Phantom ICT

I’ve spent some time testing Phantom ICT and, so far, what I appreciate most is that it feels like a genuinely different approach compared with many Gold EAs in the Market. The biggest positives for me are the clean risk architecture: no martingale, no recovery grid, no averaging into losing positions, one position at a time, and every trade has a real stop-loss. The EA is also very selective rather than forcing trades every day, which makes sense for an ICT/liquidity-based strategy. From my testing, the logic appears to revolve around higher-timeframe bias, liquidity sweeps, Order Blocks, retests and confirmation before entry. I also like that the EA is willing to show losing trades and losing streaks in backtesting instead of presenting an unrealistically perfect equity curve. The main area I think the author should continue improving is **regime detection and losing-cluster protection**. Phantom performed much better in some market conditions than others. There were periods where several valid-looking setups lost consecutively. Rather than simply adding filters to remove historical losses, I would like to see the EA become better at identifying when its ICT setup is operating in an unfavorable market regime. Some improvements I would personally like to see: • Better distinction between a genuine liquidity sweep and a normal wick/fake sweep. • Stronger displacement/MSS confirmation after the liquidity raid. • More adaptive Order Block quality filtering. • Optional premium/discount or HTF structure confluence. • Percentage-risk position sizing based on actual SL distance instead of relying mainly on fixed lots. • Optional CPI/NFP/FOMC protection. • A configurable account/equity drawdown protection. • More information on the dashboard explaining why a setup was accepted or rejected. I would also be careful about adding too many new parameters. Phantom already has a lot of filters, and I would rather see the strategy simplified and made more robust than optimized to produce a prettier historical result. Overall, I think Phantom has a **very promising foundation** because its underlying architecture is much cleaner than most high-profit Gold EAs I’ve tested. What I want to see next is more long-term and forward evidence showing that the same settings can survive different Gold market regimes without excessive optimization. For me, Phantom is currently a **promising experimental EA rather than a finished “set and forget” system**, but I appreciate the direction the author has taken—especially avoiding martingale/grid recovery and focusing on actual market structure and controlled risk. Looking forward to seeing how the strategy develops.