My AI agent planned a 2:1 trade. The exit button had other plans.

LINK long autopsy, real fills:

🤖 Plan (placed by my Hermes agent, 20x isolated):
• Entry limit 13.150, filled as maker
• Stop 12.950 (risk ≈ $1.90)
• Take-profit 13.550 (reward ≈ $3.80)
• Size 9.5 LINK ≈ $125 notional, ~$6.25 margin

😬 What actually happened:
• Closed at market 13.238, 56 minutes in
• Gross +$0.84, fees ≈ $0.09 → net ≈ +$0.75
• That's 0.44R on a trade built for 2R
• Fees ate ~10% of the win. Taker exit cost 3x the maker entry.

📈 The painful part:
• Stop at 12.950 never came close. Low after entry: 13.091
• Price traded through the 13.550 TP about 12 hours later
• 24h high 14.215, now ~13.98 (+13.6% on the day)
• Holding the plan = +$3.80. Taking the itch = +$0.75

🧮 The math I keep relearning:
With a 2:1 plan you only need to win >33% to break even.
Cut winners at 0.44R while losers still cost 1R, and break-even jumps to ~69%.
Same entries, same "edge", completely different system.

The agent's order logic was fine. The weak link was the human with a close button.

Rule going forward: once the bracket is set, the exit belongs to the bracket. Moving the stop to break-even is allowed. Grabbing crumbs isn't.

Do you let your stops and TPs do the work, or do you still babysit every trade?

#TradingLessons #AITrading
$LINK