Pipsgrowth EX18064 TrendFollow

Pipsgrowth EX18064 is a fully-discretionary Ichimoku Kinko Hyo system that waits until all four canonical Ichimoku conditions align on a closed bar before it fires. The four conditions, in the order they are evaluated inside SignalIchimoku(), are: (1) Tenkan-sen (period 9) is on the correct side of Kijun-sen (period 26) — Tenkan above for longs, below for shorts; (2) the most recent closed price sits above the upper kumo boundary (Senkou Span A and Span B maximum) for longs, or below the lower boundary for shorts, with a one-bar tolerance so that a bar that just exited the cloud still counts; (3) the Chikou Span — the close from 26 bars back, which is the displaced reading MT5 produces by default — is itself clear of the kumo at the same historical point, which filters out signals where a recent pullback would still be tangled with the cloud; (4) the future cloud 26 bars ahead, computed from Senkou A and Senkou B at index 1+InpKijun, is green for longs or red for shorts. All four must agree; partial consensus is recorded as a confidence score (0–100) but does not open a trade. This is the strictest application of the Ichimoku system in the TrendFollow family — there is no averaging, voting, or scoring fallback.

The signal is gated by a 7-state regime classifier in ClassifyRegime(). It reads ADX(14) main, ATR(14) as a percentile rank over the last 200 bars, and Bollinger Band width(20, 2.0). The thresholds are tunable through #define constants: ADX at or above 35 with an ATR percentile at or above 0.75 returns REG_EXPAND, ADX at or above the InpAdxThreshold input (default 22) returns REG_STRONG_TREND, ADX at or above 20 returns REG_WEAK_TREND, ATR percentile at or below 0.30 returns REG_COMPRESS, and ADX below 18 returns REG_CHOPPY. Of these seven states, only four permit entries: STRONG_TREND, WEAK_TREND, EXPAND, and BREAKOUT. A flip into CHOPPY during an open position is treated as a full-book close signal by ManageOpen() — the EA does not wait for the stop to be hit, it exits at market.

Two HTF filters run in ConfirmEntries(). The first checks that the close on the higher timeframe (default H4 — InpHtfTimeframe=9 maps to PERIOD_H4 via the same MapTimeframeInt switch) is on the correct side of an EMA(50) computed on that HTF, with a small ±0.3% slack to forgive near-edge price. The second reads the Tenkan/Kijun cross on the HTF Ichimoku and requires the same direction. A failure on either filter is logged and the trade is rejected for that bar. The third guard, IsHtfOpposing(), runs separately and blocks entries when the HTF EMA disagrees by more than 0.3%. Spread is filtered by IsSpreadBad() against both an absolute cap (InpMaxSpreadPoints, default 35 points) and a rolling 50-bar average — a current spread more than 2× the recent average rejects the entry even if it is under the absolute cap.

Risk is sized from effective capital, where effective capital is the live account equity minus realized losses today, optionally capped at InpCapitalCapAmount. Each trade risks InpRiskPercent (0.5%) of that effective capital. The stop distance is InpAtrMultSL × ATR(14), default 2.0 × ATR, which lands a typical XAUUSD H1 SL in the 1.5–2.5 USD range. TP1 sits at 1R and TP2 at 2R; the order is sent with TP2 as the broker-side target. In ManageOpen(), when the trade reaches 1R the position is partially closed for InpPartialPctR (default 50%) via TryClosePartial_EX18064 — this is a one-shot partial that does not repeat. At the same 1R threshold the stop is moved to break-even with a 10-point offset (BE_OFFSET_POINTS) — long stops move to entry + 10 points, short stops move to entry − 10 points. From that point on, a forward-only ATR trail at 2.5 × ATR(14) (InpTrailAtrMult) ratchets the stop every tick; the ratchet only moves in the favorable direction. All three retry helpers — TryClose_EX18064, TryClosePartial_EX18064, TryModify_EX18064 — use a 3-attempt loop with 200ms backoff on requote/timeout/price-changed conditions.

Three extra exits are layered on top of the trailing logic. First, a max-hold time exit at MAX_HOLD_BARS=500 (the time exit uses Bars() between the open time and the current bar) closes the trade without waiting for a stop or target hit. Second, an opposite-signal exit re-runs SignalIchimoku() on every tick and, if the consensus flips, closes the trade at market. Third, the regime-change exit described above. The loss-cooldown is implemented through g_lastLossBarTime plus the COOLDOWN_BARS=3 constant — after a losing close, the next three bars on the working timeframe are blocked from entry. Daily loss is capped at 3% of effective capital and weekly loss at 6%; both are computed from closed deal history (RealizedSince(dayStart) and RealizedSince(weekStart)). InpMaxConcurrent defaults to 2 and the symbol-level cap MAX_SYMBOL_EXPOSURE also defaults to 2, so the EA will not stack more than two XAUUSD positions at a time. The OnTester() fitness function returns (net * pf) / (1 + dd), with a 30-trade floor — backtests under 30 trades or with non-positive drawdown score 0, so the optimizer will not pick degenerate parameter combinations.

In practical terms this EA is built for a trader who already trusts Ichimoku as a complete system and wants the discipline of never breaking its rules. The 2R target is realistic for a trend continuation on H1, the 1R partial turns the win rate into something measurable, and the regime filter keeps the EA out of the sideways conditions where Ichimoku whipsaws badly. The 500-bar max-hold is generous (≈3 weeks on H1) but it acts as a circuit breaker for trades that simply never develop. Like all of the TrendFollow EAs in the Pipsgrowth library, the file defaults to InpDryRun=true so a new installation does not place real orders until you have reviewed the inputs and the EA's behavior in the Strategy Tester. The order comment is hard-coded to "Psgrowth.com Expert_18064" and the magic number is 22218064 — both are required for the EA to recognize its own positions during manage/exit passes.