Bottom line up front
This account is profitable at +0.003% ($221 on a $6.9M base), but the headline obscures a volatile and behaviourally compromised trading pattern. The account peaked at $8.08M before declining 26.68% to a trough of $5.92M, then recovered to close near breakeven. Short-side trades (73% win rate) carried the PnL; long-side attempts (66% win rate) leaked value. The core problem is not edge but execution: five revenge trades, five FOMO re-entries, and two catastrophic ETH losses (−$128 and −$108) that dwarf median losses by 70–84x, all compressed into the first 24 hours of trading.
What the data shows
The account opened on 23 May 2026 and has traded 239 closed episodes over 31 days. The first 12 hours were chaotic: rapid-fire position cycling across BTC, ETH, and SOL, with overlapping long and short attempts on the same coins. This period generated five revenge trades (opening new positions immediately after small losses on unrelated coins) and established the two largest losses of the entire window.
Money was made on the short side. Short-trade PnL totalled $194.20 with a 73% win rate; long-trade PnL was $26.95 at 66% win rate. SOL and BTC were profitable (SOL: $139.60 realised PnL, 71% win rate across 55 episodes; BTC: $68.88 realised PnL, 75% win rate across 109 episodes). ETH was a net loser at −$27.83 realised PnL despite a 60% win rate, because three trades (two shorts, one long) lost $238.66 combined. The largest single loss was an ETH short entered at 2109.78 on 23 May, exited at 2114.46 after 10 minutes, for −$128.53 notional on a $39.6k position. The second-largest was an ETH long entered at 2119.64, exited at 2113.05 after 5 minutes, for −$107.69 on an $18k position, flagged as both a revenge trade and an averaging-down episode.
Fees consumed 17.51% of realised PnL. Gross fees paid were $69.41, but the account received a net rebate of $39.65 (98.25% maker rate), so net fee drag was −$39.65. This means realised PnL before fees was $396.45; fees and rebates netted to −$175.30, leaving the final $221.15. Without the maker rebate, the account would have closed at −$135.
The account's peak-to-trough cycle (23 May to 9 June) saw the balance swing from $8.08M to $5.92M, a $2.16M decline. The recovery from trough to close was partial, ending $1.18M below peak. This volatility was driven by concentration: the largest open position at any point reached $347k notional (a synthetic instrument trade on 2–3 June that closed profitably at +$71.67), and multiple positions in the $10–40k notional range were held simultaneously during the first day.
Trade quality
Win rate of 69.46% is strong. Profit factor of 1.3 means gross wins were 1.3x gross losses. Expectancy of $0.93 per trade is marginal given the $6.9M account size and the fee structure. The win/loss ratio of 0.57 (average win $5.77 vs. average loss −$10.09) shows losses are nearly twice the size of wins. This is the core fragility: the account wins frequently but small, and loses infrequently but large. The max loss streak was 3 consecutive losses; the max win streak was 12.
Post-mortems
ETH long, 23 May, 21:24–21:33 UTC
Entered at $2119.64, exited at $2113.05, −$107.69 loss on $18.04k notional. Flagged as averaging-down, oversized loser (70x median loss), and revenge trade (opened after a −$4.30 BTC loss). The position was held 5 minutes. This was the second-largest loss in the window and set the tone for the session.
SOL long, 23 May, 21:32–22:01 UTC
Entered at $86.13, exited at $85.86, −$40.62 loss on $9.49k notional. Flagged as averaging-down, oversized loser (4x median loss), and revenge trade (opened after a −$3.78 SOL loss). The position was held 30 minutes. This trade was part of a cluster of SOL re-entries: the account closed this position, re-entered at $86.09 (15 seconds later) for +$0.05, closed again, re-entered at $86.33 for +$0.18, and continued cycling. The pattern suggests reactive, scale-dependent decision-making rather than thesis-driven entry.
What the risk simulation reveals
Under a 1% stop-loss rule applied historically, the account would have realised $170,132 PnL with a −0.12% max drawdown. Under 2%, $340,265 PnL and −0.25% drawdown. Under 4%, $680,530 PnL and −0.5% drawdown. These are gross-of-fees figures. The simulation shows that mechanical position sizing discipline would have transformed the account from near-breakeven to substantially profitable, because the core edge (69% win rate, short-side bias) would have been preserved while the tail losses (ETH −$128, −$108, −$102) would have been capped. The actual max drawdown of −26.68% reflects the absence of any binding risk constraint during the first 24 hours.
Open positions
No open positions at close.
Honest summary
- Short-side edge is real. 73% win rate on shorts across 73 episodes, with SOL and BTC both profitable. The account can identify reversals and hold them for 100+ hours when conviction is present (e.g., ETH short 24–29 May, +$288.71).
- Revenge trading and position cycling destroy discipline. Five revenge trades and five FOMO re-entries are documented in the first 24 hours. The account opened new positions within seconds of closing losing trades on unrelated coins, and re-entered the same coin multiple times within minutes. This pattern is incompatible with systematic risk management.
- Oversized losers on ETH are structural, not random. Three ETH trades account for −$238.66 of the −$175.30 net loss. All three were opened in the first 12 hours, all three were flagged as revenge or averaging-down trades, and all three were sized 3–84x the median loss. The account has no visible position-sizing rule that scales with account volatility or recent drawdown.
- Maker rebate masked the true cost structure. The 98.25% maker rate generated a −$39.65 net rebate, which was the only reason the account closed profitably. Without it, the account would have lost $135. This is not a sustainable edge; it is fee arbitrage on a large account.