Pipsgrowth EX15010 SMC-OrderBlock

EX15010 is a swing-pivot liquidity-trap reverser for XAUUSD, built around a single mechanical idea: where the market previously carved a swing high or swing low, that extreme price often contains resting stop orders — liquidity — and price tends to revisit those levels to grab that liquidity before reversing. The EA maps those levels as zones, then fires a counter-trend market order the moment price walks back into the zone with enough speed to qualify as a sweep. The structure is deliberately lean: four moving-average handles (two for the main timeframe, two extra created only when the higher-timeframe filter is enabled), a single 1% wide zone buffer, a 0.1% price-move threshold, and a fixed-lot sizing template. No indicator is wasted; no exit logic runs after the order is sent.

The detection engine lives in IdentifySwingPoints and its helpers FindSwingHigh and FindSwingLow. With the default SwingBarCount = 1 the engine looks for the simplest possible pivot — a bar whose high is higher than the bar before and after it, or whose low is lower than its neighbours. Each newly discovered pivot is stored in swingHighPrice / swingLowPrice and then converted into a horizontal liquidity zone in AddLiquidityZone. The zone is a 1% wide price band centred on the pivot (zoneLower = lowerLiquidityZone[i] and zoneUpper = zoneLower * 1.01 for buys, or the inverse for sells), which the EA draws on the chart as a dashed OBJ_HLINE — crimson above, dodger blue below. A 0.3% de-duplication check rejects new zones that crowd existing ones, so a tight consolidation that prints several pivots within 30 points does not flood the watch list. Zones expire after MaxZoneAge = 24 hours, and CleanupInactiveZones runs every four hours (TimeCurrent() % (3600*4) < 60) to strip dead entries from the arrays and delete the corresponding chart objects.

Signal generation happens once per closed bar inside GenerateSignals. The buy path, CheckBuySignals, walks the lower-zone array looking for a zone where SymbolInfo.Ask() is currently inside the 1% band. If price is inside, the engine calls CheckPriceSurge(true), which uses the larger of two metrics: the absolute percentage change from the open of the current bar to the live ask, and the percentage range of the last five bars (highest high minus lowest low, divided by the lowest low). Whichever is bigger has to clear PriceSurgePercent = 0.001 — i.e. 0.1% — for the surge gate to flip green. IsVolumeSurge then samples 21 tick volumes, averages bars one through twenty, and asks whether the live bar's tick volume exceeds the average by VolumeSurgeMultiplier = 1.5 times; that gate is only consulted when RequireVolumeSurge = true, which is the default. The third gate, IsZoneConfirmedByHigherTimeframe, is also off by default; with UseHigherTimeframeConfirmation = false it returns true immediately, but when armed it calls DetermineTrend on g_ConfirmationTimeframe and checks the MA20 vs MA50 cross — a +1 reading (MA20 above MA50) must confirm a buy off a lower zone, a -1 reading must confirm a sell off an upper zone, and a 0 reading (perfectly flat cross) is treated as permissive. When all three gates are green, the EA computes a stop loss 0.5% beyond the zone (stopLoss = zoneLower * 0.995 for buys, zoneUpper * 1.005 for sells) and a take profit at RewardToRiskRatio = 2.0 times the distance — a fixed 1:2 reward-to-risk geometry — then calls ExecuteBuyOrder or ExecuteSellOrder. After a fill, the zone is marked inactive so the same level cannot fire twice.

There is no post-entry logic. ManageOpenPositions is essentially a bookkeeping routine: it counts buys and sells, compares the current POSITION_TICKET array against the last tick's array to detect closed trades, and hands the result to UpdateTradeStatistics which updates win/loss counters, gross profit, gross loss, and the equity-curve drawdown. When consecutiveLosses reaches MaxConsecutiveLosses = 3, pauseUntilTime is set to TimeCurrent() + 60 * 60, and the next 60 minutes of ticks return immediately from OnTick after refreshing the dashboard. Three helper functions — TryClose_EX15010, TryClosePartial_EX15010, TryModify_EX15010 — are declared at the bottom of the file with a 3-attempt retry loop on REQUOTE/TIMEOUT/PRICE_OFF/PRICE_CHANGED codes, but they are not wired into the live order flow; the EA's Trade.Buy and Trade.Sell calls go through a separate CTrade Trade; object that does not reference these helpers, so the retry layer is dormant defensive code. The EnableTrailingStop = false default and the explicit // TODO: Add trailing stops and dynamic take profit in v3.0 comments in ManageOpenPositions make it clear that exits are intended to be handled entirely by the broker-side SL and TP attached at order time.

The risk scaffolding is thin but explicit. FixedLotSize = 0.01 is the only sizing input; CalculatePositionSize reads it and hands it to NormalizeLotSize, which clamps to SYMBOL_VOLUME_MIN/MAX and rounds to the symbol's lot step. MaxPositions = 5 caps total open positions across both directions, and the position counter is checked in both CheckBuySignals and CheckSellSignals before each fill. MaxSpreadPoints = 30 filters out wide-spread conditions at the top of OnTick — anything wider than 30 points returns immediately without generating signals. EnableTimeFilter = false by default means the 08:00–20:00 TradingHoursStart/End window is parsed but not enforced; turning it on makes IsWithinTradingHours gate the tradingAllowed flag. AvoidHighImpactNews = true and the two 30-minute buffer inputs are present in the parameter block, but no CalendarValueHistory or news-fetching call exists in the source — the flag is currently a no-op. The four global MA handles (g_ma20HandleMain, g_ma50HandleMain, and the two HTF twins) are released in OnDeinit and chart objects are wiped with ObjectsDeleteAll(0, "EA_"), but UpdateDashboard re-creates its labels on every tick so removing the EA cleanly clears the visible footprint.

Backtest this with realistic tick data and a commission per lot that matches the broker you intend to run live on. XAUUSD on M30 with SwingBarCount = 1 produces a high frequency of zones (almost every meaningful pivot counts), and the 0.1% price-surge threshold is low enough that the EA will trigger often; tightening the threshold to 0.005 or raising SwingBarCount to 2 or 3 will trade far less but with cleaner setups. The RewardToRiskRatio = 2.0 means the EA needs a win rate above ~34% to be net positive at 1:2, which is achievable on gold's reversal behaviour but unforgiving on the spread component — a 30-point ceiling on the spread filter is appropriate for accounts where the average spread sits around 10–20 points. The fixed 0.01 lot size and 5-position cap mean the worst-case portfolio heat is bounded but small; traders who want to scale this EA should pre-multiply the lot or run multiple instances on different magic numbers rather than relying on the EA's internal scaling, because there is no equity-based lot escalation, no martingale, and no recovery multiplier in this build.