Pipsgrowth EX05002 Hedge

EX05002 is a single-position, single-hedge cover EA. It does not run a basket of opposite legs or stack scaling orders into a trend — every entry is exactly one seed trade, and at most one opposing hedge is attached to that seed. The hedge fires only after the seed is already in a loss, only on losing positions the EA still owns, and only once per ticket. There is no grid logic, no martingale, and no per-layer lot scaling. The EA's only job is to enter on a confirmed momentum cross, manage that one trade through break-even, partial close and trailing, and then — if price keeps running against the seed — open a single protective cover that bounds the loss while the original thesis plays out.

The seed is generated by an EMA(8)/EMA(21) crossover on the closed bar. The cross alone is not enough: ComputeSignal() also computes a confidence number from the relative spread between the two EMAs (sep * 30 + 40, capped at 100), and HtfAgrees() confirms the direction against an H1 EMA(50) filter — longs only above EMA50, shorts only below it. The regime classifier (ClassifyRegime()) runs on every tick using ADX(14), ATR(14) percentile over a 200-bar lookback, and a Bollinger Band(20, 2.0) width proxy, and resolves the market into one of seven buckets: COMPRESS, RANGE, CHOPPY, WEAKTREND, STRONGTREND, EXPAND and BREAKOUT. The EA seeds only inside RANGE, COMPRESS or WEAKTREND — strong trends and breakouts are deliberately excluded so the protective hedge has room to do its job without fighting momentum on both sides at once. The CHOPPY bucket hard-blocks all new entries via NoTradeGated().

Risk and sizing are split into a per-trade percent and a portfolio cap. CalcLot() sizes the seed at 0.5% of effective capital by default (InpRiskPercent = 0.5), where effective capital is min(cap, equity) – realized_today. The SL distance is 2.0 × ATR(14) and the lot is derived from that distance and the symbol's tick value/size, then NormalizeVolume() clamps to the broker's volume step, min and max. OrderCalcMargin() pre-checks margin before the order is sent, and the CTrade request is retried up to three times on REQUOTE, PRICE_OFF, PRICE_CHANGED or TIMEOUT with a 150 ms sleep between attempts. Filling is selected automatically: FOK if the symbol allows it, then IOC, then RETURN.

Once a seed is open, the EA runs the full management stack in ManageExits() on every tick. At +1R, it ratchets the stop to break-even plus one point (one-shot, irreversible). At the same +1R threshold, it also closes 50% of the position as a partial TP — the partial close is skipped only if the remaining volume would fall below 2× the broker's minimum lot. After +1R, an ATR(14) × 2.5 trailing stop takes over and can only ratchet forward. The exit logic also watches for an opposite EMA cross with confidence >= 50, and closes the position if the regime flips into STRONGTREND against the trade. A time-stop fires at 240 bars in trade (about 20 hours on M5) to prevent capital from being tied up in dormant positions.

The protective hedge is a separate subsystem driven by ManageHedge(). For every main position still open under the EA's magic, it measures the current adverse excursion: (open_px - bid) for longs, (ask - open_px) for shorts. When that distance exceeds InpHedgeLossATRMult * ATR(14) (default 1.5 × ATR), and the position is not already in g_hedged_tickets, and the global hedge count is below min(InpMaxHedges, 3), the EA opens a single opposing market order with volume = position.Volume() * min(InpHedgeLotMultiplier, 1.5) — i.e. the same size as the seed by default, hard-capped at 1.5× by the HEDGE_LOT_MULT_MAX define. The hedge is tagged with the comment ..._H, the seed's ticket is added to g_hedged_tickets so it can never be re-hedged, and g_last_hedge_time is set so subsequent hedges have to wait InpHedgeCooldownMin minutes (default 5). Hedges inherit no SL/TP from the seed — they live until the manual close or the basket is flattened by the equity stop.

The hardening stack is layered on top. IsMarketOpen() blocks Saturday and Friday after 21:00 server time, plus Sunday before 22:00. InActiveSession() translates server time to GMT using the auto-detected g_gmtOffset (or a manual InpServerGMTOffset), then admits bars only inside London (7–16 GMT) or New York (12–21 GMT), and explicitly rejects the Asian window (0–7 GMT) when InpAvoidAsia is true. IsNewsTime() blocks entries within ±15 minutes of the London or NY session open when InpNewsFilter is on. NoTradeGated() runs ten separate checks (choppy regime, spread > 250 points, session, news, daily trade cap, HTF disagreement, capital floor, daily loss, weekly loss, consec-loss cooldown, max concurrent) and writes a reason string to the log if any one of them fails.

The capital caps are enforced twice. On entry, EffectiveCapital() floors at the smaller of InpCapitalCapAmount (set to 0 means use full equity) and the current account equity, then subtracts today's realized P&L. The lot is sized off that effective figure, so after a losing day the EA automatically trades smaller. Daily loss at 3% of effective capital and weekly loss at 6% block all new entries; the realized figures are read straight from the deal history filtered by magic and symbol. A separate emergency check in OnTick() closes everything if equity drops below balance – (effective_capital × daily_loss_pct) — the equity stop.

The OnTester custom criterion is (net_profit * profit_factor) / (1 + |balance_dd|), gated on a minimum of 30 trades and a positive profit factor. That formula rewards both profitability per unit of drawdown and consistency across many trades — the kind of curve a single-strategy hedge-cover system can actually produce. A backtest that wins big on 12 trades and then gives it back will score poorly even if the absolute profit is high.

The minDeposit label in the catalog is $100, and the risk tier is HIGH — both correct. The EA needs enough headroom to cover one full seed plus one full hedge at the broker's lot step, and the protective cover is only useful if the account can hold both legs through a full ATR move against the seed. The 0.5% per-trade risk default assumes the operator is willing to take the seed to stop AND the hedge to stop on the same day, which is the worst realistic case for the design. Anyone running this should size for that, not for the seed alone.