Bottom line up front
Only the most recent public fills are visible, so this audit covers the data covered rather than full account history. This account is exceptional in the data covered: $785k realised PnL on $54.3m gross volume, 74% win rate, 50.71 profit factor, and $25.3k expectancy per closed trade. The headline masks a severe structural fragility: the account survived a deepest decline in this window of -99.82%, recovering from $3.6k to $697.8k, and the two largest losses—a $10.9k SOL long and a $3.2k BTC long—were both oversized revenge trades built on averaging down after prior losses. Short-side edge is real and dominant; long-side attempts are a liability.
What the data shows
The account opened on 18 November 2025 and has completed 31 closed episodes in the data covered. The arc is violent: highest balance in this window reached $1.99m on 19 November, then collapsed to $3.6k on 3 December, a swing of $1.98m in two weeks. The account has since recovered to $697.8k, but the recovery is built on a narrow edge: shorts account for $788.7k of the $790.1k realised PnL, while longs lost $3.7k. BTC shorts alone generated $353.9k; ETH shorts $241.8k; HYPE shorts $192.5k. The short-side win rate is 76.47% across 17 episodes. By contrast, the long side is 71.43% win rate across 14 episodes but deeply negative in absolute terms.
Fees are negligible relative to gross PnL: $5,059.82 paid on $54.3m volume, a 0.64% ratio to realised PnL. The account is a 94.5% maker, so rebates offset much of the cost. Net fee drag is $5,059.82, immaterial to the headline.
The three largest wins are all shorts: BTC short from 90,795 to 90,757 on 18–20 November (+$248.3k), ETH short from 3,007.92 to 2,968.83 on 18–25 November (+$241.8k), and HYPE short from 38.12 to 36.56 on 18–25 November (+$192.5k). These three trades account for $682.6k of the $790.1k realised PnL. All three were clean, unencumbered by behavioural flags.
The three largest losses are all longs, and all carry multiple behavioural red flags. SOL long from 140.17 to 136.06 on 18–25 November: -$10.9k, 67 averaging-down events, max position $288k notional, flagged as oversized loser (91x median loss). BTC long from 87,625.67 to 87,521.91 on 24–25 November: -$3.2k, flagged as averaging-down, FOMO re-entry, oversized loser (26.6x median), and revenge trade following a -$1.4k loss. BTC long from 83,887.96 to 83,676.81 on 21 November: -$1.4k, 0.34 hours, oversized loser (11.4x median), revenge trade after a -$65.97 loss.
Trade quality
Win rate of 74.19% across 31 closed episodes. Profit factor of 50.71: for every dollar lost, the account made $50.71. Average win of $34,817.63; average loss of -$1,974.02. Win/loss ratio of 17.64. Expectancy of $25,323.01 per closed trade. These numbers reflect a heavily skewed distribution: a small number of large, clean shorts dominate; a larger number of small, noisy longs and revenge trades drag the average down but do not overwhelm the edge.
The 7-win streak (22–24 November) and 3-loss streak (21 November) both occurred in the same 72-hour window, suggesting high volatility and reactive position-taking rather than systematic execution.
Post-mortems
SOL long, 18–25 November, entry 140.17, exit 136.06, -$10,894.59. Opened with 67 averaging-down events over 160 hours, max position $288.2k notional. Structural stop (ATR 14 1h) was 2.21% away; the trade ran 2.93% against entry before closing. This was the largest loss in the data covered and is flagged as both an oversized loser (91x median loss) and a revenge trade. The averaging pattern suggests the trader was fighting a losing position rather than managing it. No exit discipline visible.
BTC long, 24–25 November, entry 87,625.67, exit 87,521.91, -$3,186.14. Opened 33.69 hours after a -$1.4k BTC long loss on 21 November. Max position $1.126m notional, flagged as averaging-down, FOMO re-entry, oversized loser (26.6x median), and revenge trade. Structural stop was 1.35% away; the trade moved 0.12% against entry and closed immediately. This is a textbook revenge trade: large notional, short duration, tight stop, opened after a prior loss in the same coin.
What the risk simulator reveals
Under a 1% hard stop rule applied historically, the account would have realised -$68.2k with a -78.16% deepest decline in this window, win rate collapsing to 58.62%. Under 2%, the loss widens to -$136.4k and deepest decline to -135.27%. Under 4%, -$272.8k and -213.16% deepest decline. The simulator stopped 6 episodes early in all three scenarios, indicating that the largest positions would have been cut before they matured. The account's actual profitability depends entirely on the absence of hard stops; the moment you enforce risk discipline, the edge inverts.
Open positions
No open positions at the time of this snapshot.
Honest summary
- Short-side edge is real and repeatable. Three large, clean shorts (BTC, ETH, HYPE) generated $682.6k of the $790.1k realised PnL. Win rate on shorts is 76.47%. These trades had structural stops in place and were closed at profit without averaging or re-entry.
- Long-side attempts are a liability masked by short-side dominance. Longs are -$3.7k net across 14 episodes. The two largest losses are both longs with severe behavioural flags: averaging down, revenge trading, and oversized notional relative to median loss. The SOL long alone is 91x the median loss.
- Risk management is absent on losing trades. Structural stops are defined but not enforced. The SOL long and both BTC revenge longs all breached their ATR-based stops without triggering exits. Averaging down on losers is systematic, not accidental.
- The account's survival depends on short-side edge overwhelming long-side leakage. The deepest decline in this window of -99.82% was recovered by re-establishing shorts, not by improving long-side discipline. A risk simulator shows that any hard stop rule would have inverted the PnL to losses.
- Sample is concentrated in a two-week window. The data covered covers 18 November to 30 March, but