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 near break-even in that window: +0.08% realised PnL ($209.74 on a $255k balance), but the headline masks severe execution and behavioural friction. Fees consumed 61% of gross trading profit, and a catastrophic loss cluster on 10 July—five oversized reversals in rapid succession—nearly wiped the month's gains. The deepest decline in this window reached -24.45%, and the account recovered only because short-side edge in INTC and EWY offset systematic losses in AMD, DRAM, CRCL, and NVDA.
What the data shows
The account opened on 6 May 2026 and has completed 641 closed trades across 65 days. Net realised PnL is $599.50 before fees; fees of $366.64 reduce that to $232.86, leaving the account at +0.08% after execution costs. The starting balance estimate is $255,181.94; current balance is $255,391.68. The highest balance in this window was $253,088.54 on 3 July; the lowest balance was $104,287.82 on 20 May—a swing of nearly $149k in the first fortnight.
Money was made in four instruments: cash:INTC ($265.56 realised), xyz:EWY ($153.47), xyz:INTC ($148.76), and xyz:SKHX ($69.65). These four accounted for $637.44 of the $599.50 gross realised profit. Losses came from AMD (-$265.16), SMSN (-$162.01), DRAM (-$112.07), CRCL (-$84.15), and NVDA (-$50.18). The long side lost $57.83; the short side gained $267.56. This asymmetry is material: long-side win rate was 39.32%, short-side win rate was 43.71%.
Fees are the second-order problem. At $366.64 on $6.74m in closed trade volume, the fee ratio is 5.4 basis points—reasonable for retail—but the fee-to-PnL ratio of 61% reveals the core issue: the account is generating small, frequent wins and losses, and execution costs consume more than half the edge. The profit factor of 1.09 means gross wins are only 9% larger than gross losses; after fees, the margin evaporates.
Trade quality
Win rate is 41.5% across 641 closed episodes. Profit factor is 1.09. Expectancy is $0.33 per trade. The win/loss ratio is 1.53: average win is $10.05, average loss is $6.57. These numbers describe a marginal edge that exists only before fees. the account is only marginally profitable in the data covered; it is fee-neutral with a small realised gain that fees nearly eliminate. The max win streak was 9 trades; the max loss streak was 13. The longest losing run is longer than the longest winning run, and the account spent the first two weeks of May in sustained deepest decline in this window.
Post-mortems
xyz:AMD short, 10 July, 2.26 hours, entry 553.46, exit 558.48, -$135.25. This is the single largest loss in the data covered and carries both oversized-loser and revenge-trade flags. The position was opened at $553.46 and exited at $558.48 after 2 hours 16 minutes, a 0.9% adverse move. The notional was $8,836.24—well above the median loss size. This trade followed a loss in cash:META and was part of the 10 July collapse sequence.
xyz:AMD short, 10 July, 0.4 hours, entry 555.56, exit 559.03, -$62.33. Opened 2 hours after the first AMD short closed, this trade averaged down into the same instrument at a higher price (555.56 vs 553.46) and exited at 559.03 after 24 minutes. The notional was $8,826.93. The averaging-down flag and oversized-loser flag indicate the account was doubling down on a failed thesis without waiting for mean reversion or new signal.
Both AMD trades on 10 July were part of a broader pattern: five oversized losses closed within 6 hours (AMD -$197.58, CRCL -$71.86, DRAM -$68.61, SKHX -$50.28, and one other). This cluster suggests panic liquidation or forced exits under margin pressure, not deliberate trade management.
What the risk simulator reveals
Under a 1% hard stop rule applied historically, the account would have realised -$10,605.54 PnL with a deepest decline in this window of -5.08%. Under a 2% rule, -$21,211.07 and -10.09%. Under a 4% rule, -$42,422.15 and -19.91%. The simulator stopped 8 episodes early in each scenario due to data quality constraints. The win rate remained stable at 43.15% across all three rules, meaning stops would not have improved hit rate—they would have simply capped losses and forced earlier exits. The actual account, with no hard stops, suffered a -24.45% deepest decline and recovered only because the short-side edge held.
Open positions
Five positions are currently open. The largest by notional is xyz:PLTR long: $6,635.43 notional, 52.425 contracts, entry 126.688, mark 126.57, unrealised -$6.19 (-0.37% ROE), 4x leverage, isolated margin, held 0 days (opened 10 July), liquidation price 100.0189. No stop is in place. The position is underwater by 37 cents and has 26.67 points of downside to liquidation.
xyz:HOOD short is the second-largest: $4,043.82 notional, 36.3 contracts, entry 112.814, mark 111.4, unrealised +$51.36 (+5.02% ROE), 4x leverage, isolated margin, held 64 days, liquidation price 134.545. No stop is in place. This position is profitable and has been held since 7 May.
xyz:TSLA short: $3,471.47 notional, 8.525 contracts, entry 409.336, mark 407.21, unrealised +$18.13 (+2.08% ROE), 4x leverage, cross margin, held 64 days, liquidation price 16670.0411. No stop.
xyz:COIN short: $1,188.66 notional, 7.466 contracts, entry 158.94, mark 159.21, unrealised -$2.02 (-0.68% ROE), 4x leverage, isolated margin, held 0 days, liquidation price 189.3088. No stop.
xyz:META long: $4,467.85 notional, 6.686 contracts, entry 668.589, mark 668.24, unrealised -$2.34 (-0.21% ROE), 4x leverage, cross margin, held 0 days, no liquidation price. No stop.
Total open notional is $19,807.63 across five positions. Combined unrealised PnL is +$58.94. None of the five positions have stops in place.
Honest summary