Pipsgrowth EX17022 Volatility

Pipsgrowth EX17022 sits in the middle of the GB10_5_2 family — a stripped-down, single-purpose version of the Gaussian-band breakout framework. Where the heavier EX17020 ships with a 4-MA slope stack, an inline Heiken Ashi check, an eight-stage auto-filter, anti-intervention gates, a 3-phase market-close, and Fibonacci confluence scoring, EX17022 keeps only the core breakout engine plus three layers of trade management and one safety gate. The file itself is just under 400 lines, the entire strategy input set is ten parameters, and the EA touches three indicator handles in OnInit. That compactness is the design point: this is the EA you reach for when you want the Gaussian-band thesis — a smooth central line wrapped in a volatility-scaled envelope, broken only when candle structure agrees — without the multi-stage filtering, regime classification, or time-based session controls of its siblings.

The breakout core is unchanged from the rest of the family. In OnTick() the EA reads a 10-period EMA of close (maHandle = iMA(_Symbol, g_Timeframe, 10, 0, MODE_EMA, PRICE_CLOSE)) which acts as the Gaussian midline. It then reads a 14-period ATR (atrHandle = iATR(_Symbol, g_Timeframe, 14)) and computes an upper and lower band as gaussNow ± DistanceMultiplier × ATR. With the default DistanceMultiplier = 0.85, the band sits 85% of one ATR away from the smoothed midline on each side. A long signal requires price > upperBand while both pricePrev and pricePrev2 were still inside the band on the previous two M5 bars — a two-bar inside-then-outside break that prevents the EA from firing on the same bar that punches through. A short signal is the mirror: price < lowerBand with the two prior closes above the lower band.

Two filters gate the breakout before the order goes out. First, the HTF trend filter (emaHandle = iMA(_Symbol, g_TrendFilterTF, 50, 0, MODE_EMA, PRICE_CLOSE), default H4) — longs only fire if price > emaHTF, shorts only if price < emaHTF. This is a directional gate, not a slope check: the EA simply refuses to take long breakouts below the H4 50-EMA and short breakouts above it. Second, the slope of the Gaussian line itself — gaussNow > gaussPrev for longs, gaussNow < gaussPrev for shorts — ensures the midline is moving in the direction of the break, ruling out flat-EMA false breaks. A minimum break-distance of MinBreakDistance × ATR (default 0.3) means the close has to be at least 0.3 ATR beyond the band, not just barely on the wrong side. A 60% body gate (body / candleSize ≥ 0.6) rejects breakouts on dojis or pin bars where the close crosses the band but the candle is mostly wick.

The two refinements that distinguish EX17022 from a vanilla band-breakout bot are the auto-tuner and the dynamic lock-profit. AutoTuneFilters() looks at the current spread ((ask − bid) / _Point) and decides whether the symbol is a 4-decimal forex pair (spread < 3 points) or a 2-decimal metal pair like XAUUSD. For a 4-decimal symbol it sets pipFactor = 0.0001 and minATR = 0.0003; for a 2-decimal symbol pipFactor = 0.01 and minATR = 0.03. The effect is that the EA refuses to trade when ATR is smaller than three pips of the symbol's quote precision, which kills off a chunk of low-volatility chop on both XAUUSD and on EURUSD-style majors without you having to hand-tune the threshold. The UseAutoFilter input lets you turn this off and fall back to the static MinATR input instead.

The dynamic lock-profit (EnableDynamicLockProfit = true by default) is the most configurable money-management feature in the EA. Every tick, the EA reads PositionGetDouble(POSITION_PROFIT) and divides it by LockProfitEvery_X_Dollars (default $2). It then sets a new stop loss to entry ± (stepCount × LockProfitEvery_X_Dollars − LockMinusBuffer) / tickValue × tickSize. With defaults that means: at $2 of profit the stop locks at +$1, at $4 it locks at +$3, at $6 it locks at +$5, and so on. The stop is forward-only — it never moves against the position — and the $1 buffer keeps the trade from being closed by a one-tick drawdown back to entry. On XAUUSD at 0.1 lot each $2 of profit is roughly 20 cents of price, so the stop steps up in 10-point increments as the trade moves in your favour.

Position management runs in three modes that can be toggled independently. UseTrailingStop = true ratchets the stop to bid − 1.2×ATR for longs and ask + 1.2×ATR for shorts on every tick where the new stop is better than the current one. UseBreakeven = true moves the stop to entry once the position is more than 1.2×ATR in profit, removing the loss scenario from the trade. And the open-trade mid-line exit: if the position is a long and price closes back below gaussNow, or a short with price above gaussNow, the position is closed via TryClose_EX17022() — the Gaussian line is the invalidation level. All three layers stack, so a typical long trade has its stop ratcheting up via the ATR trail, hits breakeven around the first $0.50 of profit, and then starts locking profit in $2 increments via the dynamic lock — each layer tightening the risk envelope without ever loosening it.

Risk and safety are intentionally minimal. CheckMaxDrawdown() reads the live equity-to-balance percentage on every tick and refuses to evaluate any new entry once drawdown exceeds MaxDrawdownPercent (default 10.0%). The canTrade gate adds three more conditions: OneTradeOnly = true means at most one position per symbol (flip this off for grid-style accumulation), CountOpenTrades() < MaxOpenTrades (default 5) caps the per-symbol count, and AllowOnlyProfitableAdditions = true with MinProfitPerTradeToAdd = 5.0 means new positions in the same direction are only allowed when every existing position on the symbol is showing at least $5 of profit. That last rule effectively turns pyramiding into a trend-confirmation feature: you can only add to a position that is already working.

Order execution goes through TryClose_EX17022() and TryModify_EX17022() retry wrappers that attempt the close or modify three times with 200ms / 100ms sleeps on requote, timeout, price-off, and price-changed retcodes. Filling is auto-detected per symbol via trade.SetTypeFillingBySymbol(_Symbol). There is no partial-close path in the live code path, no news filter, no session filter, no day-of-week filter — if the spread is sane, the ATR is above the floor, and the candle passes the 60% body gate, the trade goes in. That makes the EA predictable to backtest but also means you are responsible for running it on a low-spread ECN or RAW account and being aware that the GAussian band thesis can be whipsawed in low-volatility regimes where the price keeps tapping the band but never breaks it cleanly. The default timeframe input is 0 (which maps to H1 via the MapTimeframeInt helper), although the EA was registered as M5 in the catalog and the source header recommends M5-M5 to M5-H1. Pair it with a $100 minimum balance, a low-spread broker, and you have the entire configuration surface.

In backtest, expect the equity curve to look like a stair-step rather than a smooth trend: the EA takes relatively few trades (a clean Gaussian-band break only happens a handful of times per session on M5 XAUUSD), but most of them either reach the 160-point fixed TP or get stopped at the 1.2×ATR stop with the dynamic lock taking a chunk out of the loss. The win rate tends to be above 50% in trending conditions and the average winner slightly larger than the average loser once the lock-profit steps start engaging. In chop, the 60% body gate and the MinBreakDistance filter do most of the heavy lifting by suppressing the entries that would otherwise pile up against the band. As with any band-breakout strategy, the worst regime for this EA is a slow, compressing range that resolves in a false break followed by a snap-back — the EA will fire on the break, get stopped at the ATR-based SL, and then watch price reverse the other way.