I built the same strategy in FXDreema, StrategyQuant, and an AI generator — manual fix counts inside

@pips_and_pythonjoined Jun 20, 2025Jul 30, 2026414 views4 replies

Same strategy, three tools, one weekend: EMA 50/200 cross on EURUSD H1, ATR(14) trailing stop, no trades during red-folder news, 0.5% risk per trade.

I counted every manual intervention needed to get from "generated" to "compiles and behaves correctly in the visual tester":

FXDreema: 0 compile fixes, 6 logic adjustments. The news filter block alone took 40 minutes to wire correctly.
StrategyQuant: it is a generator-miner, not a builder — it produced 214 strategies and I rejected 211 as obvious overfits. The 3 survivors needed spread-stress retests. Different tool, different skill.
AI text-to-MQL (wobr.ai EA Generator): 2 compile fixes, 3 logic fixes. Fastest to a first draft by far, but it wired the entry signal to fire on every tick instead of new bar — the classic bug — and invented a nonexistent MQL5 constant.

Takeaway: none of them remove the need to understand what the EA does. They shift the work from writing code to reviewing it.

4 Replies

@ea_architectjoined Mar 23, 2025#1 · 6 days ago
+2
The every-tick-instead-of-new-bar wiring is the single most common generator bug I see, across every tool. My habit: attach any generated EA to the visual strategy tester and just watch it trade one month of ticks. Logic errors become obvious in about two minutes.
For what it's worth, my hand-written first drafts have a similar fix count — they're just different bugs, and I find them slower to spot in my own code than in generated code.
@mt5_tinkererjoined Jul 14, 2025#2 · 6 days ago
+1
Your StrategyQuant rejection rate rings true. My filter: discard anything with in-sample PF above 2.5 or under 300 trades, which kills roughly 95% of output before I even look at it. The tool is a hypothesis factory — treat it that way and it earns its keep; treat it as a strategy vending machine and it will bankrupt you politely.
@grid_goblinjoined Aug 15, 2025#3 · 6 days ago
0
I ran the same exercise last year with a grid EA and all three tools failed at basket-level state management — tracking aggregate exposure across open grid levels and sizing the next level off total floating P&L. Some strategy types are still hand-code only. Has anyone gotten a generator to handle multi-position basket logic correctly?
@pips_and_pythonjoined Jun 20, 2025#4 · 6 days ago
0
Confirmed — basket logic broke all three of mine too, each in a different and creative way. Next test on my list is a multi-symbol portfolio EA; I expect similar failure and I will log the fix counts here either way.