Bottom line up front
This account is -3.6% net of fees across 514 closed trades over 139 days. The headline obscures a sharply asymmetric pattern: NVDA shorts worked cleanly (41% win rate, $51.80 realised PnL), but XYZ100 consumed the entire edge and then some, with a 31.7% win rate across 435 episodes and a -$398.46 realised loss. The account exhibits textbook revenge-trade and averaging-down behaviour concentrated in XYZ100, with five oversized losses running 3–8× the median loss size. Fees were minimal ($11.32 net drag), so the damage is pure execution and position selection, not slippage.
What the data shows
The account opened on 14 November 2025 and has traded three primary instruments. NVDA and TSLA were tactical, high-conviction shorts that worked: NVDA generated $51.80 on 39 episodes (41% win rate), and TSLA closed flat at -$5.29 on 39 episodes. XYZ100, by contrast, became a compulsion. Across 435 episodes—85% of all closed trades—the account lost $398.46 with a 31.7% win rate. The balance peaked at $5.74M on 25 March 2026, then fell to $3.22M by 21 January 2026, a -39.8% decline. The account recovered to $9.68M by the end of the window, but that recovery is driven by the opening balance estimate, not by trading edge.
Long trades underperformed short trades: longs returned -$203.06 at 30.9% win rate, while shorts returned -$148.89 at 34.1% win rate. The short-side edge was marginal and concentrated in NVDA. The account's win rate of 32.5% across all episodes sits well below breakeven for a 0.7-to-1.35 win-to-loss ratio structure. Average win was $0.70; average loss was $1.35. That asymmetry, combined with the 32.5% win rate, produced an expectancy of -$0.68 per trade.
Fees were negligible: $11.32 net drag on $1.11M gross volume, with 94.3% maker fills. The real cost was position selection and re-entry discipline.
Trade quality
Win rate of 32.5% with a profit factor of 0.25 means the account lost $0.75 for every $1.00 it made. The win-to-loss ratio of 0.52 reflects the structural problem: when the account won, it won small ($0.70 average); when it lost, it lost larger ($1.35 average). Expectancy of -$0.68 per trade is the mathematical consequence. Over 514 closed episodes, that expectancy compounds into the observed -$361.75 realised loss.
The account had a max win streak of 17 and a max loss streak of 17, indicating no consistent edge and high variance. The profit factor of 0.25 is the most damning metric: it means the account's gross wins covered only one-quarter of its gross losses.
Post-mortems
XYZ100 short, 18 November 2025, 24635.22 entry, 24710.0 exit, -$10.38 loss
This trade carried four behavioural flags: averaging down, FOMO re-entry, oversized loser (4.5× median loss), and revenge trade. It was opened and closed within 0.1 hours at a max notional of $3,463.82. The structural stop was 0.86% away (ATR-based). The trade lost $10.38 on a 75-pip move against the position. This was a revenge trade following a loss in NVDA, opened into a position already sized above median, and closed in panic.
XYZ100 short, 17 November 2025, 25157.67 entry, 25240.0 exit, -$9.69 loss
Four flags again: averaging down, FOMO re-entry, oversized loser (8.4× median loss—the worst in the set), and revenge trade. Held 0.37 hours at $2,998.95 notional. The structural stop was only 0.25% away. This trade lost $9.69 on an 82-pip move. It was opened after a loss in TSLA and closed within minutes, consistent with panic liquidation rather than planned exit.
Both post-mortem trades are XYZ100 shorts from mid-November, both flagged as revenge trades following losses in other instruments, both oversized relative to the account's median loss tolerance, and both closed in under an hour. They exemplify the pattern: after a small loss elsewhere, the account would open an outsized XYZ100 position and exit at the first sign of adverse movement.
What the risk simulation reveals
Under a 1% stop-loss rule applied historically, the account would have realised -$3.43M PnL with a -35.6% max drawdown and 31.1% win rate. Under 2%, the simulated loss would have been -$6.86M with a -71.1% drawdown. Under 4%, -$13.71M with a -142.1% drawdown. These are gross-of-fees counterfactuals. The simulation stopped 15 episodes early across all three scenarios, indicating that strict stops would have prevented some of the worst losses but also locked in the account's structural negative expectancy faster. The fact that simulated losses scale linearly with stop width suggests the account's losses are not clustered in a few catastrophic trades but distributed across many small underwater positions that would have been stopped out.
Open positions
No open positions at the time of analysis.
Honest summary
- Strength: NVDA shorts demonstrated edge. A 41% win rate on 39 episodes with $51.80 realised PnL shows the account can identify and execute a directional thesis. The trade was held for 101 hours, entered at 190.36, exited at 181.34, and let the position breathe without re-entry noise.
- Weakness: XYZ100 became a loss-making obsession. 435 episodes, 31.7% win rate, -$398.46 realised loss. The coin appears in five of the top five losses, all flagged as revenge trades or oversized losers. The account chased losses in this instrument repeatedly, opening positions after losses in other coins and closing them in panic within minutes.
- Weakness: Averaging down and FOMO re-entry are endemic. Five averaging-down episodes are flagged, and five FOMO re-entries are documented. The re-entries in TSLA on 14 November (three separate re-entries within hours) and the XYZ100 revenge trades show a pattern of emotional re-engagement after losses rather than systematic position management.
- Structural issue: The account's negative expectancy (-$0.68 per trade) means profitability requires either a dramatic shift in win rate or a substantial increase in average win size relative to average loss. Neither is visible in the data. The account is trading too many small-edge or no-edge instruments (XYZ100, TSLA) and concentrating capital in them after losses, not after wins.