Pipsgrowth EX07014 MA

Pipsgrowth EX07014 sits at the most minimal end of the MA-cross family. The whole indicator stack is one line: a single Simple Moving Average built from the close, period 12, shifted six bars into the past. No RSI, no ADX, no ATR, no Bollinger band, no HTF bias, no multi-MA ribbon. The author leaned in the opposite direction of EX07011's eight modules and EX07001's seven-state regime classifier — they gave the trader a textbook mean-reversion cross and asked the price action to do the work.

The entry rule lives in CheckForOpen and is intentionally narrow. Each new bar, on its very first tick (rt[1].tick_volume must be 1 or smaller — anything else means a later tick of the previous bar, and the call returns immediately), the EA pulls the MA value and looks at the current candle's open-to-close relationship to that line. If the bar opened above the MA and closed below it, the EA sends a market SELL. If the bar opened below the MA and closed above it, the EA sends a market BUY. There is no candle pattern, no wick filter, no volume confirmation, no spread cap, and no minimum body size. The cross itself is the signal.

The shift of six bars is the design choice that gives the system its character. iMA(_Symbol,_Period,MovingPeriod=12, MovingShift=6, MODE_SMA, PRICE_CLOSE) returns a value that was computed six candles ago, not the live MA. That deliberate lag means the EA is not chasing the average — it is reacting to a stale reference. On a five-minute chart the line is effectively smoothed across an hour of price action, which filters a lot of the noise that would otherwise whip a faster cross. On H1, six bars is six hours; the MA becomes essentially a daily/weekly-range anchor. The trader who picks M5 gets a different rhythm than the trader who picks H1, and both are valid.

PositionOpen is called with both SL and TP equal to 0 — the server never sees a hard stop, and the EA never sets a take-profit. The only exit is in CheckForClose: when a position is open and the next bar prints a candle that closes on the opposite side of the MA relative to its open (a SELL triggers when a buy-position's bar opens above MA and closes below; a BUY triggers when a sell-position's bar opens below and closes above), TryClose_EX07014 fires with a 3-pip-equivalent deviation. The retry wrapper walks the same three attempts at 200 ms intervals on REQUOTE, TIMEOUT, PRICE_OFF, or PRICE_CHANGED retcodes that the rest of the EX07xxx family uses. Net result: the EA lives and dies by the reverse cross. It can sit in a trade for five bars or for five hundred; the only way out is the next opposing cross.

TradeSizeOptimized is the most interesting function in the file. It starts from the same risk-percentage-of-free-margin pattern that the other EX07 EAs use, but instead of an equity cap it uses orderCalcMargin against 1.0 lot to derive margin per unit, then sets lot = free_margin × MaximumRisk / margin. MaximumRisk defaults to 0.02 (2%). After that, DecreaseFactor kicks in: the function walks HistorySelect from the most recent deal backwards, counting consecutive losses tagged with this EA's magic 22207014, and the moment it sees a profitable deal (or runs out of history) it stops. If the count is greater than 1, the lot is reduced by lot × losses / DecreaseFactor, so a DecreaseFactor of 3 and a string of three losers gives you lot × 1.0 reduction — the next trade is at minimum lot. The mechanism is the opposite of martingale: a losing streak shrinks position size, and the streak only resets on a winning trade. The final lot is then floored to SymbolInfoDouble(SYMBOL_VOLUME_MIN) and capped to SYMBOL_VOLUME_MAX, with the standard stepvol rounding pass to keep it on the broker's grid.

OnInit is the third piece worth noting. It detects the account margin mode — ExtHedging = (ACCOUNT_MARGIN_MODE == ACCOUNT_MARGIN_MODE_RETAIL_HEDGING) — and SelectPosition() branches on that. Hedging accounts loop PositionsTotal() and match by symbol and magic, so multiple positions in the same direction can stack if the broker permits. Netting accounts use PositionSelect on the symbol directly, and the magic must match. ExtTrade.SetExpertMagicNumber(MA_MAGIC) and SetTypeFillingBySymbol() are the standard wrappers; if the iMA handle fails to create, OnInit returns INIT_FAILED and the EA will not start. There is no global OnTradeTransaction filter, no news blackout, no session window, no equity floor, no daily-loss cap, no time stop, no breakeven, no partial close, no trailing ratchet, no pyramiding. The Bars > 100 sanity check is the only safety net in the entry path.

What the trader gets is a clean test bench for a single MA. Drop it on XAUUSD M5, leave the defaults, and the EA will fire when gold crosses its six-hour shifted SMA. Tighten MovingPeriod to 6 and you get a faster, choppier cross that will trade more often. Push MovingShift back to 0 and the MA becomes current — entries tighten, whipsaws multiply. Increase DecreaseFactor to 6 to slow the post-loss shrink, drop it to 2 to compress the curve. There is no optimizer function in the source (no OnTester), so backtest the parameters across different volatility regimes manually, and accept that on a ranging market this EA will pay for every cross with a loss.

For a backtest expectation: the equity curve on a trending XAUUSD M5 will be lumpy but generally upward — long runs of small profits separated by occasional larger givebacks as the lag-filtered MA falls behind a fast reversal. The decrease factor means drawdowns grow in length, not in lot size, which is the gentler failure mode. On EURUSD H1 the same code is a slow grinder; on M1 the shift=6 will look like nonsense. Treat EX07014 as a price-action baseline — it is the stripped-down reference point the rest of the EX07 MA family is measured against.