Bottom line up front
This account is -3.02% in the analysed window, down $23,440 on a starting balance of $775,707. The headline obscures a catastrophic pattern: two oversized losses on 6 June—a $18,264 short on XYZ100 and a $3,082 long on GOLD—triggered a cascade of revenge trades (BRENTOIL, SP500) that compounded the damage. The account recovered marginally with small wins on micro-cap longs, but the core issue is structural: position sizing exploded on the two largest losses (34× and 6× median loss magnitude), and revenge trades followed within minutes. Fees were immaterial; the damage was pure execution and impulse control.
What the data shows
The account opened on 6 June and closed 14 episodes by 7 June. The arc is stark: starting balance of $775,707, peak of $774,141 (essentially flat), trough of $751,734 (a 2.89% decline), and final balance of $752,267. All material losses occurred on a single day.
The XYZ100 short on 6 June at 29,108.42 lost $18,264 on a $3.13M notional position—the largest single trade and 34× the median loss. This trade ran for 2.84 hours. Within minutes of closing, the account opened a $321K long on GOLD at 4,310.22, which lost $3,082 in 1.04 hours. This was a textbook revenge trade: opened at 17:42 on 6 June, immediately after the XYZ100 loss. GOLD itself was flagged as an oversized loser (5.82× median). The GOLD loss then triggered another revenge trade: a $159K short on BRENTOIL at 94.48, opened at 17:42:18 on the same day, closed in 11 minutes for -$1,583. A third revenge sequence followed: NFLX lost $0.94, then SP500 opened at 18:47:54 for a $284K short, closed in 43 minutes for -$571.
The account's only profitable side came from small-notional longs on SKHX (+$468), SMSN (+$65), GRASS (+$7), and a short on RENDER (+$10) and NEAR (+$0.66). These wins totalled $553 across four instruments. The remaining six trades were losses totalling -$24,000. Long trades lost $3,032 (42.86% win rate on 7 episodes), while shorts lost $20,408 (42.86% win rate on 7 episodes).
Fees paid were $218.49 gross, with net fee drag of $201.45. Fees are negligible relative to the realised loss of -$27,726. The account was not killed by execution costs; it was killed by position sizing and emotional re-entry.
Trade quality
Win rate: 42.86%. Profit factor: 0.02 (for every dollar won, 50 dollars were lost). Expectancy: -$1,674 per trade. Win/loss ratio: 0.03 (average win of $92 versus average loss of $2,999). These metrics describe an account with no edge and severe asymmetry between winners and losers. The max loss streak was 4 consecutive losses. The max win streak was 3, but those wins were micro-positions that could not offset the oversized losses.
Post-mortems
XYZ100 short, 6 June, 2.84 hours, exit 29,108.42, -$18,264. This was the catalyst. The position reached $3.13M notional—nearly 4× the starting balance. No structural stop is recorded. The loss is 34× the median loss size, marking it as an outlier in position construction. This single trade consumed 78% of the session's total realised loss.
GOLD long, 6 June, 1.04 hours, exit 4,310.22, -$3,082. Opened at 17:42:00 on 6 June, immediately after XYZ100 closed. Notional reached $321K. This trade is flagged as both an oversized loser (5.82× median) and a revenge trade. The duration (1 hour) and timing (within seconds of the prior loss) indicate reactive positioning rather than planned entry. The loss compounded the session's damage.
BRENTOIL short, 6 June, 11 minutes, exit 94.48, -$1,583. Opened at 17:42:18, seconds after GOLD opened. Notional $159K. Flagged as a revenge trade following the GOLD loss. Closed in 11 minutes. This pattern—opening large positions in rapid succession after losses—is the defining behavioural signature of the session.
SP500 short, 6 June, 43 minutes, exit 7,393.75, -$571. Opened at 18:47:54 following a $0.94 loss on NFLX. Notional $284K. Flagged as a revenge trade. The account was still chasing after the morning's losses had already exceeded $22,000.
What the risk simulation reveals
Under a 1% stop-loss rule applied historically, the account would have realised -$50,479 with a max drawdown of -6.71%. Under 2%, simulated loss would be -$100,957 with -13.42% drawdown. Under 4%, simulated loss would be -$201,915 with -26.84% drawdown. These are gross of fees. The simulation shows that without any structural risk control, the account's losses scale linearly with position size. The actual max drawdown of -2.89% reflects the fact that the account did not deploy full leverage on every trade; had it done so, the drawdown would have been catastrophic.
Open positions
No open positions at the time of analysis.
Honest summary
- Micro-cap execution worked. The five smallest trades (SKHX, SMSN, RENDER, GRASS, NEAR) were all profitable or breakeven. The account can identify entry points; the problem is not signal quality on small positions.
- Position sizing on losses was uncontrolled. The two largest losses (XYZ100 and GOLD) were 34× and 6× the median loss, and both reached notional sizes of $3.13M and $321K respectively on a $775K starting balance. There is no evidence of a position-sizing rule or maximum notional limit.
- Revenge trading is the dominant behavioural pattern. Three separate revenge sequences are visible: GOLD opened within seconds of XYZ100 closing; BRENTOIL opened within 18 seconds of GOLD closing; SP500 opened after NFLX. All three revenge trades lost money. The account was active for only 29 hours; the clustering of losses and re-entries on a single day is not noise.
- Sample is too small for long-term inference. The account has been active for 1 calendar day with 14 closed episodes. Sample is too small to assess whether this session represents a systematic edge or a single catastrophic day.