Pipsgrowth EX18063 TrendFollow

Pipsgrowth EX18063 TrendFollow is a three-leg Hull moving-average consensus strategy on M15, with M30 and H1 acting as confirmation telescopes. Each leg carries its own ADX threshold, and the EA only fires when at least two of the three timeframes slope in the same direction with no opposition vote. The whole system is built around the idea that a single Hull cross on a fast chart is noise, but a fast chart, a medium chart, and a slow chart all leaning the same way is information worth trading.

The Hull values are computed in-line from raw close arrays rather than pulled from a child indicator. CalcLWMA builds a linear-weighted average of N consecutive closes with weights descending from period to 1, and CalcHullValue returns 2*LWMA(period/2) - LWMA(period). HullDirection then samples that result at offsets 0, 1, and 2 on the closed-bar series and demands a monotonic slope across two bars and a close beyond the current Hull value before declaring +1 or -1. The per-timeframe ADX gate sits inside HullDirection, so a chart that is technically rising but lacks the trend strength to clear its floor is treated as a non-vote rather than a vote. The M15 fast leg has a 24 ADX minimum, the M30 medium leg 28, and the H1 slow leg 30; raising any of those floors is the cleanest way to dial selectivity.

With InpMinHullsAgree=2 the EA only trades when the count of agreement votes is at least 2 and the count of opposite votes is 0. A clean long setup needs M15 and M30 both returning +1 (with H1 neutral or agreeing), and a short needs the inverse pattern. The asymmetric rule is deliberate: a +1/+1/0 signal is allowed, but a +1/0/-1 signal is rejected. That veto-on-opposition clause is what filters out the period after a fast chart flips bullish while the dominant chart is still in a downtrend, which is the single most common source of false breakouts on M15.

Confirmation runs after the vote, not before. ConfirmBlock checks four gates: H1 EMA(50) agrees with the direction and is itself sloping in the same way, the ATR-percentile sits inside the 35-92 band (excluding both compressed and climactic volatility), server time is in Asia, London, or New York, and the planned R-multiple of TP2 over the stop is at least 1.5. InpMinConfirms=2 means a trade can fire when only two of the four gates are clear, and in practice the session and R-multiple gates pass almost trivially while the HTF EMA gate is the one that does the real filtering against counter-trend signals.

The regime classifier pulls from three different sources: H1 ADX for trend strength, M15 ATR(14) percentile over a 50-bar window for volatility context, and 20-period Bollinger bandwidth versus its own rolling average. The output is one of seven states. StrongTrend is ADX >= 30 and ATR-percentile >= 60, WeakTrend is ADX >= 22 and ATR-percentile >= 40, Breakout is BB width above 1.4x its rolling average with ATR-percentile >= 70, Compress is BB width below 0.7x its average, Expand is ATR-percentile >= 80, Choppy is ADX < 18, and Range is the catch-all. Choppy is the only state that blocks entries outright; the others modulate signal quality but do not veto the trade on their own.

Risk and the no-trade gate form the standard safety stack. Risk per trade is 0.75% of effective capital, which sits above the 0.5% floor that most of the TrendFollow family uses; position size is sized against the ATR(14) stop distance through CalcLotByRisk and clamped to the broker's lot step and limits via NormVol. Daily loss limit is 3% and weekly is 9%, with the weekly cap deliberately looser than the 6% that other EAs in the corpus tend to enforce. The capital cap is opt-in: InpCapitalCapAmount=0 disables it, but when set the EA uses min(cap, equity) as effective capital and refuses to take new entries below a $50 floor. Cooldown fires after 2 consecutive losses and lasts 3 M15 bars. Friday at hour 20 server time forces a close of every open position, Saturday and Sunday are no-trade, and the spread filter compares the current spread to a rolling average with a 3.0x multiplier.

Position management is the most configurable block on the EA. Break-even triggers at profitR >= 1.0 and ratchets the stop to open + 5 points (open - 5 for sells); the trailing stop kicks in at profitR >= 0.8, which is unusually early in the cycle, and walks the stop by 2.0x ATR(14) per new bar. The 50% partial close fires at 1.0R. Three independent exit triggers run in parallel and can cut the trade before TP2 is reached: a stochastic K% cross above 80 on longs or below 20 on shorts (with a 15-point minimum profit floor), an ATR spike above 2.2x its 50-bar average, and a reversal signal where the consensus flips to the opposite direction. The full TP2 at 3.0R is sent on the entry ticket; the trail and the three exit triggers can close the position before it is hit.

Pyramiding is enabled by default and capped at 3 legs in the same direction. The pyramid gate requires every existing same-direction position to be at least 15 points in profit before another leg can be added, which means the EA only adds to winners that are already moving. OrderCalcMargin is checked before each send, slippage is capped at 30 points, and SetTypeFillingBySymbol handles the filling-mode question automatically. InpDryRun=true is the default, which means the EA logs hypothetical orders until you turn it off in the inputs; InpKillSwitch is the panic button that stops new entries without removing the EA from the chart.

For deployment, the EA is tuned for XAUUSD M15 but the M30 and H1 cross-checks mean the strategy behaves best on instruments with sustained directional trends rather than the kind of chop that some FX crosses deliver. A 10-trade cold streak at maximum loss lands at -7.5% of effective capital, well inside the 9% weekly cap, and the 3-bar cooldown after 2 losses is short enough to keep the EA trading but long enough to avoid the revenge-trade pattern. Expect a low-trade-frequency trend follower with selective entries - two or three trades per week on XAUUSD M15 is a realistic baseline, with most of the equity curve coming from a small number of medium-duration trends rather than a high count of small wins.