Pipsgrowth EX15012 SMC-OrderBlock

Pipsgrowth EX15012 is an XAUUSD liquidity-trap scalper that maps the chart as a sequence of swing-anchored demand and supply pools, then waits for price to return into one of those pools and reject from it. The whole logic is anchored to liquidity zones, not to candles or indicators: the EA keeps an internal list of swing-high prices (upper zones) and swing-low prices (lower zones), drawn on the chart as dashed horizontal lines, and every decision on every new bar starts by asking 'is the current price touching any active zone?' If yes, the four-gate confirmation chain runs. If no, the EA does nothing that bar.

The 3-bar swing detector is implemented by FindSwingHigh and FindSwingLow. With SwingBarCount = 1 (the default), an index i is considered a swing high only if high[i] > high[i-1] and high[i] > high[i+1], and the swing low mirror is the same. The function walks forward from barsEachSide and returns the first qualifying index. Whenever IdentifySwingPoints finds a swing it has not yet recorded, it stamps the price and time into the upperLiquidityZone[] or lowerLiquidityZone[] arrays and calls AddLiquidityZone, which de-duplicates by skipping any new zone within 0.3% of an existing one and then calls MarkLiquidityZone to draw the dashed OBJ_HLINE chart object (crimson for upper, dodger blue for lower). Every 4 hours, CleanupInactiveZones walks the arrays, deletes the chart objects for deactivated zones, and rebuilds the global arrays. Zones also self-expire after MaxZoneAge = 24 hours, computed from zoneCreationTime against the current time.

Entry is gated. When OnTick detects a new bar (or the first tick), GenerateSignals calls CheckBuySignals and CheckSellSignals. For a buy, the EA asks whether the ask price sits inside the lower zone's 1% band — that is, currentPrice >= zoneLower AND currentPrice <= zoneLower * 1.01. For a sell, the mirror is the upper zone's 1% downward band. If the price is in the band, four sub-checks run: CheckPriceSurge (open-to-current percent change over 10 bars, or 5-bar high-low range percent, whichever is larger, must exceed PriceSurgePercent * 100 = 0.1% by default), IsVolumeSurge (current tick volume ≥ 1.5x the 20-bar average — bypassed if RequireVolumeSurge = false, which is the default), IsZoneConfirmedByHigherTimeframe (calls DetermineTrend on the H4 default timeframe, comparing MA20 vs MA50 — bypassed if UseHigherTimeframeConfirmation = false, the default, and permissive when the trend is 0), and IsBullishEngulfingPattern / IsBearishEngulfingPattern (walks back EngulfingLookback = 3 bars looking for a candle whose body engulfs the prior opposite-color candle's body — this is the only one of the four gates that is hard-on by default). All four must pass. After the order fires, zoneActive[i] = false deactivates the zone so it cannot be re-traded.

Sizing and stops are deliberately simple. CalculatePositionSize ignores the stopLossPoints argument and returns the fixed FixedLotSize = 0.01 lots, normalized to the symbol's lot step. For a buy, the stop is zoneLower * (1 - 0.005) — 0.5% below the zone. For a sell, the mirror is 0.5% above. The take-profit is then currentPrice + (currentPrice - stopLoss) * RewardToRiskRatio for buys, with RewardToRiskRatio = 2.0 default producing a clean 1:2 reward-to-risk. The order is sent via Trade.Buy / Trade.Sell with the comment Psgrowth.com Expert_15012 and magic 22215012.

There is no progressive profit lock, no post-entry management that re-runs the swing logic, and no time-based exit. The trailing stop input exists (EnableTrailingStop, TrailingStopPercent = 50%) but the code path that would actually move the stop is not wired into ManageOpenPositions — the only thing that function does is count buys/sells, walk the last 24 hours of HistorySelect to update UpdateTradeStatistics, and log position counts. The high-impact-news input AvoidHighImpactNews = true (with NewsBufferMinutesBefore = 30, NewsBufferMinutesAfter = 30) is similarly a dead switch — there is no news API integration in the file. The three retry helpers at the end of the source — TryClose_EX15012, TryClosePartial_EX15012, TryModify_EX15012 — are defined against a separate CTrade g_trade_wrappers_EX15012 instance, retry up to 3 times on REQUOTE / TIMEOUT / PRICE_OFF / PRICE_CHANGED, and are not called from OnTick, ManageOpenPositions, ExecuteBuyOrder, or ExecuteSellOrder. They are present as scaffolding for a v3.0 build that the source explicitly TODO-marks: // TODO: Add trailing stops and dynamic take profit in v3.0 and // TODO: Add additional market analysis in v3.0.

Higher-timeframe trend, when used, is read by DetermineTrend which pulls MA(20) SMA and MA(50) SMA from either the main or the higher timeframe handle set. The MA handles — g_ma20HandleMain, g_ma50HandleMain, and the optional g_ma20HandleHTF, g_ma50HandleHTF — are created in OnInit only if the corresponding input is enabled, and they are released by ReleaseIndicatorHandles in OnDeinit. The function uses a progressive-delay retry loop (5 attempts, 100ms × attempt number) for the CopyBuffer call, and resets the handle to INVALID_HANDLE on persistent failure so the next call re-creates it. The MapTimeframeInt helper that maps the integer Timeframe input (1=M1, 2=M5, 3=M15, 4=M30, 5=H1, 6=H4, 7=D1) to an ENUM_TIMEFRAMES constant is defined five separate times in the source — a code-quality issue, not a logic bug.

Risk controls in the EA are intentionally minimal. MaxConsecutiveLosses = 3 triggers pauseUntilTime = TimeCurrent() + PauseMinutes * 60 (60 minutes default). MaxPositions = 5 is the cap. MaxSpreadPoints = 30 blocks any tick whose current spread exceeds 30 points (XAUUSD on a tight broker will run around 10–20 points, so this is mostly a stale-quote and rollover guard). The EnableTimeFilter switch is off by default; when turned on it parses the TradingHoursStart (08:00) and TradingHoursEnd (20:00) strings into minute-of-day and gates signals to that window in broker time. OnInit also runs an ECN detection heuristic: if ask - bid < 3 * point, the EA sets Trade.SetDeviationInPoints(10) for slippage. The UpdateDashboard routine writes a six-line label panel showing title, status, lot+RR, active upper/lower zone count, total trades + win/loss, profit factor (color-coded green ≥ 2.0, yellow ≥ 1.0, red below), and current signal. ObjectsDeleteAll(0, "EA_") cleans both the zone lines and the dashboard on deinit.

The .mq5 ships with the conservative defaults that a conservative trader would want: 0.01 lots, 1:2 reward-to-risk, 3-bar swing detection, 0.3% de-dup between zones, 1% zone-band entry, 24-hour zone lifetime, only the engulfing confirmation hard-on out of the four gates, a 3-loss / 60-minute pause, and a 5-position cap. The progressive profit lock, trailing stop, and news filter that the family name suggests are visible in the input list but not actually wired to the live trade path — that is the single most important operational fact for anyone considering this EA, and it is what the source file's own TODO comments confirm.