Retail traders naturally assume that if a bot wins 9 out of 10 trades, long-term profitability is virtually guaranteed. They fail to inspect the underlying risk-to-reward ratio, assuming that modern code has solved the risk equation.
Expectancy = (Win Rate × Average Win) - (Loss Rate × Average Loss).
If an EA wins 95% of the time, but the average win is +$5 while the average loss is -$150 (a 1:30 inverted Risk-to-Reward ratio), the mathematical expectancy per trade is: (0.95 × 5) - (0.05 × 150) = 4.75 - 7.50 = -$2.75 per trade.
Every time the bot places a trade, the account bleeds -$2.75 in expected value. The 95% win rate simply creates a psychological illusion of safety until a single directional market shock triggers margin call liquidation.