Pipsgrowth EX18062 TrendFollow

Pipsgrowth EX18062 TrendFollow (HullTest_Fast) is a multi-timeframe Hull Moving Average consensus EA built around an unusual premise: instead of trusting a single Hull-MA on one chart, it asks three independent Hull calculations on three different timeframes to vote on direction, and only fires when a configurable majority agrees. The three H1/M30/M15 Hull periods are 300, 200, and 130 respectively, which is far longer than the typical 20/55 or 34/89 windows used elsewhere in this family. Those long periods are deliberate — the EA is not trying to scalp or catch the early stages of a swing. It is designed to ride the body of trends that have already been running for hours on the H1 chart and to filter out almost everything that does not align with that body.

The signal path is a five-stage filter. First, each timeframe computes its own Hull value inline from LWMA primitives — CalcHullValue() builds the standard 2×WMA(period/2) − WMA(period) formula and the EA never calls iCustom. The slope check is strict: the EA requires three consecutive rising or falling Hull values plus a closed bar that closes on the correct side of the Hull line. If that test fails on any of the three timeframes, that timeframe abstains. Second, each timeframe's vote is independently gated by its own ADX(14) reading using a 24.0 threshold (InpHullAdxThresh). Timeframes whose ADX is below 24 are silenced — they abstain instead of voting against. Third, a MinHullsAgree of 2 of 3 (the default) is required for the consensus to fire; a 1-1-1 split yields no trade. Fourth, the H1 EMA(50) trend direction is fetched through HtfEmaSide() and the EA refuses to take a signal that disagrees with the EMA — for example, a long vote on the lower timeframes is killed if the H1 EMA is sloping down with price below it. Fifth, the regime classifier has to approve: STRONG_TREND, WEAK_TREND, BREAKOUT, and EXPAND are all permitted, but CHOPPY, RANGE, and COMPRESS block entries.

Regime classification itself uses three M15 indicators: an ATR(14) with a 100-bar percentile lookback, Bollinger Bands(20, 2.0) with a 100-bar width percentile, and an ADX(14) for trend strength. The thresholds are split out in the code: ADX ≥30 with ATR pctile ≥70 reads as STRONG_TREND, ADX ≥22 alone is WEAK_TREND, width percentile ≤20 is COMPRESS, width percentile ≥80 combined with ATR pctile ≥60 is BREAKOUT, ATR pctile ≥70 without an ADX spike is EXPAND, and ADX <18 with ATR pctile <40 is RANGE. Anything that falls through all the buckets ends up as CHOPPY. The fact that COMPRESS and RANGE are blocked alongside CHOPPY means the EA sits on its hands during low-volatility consolidations and during quiet ranging hours — the only environments it really likes are persistent trends and the expansion phase that comes out of a quiet base.

The risk module is conservative for a 130/200/300-period trend system. Default risk per trade is 0.75% of effective capital, daily loss is capped at 3.0%, weekly at 6.0%, and an effective-capital floor of $50 blocks new entries when the account is depleted. InpCapitalCapAmount is 0.0 by default — the cap is opt-in. The EA also has a hard cooldown: after 3 consecutive losses, new entries are blocked until a profitable close resets the counter. The session filter is one of the most distinctive choices in the file: IsSessionAllowed() returns true only between broker hours 0 and 12, which targets the Asian and London sessions and skips the entire New York window. This is paired with an end-of-day close at 20:00 broker time (EOD_HOUR) that liquidates everything before the rollover, including positions that have not hit their stops or targets. The 0-12 window means the EA does not see US economic releases at all — that is by design, not a limitation.

Trade management runs the standard toolbox. A one-shot break-even at 25 points of profit moves the stop to entry + 5 points. A 50% partial close at 1R (PARTIAL_TP1_R_MULT=1.0, PARTIAL_TP1_PCT=50) banked half the position size once the trade reaches the SL distance. From that point, a 3.0×ATR(14) trailing stop activates and ratchets forward, gated to only move in the favorable direction. A volatility-spike exit is layered on top: when current ATR is more than 2.2× the 50-bar ATR average, the EA closes the position regardless of whether the spike is in the trade's favor. The fixed take-profit is 180 points with a 40-point stop, which is 4.5:1 — the widest fixed R:R in this family. The InpMinRR=1.2 sanity gate just confirms the configured TP/SL ratio is reasonable, not that the trade itself hit 1.2R. The opposite-signal exit is built into CloseOpposite() and fires before a new entry is opened in the reverse direction.

Execution is broker-portable. All seven indicator handles (three ADX TFs, ATR, regime ADX, BB, HTF EMA) are created in OnInit, and SetDeviationInPoints(30) plus SetTypeFillingBySymbol() are called up front. Three retry helpers — TryClose_EX18062, TryClosePartial_EX18062, and TryModify_EX18062 — handle requotes, timeouts, and price changes with 200ms and 100ms backoffs before giving up. Slippage tolerance is 30 points. OrderCalcMargin() is called before every entry, and the order is refused if required margin exceeds free margin. By default, InpDryRun=true — the EA logs the trades it would have taken but does not send them; the user has to flip the switch to go live.

For backtest selection, OnTester() returns (net × pf) / (1 + dd) with a 30-trade minimum, which strongly rewards strategies with high profit factor and low relative drawdown. Inputs that matter for tuning: the three Hull periods (raise them to make the EA slower, lower them to make it faster), InpHullAdxThresh (raise it to demand stronger trends per timeframe), InpMinHullsAgree (raise to 3 to require unanimous consensus, drop to 1 to weaken the filter), the SL and TP points (must keep TP/SL ≥1.2 per InpMinRR), and the session window defined by SESSION_START_HOUR=0 and SESSION_END_HOUR=12 in the define block. Be aware that the vol-spike exit, partial close, BE, and trail all stack on the same position — the partial close runs at 1R, the BE at 25 points, and the trail starts once profit exceeds 1.0× the SL distance. On a 40-point SL, the trail begins around 40 points of profit, which is the same point the partial fires; the trail takes over managing the remaining half.

A practical note on the H1 EMA filter: the EA only blocks signals that disagree with the EMA — a long vote with the H1 EMA flat is allowed through. If the operator wants a stricter filter that blocks flat-EMA signals too, that has to be patched into HtfEmaSide() directly. Also, the M15 signal operates on closed bars only, so the EA does not repaint on intra-bar ticks. The default InpDryRun=true state means the chart will show trade logs but no positions; flip it to false to go live.