Pipsgrowth EX08003 MACD

Pipsgrowth EX08003 MACD is built around a single, specific idea: trade the zero-line crossover of a fast MACD, but only when two slower MACDs agree on the prevailing direction. The fast pair is 5/8, the medium pair is 13/21, the slow pair is 34/144 — all three computed on the working timeframe (M5 by default) with a 1-bar signal line. The triple stack is what makes this EA different from a vanilla MACD cross. A buy fires only when the 34/144 MACD is above zero, the 13/21 MACD is above zero, and the 5/8 MACD has just crossed up through zero on the most recent closed bar. A sell mirrors that condition in the negative direction. Both checks use the bar 1 vs bar 2 readout from CopyBuffer so the signal cannot repaint on the live tick.

That gating sequence is encoded in the CheckSignal() function. The fast 5/8 MACD acts as a trigger, the 13/21 acts as a medium-term trend filter, and the 34/144 acts as a long-term regime filter. The trigger alone produces a constant stream of false signals; the two filters cut the trade rate to only the moments when all three layers are aligned. The EA also blocks any new signal until a fresh bar prints — the static datetime lastBar gate inside OnTick() means the same signal is not re-evaluated dozens of times per second.

Position sizing uses a percentage-of-equity model. CalculateLotSize() divides the risk percent (default 1.0% of effective capital) by the per-lot loss at the configured stop distance, then floors the result to the broker's lot step. The minimum is the symbol's LotsMin() and the maximum is LotsMax(). When InpRiskPercent is set to 0, the EA falls back to a fixed 0.01 lot. The effective capital itself is governed by the optional capital cap: when InpCapitalCapAmount is greater than 0, the EA treats the smaller of (account equity, cap amount) as the base for risk math, so a user who funds the account with $10,000 but sets cap to $2,000 will see lot sizes that match a $2,000 book, not a $10,000 book.

The trade is opened with a fixed stop and take-profit, not a structural swing stop. Default values are SL 400 points and TP 1200 points, a 1:3 risk-reward ratio. Order construction in OpenBuy() and OpenSell() calls OrderCalcMargin first and refuses the order if free margin cannot cover the lot size, then sends a market order through the CTrade wrapper with up to 3 retry attempts 150 ms apart. The comment field is fixed to "Psgrowth.com Expert_08003". The filling mode is negotiated at init: the EA probes the symbol's SYMBOL_FILLING_MODE, prefers FOK, falls back to IOC, then RETURN. This avoids the common 10014 error you get when the strategy assumes FOK on an IOC-only broker.

The trade management layer is a pure trailing stop — there is no break-even step, no partial close, no scaling out. ManageProfitLock() iterates the position list every tick and, for each position on the magic, lifts the stop-loss to Bid − InpTrailDistance (150 pts) once the position is more than InpTrailStart (300 pts) in profit. For sells the math mirrors with Ask. The check is one-sided: the new SL must be strictly above the existing SL for a buy (below for a sell) — there is no step-size gating, so the stop ratchets every tick the moment the threshold is crossed. Note that InpTrailStep is exposed in the input panel as 100 but is not actually consumed by the trailing code, so changing it has no effect on the live behaviour.

Pyramiding is the second scaling layer. When InpEnablePyramid is true and the same-side count is below InpMaxPositions (default 3), the EA will open an additional position on a fresh M1-cross signal. The gate is in IsGridSafe(): it scans the existing same-side positions and refuses to add unless the oldest qualifying position is at least InpMinProfitPoints (300 pts) in profit. That means the EA can stack up to three longs in a confirmed uptrend, but it will not blind-add to a losing leg. There is no martingale, no lot multiplier, no anti-martingale — every layer is sized independently through the same 1%-risk formula.

The hardening stack is unusually heavy for a single-symbol M5 EA. IsSafeToTrade() rejects new entries when: equity is below 95% of the original starting balance, the realised loss for the day exceeds 3% of effective capital, the realised loss for the week exceeds 6%, three consecutive losses have triggered a 30-minute cooldown, the trade counter for the day has hit 50, the broker server is in the weekend gap (closed from Friday 22:00 GMT through Sunday 22:00 GMT), the current GMT time is outside the active session windows, or the news filter is enabled and a session open is within InpNewsFilterMinutes (15). Session windows are London 7–16 GMT, New York 12–21 GMT, and Asian 0–7 GMT. The Asia block is on by default; the EA trades only when the time is inside London or New York, and it does not trade the Asian session by itself. GMT offset is auto-detected at init by DetectGMTOffset(): it scans H1 bars looking for the weekend gap, then computes the offset from the prior bar's hour. If the user knows the broker clock, InpServerGMTOffset overrides the auto-detect.

A practical note on the InpNewsFilter input: it ships disabled. The 15-minute buffer around the London open (07:00 GMT) and the New York open (12:00 GMT) is a low-cost toggle for traders who do not want to be in a fresh position into the NFP / CPI print. The default-off state reflects the EA's design assumption that the trend filter (34/144 MACD above zero) already absorbs most event-driven noise.

The optimiser fitness function in OnTester() is the standard Pipsgrowth composite: (net * profit_factor) / (1 + balance_dd_percent), with a hard minimum of 30 trades. This is a long-form trend-fitness metric: it rewards raw profit, weights it by profit factor (so high-win-rate low-payoff curves lose to a few large winners), and divides by drawdown to penalise wide equity swings. In a Strategy Tester pass on XAUUSD M5, expect a relatively low trade count (the triple-MACD filter is strict) with comparatively large average win size, because the EA only enters when all three MACD layers are aligned. The 1:3 fixed TP and the 150-point chandelier-style trail mean each trade has a generous upside but the EA will still take the full 1:3 payout whenever price reaches the TP before the trailing stop catches it.

For live use, the realistic deployment is XAUUSD on M5, on an ECN or raw-spread broker, with a $100 minimum balance (per the database field) — though a $300–$500 buffer is more honest because the 1% risk on $100 means a 400-point stop is roughly 0.01 lots, which is the floor on most gold accounts. The capital cap input makes this EA well suited to prop-firm style scaling: a trader who wants to risk only $1,000 of a $20,000 account can set InpCapitalCapAmount = 1000 and the lot sizes will compute against that $1,000 ceiling, not the full account. The 95% equity floor, the 3%/6% loss caps, and the 3-consecutive-loss cooldown work together as a drawdown firewall that should keep the EA from bleeding the account through a regime shift that the slow MACD has not yet adapted to.