Stoploss hunter

Stoploss hunter 2.00 is a correction release. The strategy is unchanged: a rejection-candle reversal, filtered by a 200-period DEMA trend, an RSI gate and a Bollinger extreme. What changed is everything around it.

The short side works again.
In the previous build, three lines created indicator handles and compared them to prices. The effect was that one condition was always true, and the expert returned before every sell order. Measured in the Strategy Tester on EURUSD H1 over three and a half years, the previous build opened 149 long trades and zero short trades. Version 2.00 removes those lines. The expert trades both directions, as it was originally written to.

There is now a stop loss.
The previous build placed its stop at 40% of the instrument's price. On EURUSD that is roughly 4,320 pips away, below the pair's all-time low. It could not be reached. Version 2.00 places a stop a configurable number of ATR from entry, clamped to your broker's own minimum distance.

The drawdown guard works.
The previous guard measured floating loss against account balance. Balance moves with realised profit, so the reading reset to zero after every close and a slow bleed showed no drawdown at all. It also had no stand-down, so the expert re-entered on the next tick at the extreme that had just triggered it. Version 2.00 measures against the equity peak and stands down afterwards.

It talks to your broker properly.
Volume is rounded onto your broker's own step after the multipliers are applied, not before, so the multiplied sizes are valid orders. Free margin is checked before sending. The filling mode is set for opening as well as closing - on an IOC-only ECN account the previous build could close positions but never open one. Every order's return code is checked and reported in the journal instead of being discarded.

It manages only its own trades.
Position counts and the close-all routine are scoped to this symbol and this magic number. Previously a single unrelated position anywhere on the account was enough to stop the expert trading, and two charts sharing a magic number could liquidate each other.

Also in 2.00: day-of-week filters that apply while a position is open rather than only when flat; a spread limit; input validation at start-up instead of a runtime error; indicator handles created once and released; new-bar detection by bar time rather than by closing price; an optional tick-volume filter, switched off by default; and a refusal to start on a netting account, where the position cap cannot work.

Settings

• Pair: EURUSD. Timeframe: H1. Account: hedging.
• useDynamicLot = true, perThousandUSD = 0.01.
• Do not raise perThousandUSD. Larger positions meet the drawdown close far more often; in testing, 0.05 turned a positive result negative.
• A tight-spread broker matters, and 2.00 can enforce it - set maxSpreadPoints.

On performance
No backtest figure is published with this version, and there is no live signal for this product yet. The previous version's results came from a build that could only buy, measured on a period in which EURUSD rose. Publishing a number from it would describe something you would not be running. Test 2.00 on your own broker's data, on a demo account, before committing real money.

If you do not have a tight-spread broker, I recommend IC Markets: https://icmarkets.com/?camp=64005