smoothed Heiken Ashi values inline: haOpen = (open[2] + close[2]) / 2 and haClose = (open[1] + high[1] + low[1] + close[1]) / 4. A green HA candle is haClose > haOpen; a red HA candle is haClose < haOpen. That colour is the momentum vote. A 200-period exponential moving average on close prices, requested via a single iMA handle during OnInit, supplies the trend vote. A long entry needs a green HA bar with the actual close of bar 1 sitting above the EMA; a short entry needs a red HA bar with the close below the EMA. The bar's true close is the price being compared — not the smoothed HA close — so the trend filter is anchored to the real market, not to the smoothed candle.

Risk is sized off the stop loss, not off a fixed lot. CalculateLotSize() divides the effective capital by the configured InpRiskPercent (default 1.0%) and translates that dollar risk into a lot count using the symbol's tick value, tick size, and point. The result is floored to the symbol's lot step and clamped to the broker's minimum and maximum lot. If the trader sets InpRiskPercent to 0, the EA falls back to InpFixedLot (0.01). Effective capital is normally the account equity; if InpCapitalCapAmount is non-zero, the EA sizes off the smaller of equity and that cap, which lets a user pre-commit a fixed dollar amount to the strategy. Before any order is sent, OrderCalcMargin checks that the required margin fits inside free margin; if it does not, the order is dropped silently.

The stop loss is 400 points and the take profit is 800 points — a fixed 1:2 reward-to-risk ratio. The EA does not move the stop to break-even, does not trail it, and does not take partial closes. The TryClose, TryClosePartial, and TryModify helpers in the source wrap trade.PositionClose, trade.PositionClosePartial, and trade.PositionModify with three retries spaced 200ms apart on REQUOTE, TIMEOUT, PRICE_OFF, and PRICE_CHANGED return codes, but those helpers are not actually wired into the bar logic — the trade management is just SL and TP at open. A position that survives the SL/TP and stays in the market can only exit when price hits one of those two levels or the broker forcibly closes it.

The position accounting is one trade at a time. CheckSignal() short-circuits and returns immediately if PositionsTotal() > 0. There is no basket logic, no opposite-signal close, and no pending-order queue. When the current position closes, the EA waits for the next bar to form before it re-evaluates — the static datetime lastBar and iTime check in OnTick enforce this. That is by design: the EA is a swing-style trend follower, not a scalper. Trades that close at a loss feed OnTradeTransaction, which increments a running counter of consecutive losses; the moment that counter hits 3 (the default for InpMaxConsecLosses), the EA sets g_cooldownUntil = TimeCurrent() + InpCooldownMinutes * 60 and refuses any new entries for the next 30 minutes. A winning trade resets the counter to zero, so a single green bar breaks the streak.

The capital-protection stack is six layers deep and runs on every tick. IsSafeToTrade() first checks the optional capital cap and floor (InpCapitalCapAmount and InpCapitalCapFloor); then an equity floor at 95% of the initial balance; then a 3% daily realised-loss cap and a 6% weekly realised-loss cap, both expressed as a percentage of effective capital; then the consec-loss cooldown; then a hard cap of 50 trades per day; then IsMarketOpen() which blocks Saturday and the first 22 hours of Sunday plus the last hour of Friday in GMT. On top of that, InActiveSession() only allows trading during London 7–16 GMT or New York 12–21 GMT, and the InpAvoidAsia flag is on by default, so the Asian session (0–7 GMT) is excluded entirely. IsNewsTime() watches the start of London and New York and can pause entries for 15 minutes around those opens, but InpNewsFilter is off by default — the trader opts in.

The 200 EMA is the only indicator the EA actually attaches (hMA = iMA(_Symbol, _Period, InpTrendEMA, 0, MODE_EMA, PRICE_CLOSE)); Heiken Ashi values are calculated inline and never get their own handle. The session machinery is purely time-based: DetectGMTOffset() walks the last 100 H1 bars looking for the weekend gap, then picks the broker's GMT offset by testing each candidate in -12..+12 against the Friday 22:00 close. InpServerGMTOffset lets the trader override the auto-detection with a manual hour offset if the heuristic gets it wrong. Filling mode is also broker-driven: the EA checks SYMBOL_FILLING_MODE at init and uses FOK if the broker supports it, otherwise IOC, otherwise RETURN. Slippage is fixed at 3 points. New entries are placed with up to 3 retries on REQUOTE, TIMEOUT, or PRICE_OFF separated by 150ms each.

OnTester returns (net × profit factor) / (1 + relative balance drawdown), gated on a minimum of 30 closed trades in the optimisation pass. Below 30 trades the fitness is 0, which keeps the optimiser from selecting results on tiny samples. The result is a single-number score that rewards profit and consistency while penalising drawdown.

Backtest expectations: with the defaults (1% risk, 400-point stop, 800-point target, 1 trade at a time, new-bar only, 95% equity floor, 3% daily loss cap, 30-minute cooldown after 3 losses), the EA will produce a low-to-moderate number of trades per day on XAUUSD M5 because it has to wait for HA color to agree with the 200 EMA on a closed bar, and it is only allowed to look during the London or New York sessions. The fixed 1:2 RR means winners are double the size of losers in points, but the win rate has to clear roughly 34% before expectancy turns positive after spread and slippage. Set InpRiskPercent higher and the absolute P&L scales up; set it to 0 to switch to fixed 0.01 lots, which is the safer mode for forward testing. The 200 EMA in particular should be checked against the broker's data — if the broker truncates history, the EMA will be colder and the trend vote will be wrong for the first few weeks of trading.