Pipsgrowth EX18116 TrendFollow

EX18116 QuantumGold is a 126-line MQL5 Expert Advisor that trades gold by demanding the same trend to be visible on three timeframes simultaneously before it will commit a single position. The file declares itself part of the TrendFollow family and proposes XAUUSD on M15 (with H1 and H4 confirmation candles), although the same skeleton is portable to any symbol for which the broker supplies continuous tick data. Magic 22218116 marks every order, and the trade comment is always suffixed with Psgrowth.com Expert_18116 Buy or Sell so the position is unambiguous in the terminal journal.

The indicator stack is intentionally narrow. Three Exponential Moving Averages of period 50 are created in OnInit — hM15_EMA, hH1_EMA, and hH4_EMA — each applied to PRICE_CLOSE. A single Average True Range of period 14, sampled on H1, supplies the stop distance. There is no RSI, no ADX, no MACD, no Bollinger envelope, and no dashboard panel. The OnInit routine registers the four handles, the OnDeinit releases them, and that is the entire housekeeping footprint. Indicator lookups happen once per closed bar, and the function reads exactly one bar of data from each handle using CopyBuffer(..., 1, 1, buf), so the signal is by construction a closed-bar, non-repainting decision.

The entry rule is the file's defining feature. When InpMTFConfirm is true (the default), a long entry requires bid > M15 EMA(50) AND bid > H1 EMA(50) AND bid > H4 EMA(50); a short entry requires the symmetric three-way violation on the bid side. When InpMTFConfirm is false, the multi-timeframe guard collapses to a single condition against the H1 EMA. The price-level check — comparing bid directly against each EMA value — is the kind of alignment test that fires late in an established trend and rarely on counter-trend re-entries, which is precisely the bias a gold swing trader is usually trying to manufacture.

Stop loss and take profit are sized from the ATR(14) read on the H1 timeframe at the same time the EMAs are sampled. The stop is ATR * 1.5, the take profit is SL * 2.0, which locks in a hard 1:2 reward-to-risk ratio on every order. Both protective prices are sent to the broker with the order and are then maintained by the trailing module. The lot size is calculated by CalcLot(double sl), which uses the formula lot = max(0.01, (balance * InpRiskPct/100) / (sl/ts * tickValue)), where InpRiskPct defaults to 1.0 percent of the current balance. This is a textbook risk-percent-of-balance sizer with a hard floor of 0.01 lots, normalized to two decimals. There is no compounding equity curve, no Kelly fraction, and no maximum-lot cap, so a trader who wants to cap exposure on larger accounts has to do it via InpRiskPct itself.

Position management is the ManageTrailing() function, called every tick before any new entry is considered. It walks PositionsTotal() from the end of the list, filters by POSITION_MAGIC == InpMagicNumber and POSITION_SYMBOL == _Symbol, and then computes a candidate stop as bid - InpTrailingStop * _Point for a buy and ask + InpTrailingStop * _Point for a sell. The new stop is only accepted if it is past the open price (i.e. the trade is in profit) AND it is tighter than the existing stop. Each accepted modification is sent through TryModify_EX18116, which retries up to three times on TRADE_RETCODE_REQUOTE, TRADE_RETCODE_TIMEOUT, TRADE_RETCODE_PRICE_OFF, or TRADE_RETCODE_PRICE_CHANGED, sleeping 100 ms between attempts. InpTrailingStop defaults to 40 points (4 pips on a 5-digit gold quote), so the stop tightens every 40 points of favorable movement and never loosens.

The file is single-position by design. HasTrade() is called before any entry block; if any position with the matching magic and symbol exists, the EA simply returns for that bar. There is no pyramiding, no scaling-in, no hedge layer, and no partial take-profit. The two retry helpers TryClose_EX18116 and TryClosePartial_EX18116 are declared in the source but are never referenced from the live code path — they are scaffolding kept in case the EA is later extended to do basket closes or partial exits. The only retry wrapper actually invoked is TryModify_EX18116.

The risk guard that does run is the daily loss cap. On the first tick after the server clock hits 00:00, the EA records dayStartBalance = AccountInfoDouble(ACCOUNT_BALANCE) and updates lastDay. On every tick afterward, it computes (dayStartBalance - equity) / dayStartBalance * 100 and returns immediately if that loss reaches InpMaxDailyLoss percent, which defaults to 3.0. This is a start-of-day balance check rather than a peak-equity drawdown check, so a strong early profit does not raise the threshold — only the open-of-day figure matters. When the cap is hit, the EA continues to call ManageTrailing() on existing positions but stops opening new ones, which is the right asymmetry for a session pause.

Other claimed layers in the file's header comment are not implemented in the body. There is no OnTester function, no regime classifier, no capital cap input, no news filter, no session filter, no ATR-volatility filter, no pyramid, no breakeven, and no dashboard. The header advertises the full twelve-layer stack as a marketing summary, but the executable code is closer to a five-piece system: trend confluence, ATR-based stop, 1:2 R:R target, percent-risk sizing, and a 40-point trailing stop with a 3 percent daily loss circuit breaker.

For a backtest, the practical consequences are: each trade needs roughly 0.5× ATR of room in the stop, so the EA prefers XAUUSD sessions where 1.5× the H1 ATR is at least a few dollars — which is just the Asia/London overlap and the first half of New York. The single-position constraint means the equity curve will be a series of separate round-trips rather than overlapping grids, and the lack of a breakeven step means losing trades are typically closed by the trailing stop only when the structure has clearly turned. Testers should expect the win rate to be modest and the average winner to be twice the average loser, which is the only sustainable combination for a 1:2 R:R with a 1 percent risk budget.