Grid Guard
- Utilities
- Version: 1.0
- Activations: 20
Your grid EA wins 9 trades out of 10. That was never the question. I ran the mechanism the best-selling gold robots use — grid, recovery, no stop loss on any position, one hard equity stop — over four years of real XAUUSD data, with real spread taken from every bar, real commission and real swap. 156 winning series out of 169. A 92.3% win rate. The account finished down 86.2%, with a peak drawdown of 96.8%. Both numbers are true at the same time, and that is the whole problem. The 92% is what goes on the sales page. The 8% is what arrives on a Tuesday. GRID GUARD DOES NOT REPLACE YOUR EA. IT WATCHES IT. Attach it to the same chart as the robot you already bought. It reads the open positions, measures the grid's actual step and lot progression from them, and answers the one question your EA never shows you: If it adds four more layers, what does this basket cost me? That number, in your account currency and as a percentage of your equity, in letters large enough to read from across the room. HOW IT KNOWS It never sees your EA's plan. It does not need to. The step is recovered as the MEDIAN gap between the open layers, and the lot progression as the median ratio between them. Median, not average, so one layer that filled late on a news spike does not corrupt the estimate. From there it projects forward on the geometry your robot has actually been using — not on a setting you had to look up and type in. With a single position open there is no gap to measure. It says so, in those words, and uses the step you supply instead. It will not invent the number the whole product exists to give you. WHAT IT DOES WHEN THE NUMBER GETS BAD Two things, because on MetaTrader there are only two. ALERT — the panel turns red and your phone gets a push notification, with the projected loss as a percentage of equity. CLOSE — it liquidates the basket. This is OFF by default and stays off until you switch it on yourself. No MQL5 program can veto another program's orders, and any product that tells you otherwise is describing something the platform does not permit. WHAT IT DOES NOT DO It does not open positions. It does not generate signals. It does not predict direction — I tested that on thirteen years of EURUSD and gold, and the same models scored 53.4% on randomly shuffled labels, higher than on the real data. It will not make a losing robot profitable. Nothing will. It tells you what the robot can still take, while you can still do something about it. SETUP Attach to the chart. Set two numbers: how many more layers to assume, and the percentage of equity you are willing to see projected. Everything else has a working default. It runs alongside any EA, any magic number, any broker. Set Magic to -1 and it guards every robot on the symbol. FREE DEMO The demo runs in the Strategy Tester, where there is no other robot to guard — so in the tester it builds a grid of its own and walks the panel through OK, WARNING and BREACH. The logic is the shipping logic. Only the positions are simulated. Download it. Watch the number climb. THE MEASUREMENT Every figure on this page came out of a run I can reproduce. The core arithmetic is checked by 48 assertions under an independent compiler with no terminal present, and the forecasting code it shares with my other products passes 9,045 numerical checks against a reference implementation. I have not tuned a backtest to look good here, because there is no backtest to tune — Grid Guard does not trade. Questions before you buy: send me a message. I answer.
