Verdict only traits with a real sample on both sides are stated
Winners were held longer
19.9h vs 12.8h (1.6x)
Losers averaged down harder
0% vs 33% of adds at a worse price
Winners waited for the fill
9% taker vs 15%
Winners vs losers 29 winners · 27 losers · 52% hit rate
| Behaviour | Winners | Losers | Gap | n win / lose | |
|---|---|---|---|---|---|
| Position size | $544K | $508K | 1.1× | 29 / 27 | |
| Hold time | 19.9h | 12.8h | 1.6× | 29 / 27 | |
| Entries per trade | 2.0 | 2.0 | 1.0× | 29 / 27 | |
| Fills per trade | 268.0 | 246.0 | 1.1× | 29 / 27 | |
| Averaged down | 0% | 33% | — | 29 / 27 | |
| Taker share | 9% | 15% | 0.6× | 29 / 27 |
Median, not mean — one $8M outlier should not decide what "their usual
size" is. Highlighted rows are the ones where the two sides genuinely diverge
(1.5× or more, with at least 4 trades each side).
Excludes 25 closed round trips whose entry falls before the served fill history — size and hold are unknown for those, so they cannot answer a question about the entry.
Which side of the book
| Direction | Trades | Win rate | Net P/L | Median hold |
|---|---|---|---|---|
| Long | 53 | 53% | +$727,712 | 19.6h |
| Short | 3 | 33% | −$188,285 | 3m |
When they enter UTC, bar height is trade count, colour is net P/L
Where the money came from 24 coins traded
| Coin | Trades | Win rate | Net P/L | size against the biggest coin |
|---|---|---|---|---|
| ASTER | 4 | 75% | +$293,660 | |
| XPL | 6 | 50% | +$253,374 | |
| BTC | 5 | 0% | −$167,716 | |
| LINK | 1 | 100% | +$110,567 | |
| HYPE | 4 | 75% | +$109,055 | |
| SOL | 1 | 0% | −$89,594 | |
| xyz:SKHX | 1 | 100% | +$62,347 | |
| ETH | 6 | 50% | −$61,574 |
The shape of it
Closed round trips
56
29 up · 27 down
Sizing discipline
136%
how much size varies around its own average
Exit style
0%
of winners were one order in and one out
(losers 7%)
Top 3 winners
47%
of gross profit
Biggest position
$6.02M
BTC
· made −$39,413
Worst trade
−$617,528
XPL
· 4.2d held
Forced closes
0
liquidated out of 56
Reversals
1
flipped rather than closed flat
Closed round trips only — an open position has no outcome yet, so
including it would flatter or punish the record on an unrealised number. Reconstructed
flat→flat from the fill stream, with same-millisecond fills re-chained by
startPosition.