Bottom line up front
Open BTC long dominates this wallet: $15.9M notional, -$2.41M unrealised, -525% unrealised ROE, 40x leverage, cross margin, liquidation at $46,519. This is still open exposure; the unrealised loss is not locked in, and the liquidation level remains the downside boundary. The account is near break-even in the data covered—$27.6k realised profit across 19 closed trades—but that headline masks a catastrophic open position that has erased nearly all closed-trade gains and sits $2.4M underwater. Only the most recent public fills are visible, so this audit covers the data covered rather than full account history.
What the data shows
The account opened in mid-June 2025 and has been active for 367 days across 39 total episodes (19 closed, 20 open). The data covered captures approximately $100.2M in closed trade volume and $184.5M in gross volume including open positions. The highest balance in this window reached $26.6M on 13 August 2025; the lowest balance fell to $3.97M on 4 February 2026. The deepest decline in this window was -85.1%, a catastrophic deepest decline in this window driven almost entirely by the current open positions.
Closed trades generated $27.6k in realised profit, but this figure is misleading. The account paid $69.6k in fees, meaning the underlying edge before execution costs was approximately $97k. BTC was the primary edge: six episodes, 50% win rate, $553k realised profit. Long positions outperformed short positions significantly: longs returned $372k on 63.6% win rate; shorts lost $344k on 37.5% win rate. The short-side weakness is material—every short episode in the closed book was a loss or break-even.
The two largest closed losses reveal behavioural fractures. MKR long, opened 13 June, closed 5 September, lost $547k on a $2.27M position—7.4x the median loss size. The trade moved -19.1% against the entry before exit, with a structural stop 4% away that was never triggered. HYPE long, opened 18 July as a re-entry after a previous close on the same day at $44.54, re-entered at $44.90, and closed 27 August at $49.74 for a -$194.5k loss. The re-entry pattern is explicit in the data: closed a HYPE position at profit, then immediately re-entered at a marginally higher price into weakness.
Trade quality
Win rate is 52.6% across 19 closed episodes. Profit factor is 1.02—barely above break-even. Expectancy is $1,455 per closed trade. Win/loss ratio is 0.92, meaning average winners ($128.9k) are only slightly larger than average losers ($140.2k). These metrics describe an account with no meaningful edge in closed trades. The account is profitable only because it closed more winners than losers by a narrow margin; the edge is fragile and entirely dependent on position sizing discipline, which the open book shows has collapsed.
Post-mortems
MKR long, 13 June to 5 September, $1,920.68 to $1,705.13, -$547.1k. Opened at $1,920.68 and immediately faced -19.1% adverse move. The structural ATR-based stop was 4.05% away; the trade was allowed to run -19.1% into the red before exit. This is a textbook oversized loser: the position notional ($2.27M) was sized at a scale inconsistent with the account's risk tolerance, and the stop discipline was absent. The trade ran for 2,025 hours (84 days) in a losing state, accumulating funding costs and psychological pressure.
HYPE long, 18 July to 27 August, $44.90 to $49.74, -$194.5k. This is a FOMO re-entry. The account closed a HYPE long on 18 July at $44.54 for a profit, then immediately re-entered the same coin at $44.90 on the same timestamp. The re-entry lost $194.5k over 39 days. The structural stop was 4% away and was never triggered. This pattern—closing a winner, then chasing the same coin back in at a worse price—is a classic revenge or FOMO signal and is explicitly flagged in the behavioural data.
What the risk simulation reveals
Under a 1% stop-loss rule applied historically, the account would have realised $38.2k with a -8.9% deepest decline in this window. Under 2%, $76.4k with -17.8% deepest decline. Under 4%, $152.8k with -35.6% deepest decline. All three simulations stopped 3 episodes early, indicating that hard stops would have prevented some of the largest losses. The actual realised PnL of $27.6k sits between the 1% and 2% simulated outcomes, suggesting that the account's actual stop discipline was weaker than even a 1% rule but stronger than no stops at all. The simulation is gross of fees.
Open positions
BTC long ($15.9M notional, -$2.41M unrealised, -525% ROE, 40x leverage, liquidation $46,519) is the dominant exposure and the primary risk. Entry was $73,828.40; mark price is $64,134. The position has accrued $485.3k in funding costs since opening. No stop is in place. ETH long ($12.6M notional, -$1.91M unrealised, -329% ROE, 25x leverage, held 392 days, liquidation $1,165.19) is the second-largest open position, also underwater and without a stop. DOGE long ($619.7k notional, -$781.9k unrealised, -558% ROE, 10x leverage, held 155 days) is deeply underwater. CRV long ($338.8k notional, -$283.2k unrealised, -455% ROE, 10x leverage, held 357 days) and XRP long ($186.8k notional, -$13.2k unrealised, -132% ROE, 20x leverage) are smaller but all long, all underwater, and none have stops in place. Total open notional is $34.3M; total unrealised loss is -$6.85M. The account is leveraged long across five coins with no hedges and no stops.
Honest summary
- Closed-trade edge exists but is marginal. BTC longs worked; the long/short asymmetry is real and measurable. A 52.6% win rate and 1.02 profit factor are not nothing, but they leave zero room for execution error or fee drag.
- Position sizing and stop discipline have collapsed. The MKR trade was 7.4x the median loss; the HYPE re-entry was a behavioural red flag that cost $194.5k. The current open book shows no stops across five positions totalling $34.3M notional, with the BTC position alone carrying -525% ROE and a liquidation price $17.6k below mark.
- Funding costs are a material drag. The BTC position alone has accrued $485.3k in funding since opening. ETH has accrued $1.32M data-covered. These are not small numbers and are eroding the account's margin buffer continuously.
- The data covered is capped at 10,000 fills.