Bottom line up front
Only the most recent public fills are visible, so this audit covers the data covered rather than full account history. The account is profitable in this window: +0.03% and $2,397 realised PnL in the data covered, but the headline obscures severe behavioural dysfunction. Short-side trades (MU, NVDA, INTC, AMD) generated $2,881 profit; long-side trades haemorrhaged $484. The single dominant pattern is revenge trading and averaging down into losing XYZ100 longs—five re-entries after closed losses, each one negative—combined with oversized positions in instruments where the account has no edge. The account survived this window because two large short positions (MU +$2,215, NVDA +$286) offset systematic losses elsewhere. The risk simulator shows that a 1% hard stop rule would have produced $118k profit; the actual account took $363 in fees to make $1,616 gross, a ratio that signals poor execution discipline.
What the data shows
This is a 1-day window (18 May to 20 May 2026) with 159 closed episodes and 59 open positions. The account started with approximately $9.33M in equity and closed the window at the same level, having cycled through $12.07M in gross volume. The highest balance in this window was $9,328,162.64; the lowest balance was $9,325,088.22, a deepest decline in this window of 0.01%.
The profit came entirely from short-side execution. Shorts won 35.3% of the time and produced $2,881 net PnL; longs won 32.4% of the time and lost $484. The short wins were concentrated: MU (18.79 hours, entry unknown, exit 677.6, +$2,215), NVDA (12.46 hours, 223.4 to 222.61, +$286), INTC (18.78 hours, 109.11 to 108.21, +$266), and AMD (18.19 hours, exit 420.81, +$165). These four trades account for $2,932 of the $2,881 short profit. The long side was a graveyard: 72 episodes in XYZ100 alone, 30.6% win rate, -$218 net. SP500 (21 episodes, 19% win rate, -$81), USA500 (24 episodes, 29% win rate, -$45), and NATGAS (3 episodes, 67% win rate, -$413) added to the bleed.
Fees consumed 22.5% of realised PnL. The account paid $364 in gross fees on $1,616 realised profit, leaving $1,252 after execution costs. This ratio is poor for a $9M account and suggests either high slippage, frequent maker-taker reversals, or both. Maker percentage was 16%, indicating the account is predominantly a taker.
Trade quality
Win rate 34%, profit factor 2.71, expectancy $15.08 per trade, win/loss ratio 5.28. The profit factor means every dollar lost was offset by $2.71 in wins—respectable in isolation, but the distribution is skewed. The account won 54 trades and lost 105, so the wins were larger on average ($70.29 vs. $13.32 loss), but the frequency of losses is high. Expectancy of $15 per trade is positive, but with 218 total episodes and $363 in fees, the edge is thin. The longest loss streak was 12 consecutive losing trades.
Post-mortems
XYZ100 long, 19 May 2026, 0.36 hours, 28,881.04 entry to 28,860.39 exit: -$57.62. This trade was flagged for averaging down, oversized loser, and revenge trade. The account opened at 28,881.04 and exited 21 minutes later at 28,860.39, a 7-basis-point move against a $78k notional position. The structural stop was 4% away. This was a revenge trade following a previous XYZ100 loss and involved multiple add-ins as the position moved against the account. The loss was small in absolute terms but symptomatic of the pattern: re-entering a losing instrument with size, holding for minutes, and exiting at a loss.
NATGAS long, 19 May 2026, 7.36 hours, 3.18 entry to 3.23 exit: -$627.73. Flagged for oversized loser and revenge trade. The account held a $74k notional position for over seven hours, exited at a 1.6% profit on the move (3.18 to 3.23), yet still lost $628. This indicates the position was sized such that a small adverse move early in the hold consumed the eventual profit. The structural stop was 4% away. This was a revenge trade following a loss in a different instrument (SP500), suggesting the account was chasing size to recover from earlier losses rather than trading the setup.
What the risk simulation reveals
Under a 1% hard stop rule applied historically to this window, the account would have realised $118,098 profit with a deepest decline in that simulation of -0.37%. Under 2%, $236,197 profit and -0.74% decline. Under 4%, $472,393 profit and -1.43% decline. These are gross-of-fees figures. The actual account realised $2,397 (including fees) with a -0.01% decline. The gap between simulated and actual is the cost of not using stops: the account allowed losers to run and averaged into them, which occasionally worked (MU, NVDA) but more often compounded losses (XYZ100, NATGAS, SP500). The simulator assumes stops are hit; the actual account ignored them.
Open positions
Five open positions, none with stops in place:
- SOL short, 10x leverage, entry 84.2521, unrealised +$4.93. Minimal exposure, small profit.
- BNB long, 10x leverage, entry 655.531, unrealised -$97.10. Underwater, no stop.
- LTC long, 10x leverage, entry 54.6578, unrealised -$68.82. Underwater, no stop.
- ARB long, 10x leverage, entry 0.09296, unrealised +$0.01. Negligible.
- DOGE short, 10x leverage, entry 0.10446, unrealised +$70.20. The only meaningful winner.
The BNB and LTC longs are both in deepest decline in this window with no protective stops. Combined unrealised loss is $166. All positions are 10x leverage with no stops, consistent with the account's observed behaviour of letting losers run.
Honest summary
- Short-side execution is the only edge visible in this window. MU, NVDA, INTC, and AMD shorts accounted for $2,932 of $2,881 total short profit. The account can identify short setups and hold them. Long-side trades are a liability.
- Revenge trading and averaging down are systematic losses. Five XYZ100 re-entries after closed losses each lost money. The NATGAS and XYZ100 oversized losers were both revenge trades. The account is chasing losses with size rather than respecting the setup.
- Absence of stops despite structural defaults is the critical vulnerability. The risk simulator shows that a 1% rule would have produced 49x