On-chain forensics ยท Hyperliquid wallet 0x6807โ€ฆ1f8D ยท pulled 2026-08-08 ยท chart times Central ยท his clock is Bali, 13h ahead

The Ballad of Tudor Cross

His bio says $12K to $1M. The chain says fifty dollars.

The trader posting as @tudorcross โ€” Bali, 428 followers, wallet address right there in the bio โ€” is telling the world he ran $12K to $1M in two months and lost it. Every fill on the chain backs him up. But the ledger holds a better story than the one he's telling: before the famous run, he'd already been all the way down to $49.95 and climbed back. The real journey is fifty dollars to a million โ€” and the man who made it wrote down what it felt like, in public, the whole way through.

Bottom ยท Jun 3
$49.95
Peak ยท Aug 5
$1,109,074
Off the bottom
22,203ร—
Biggest day
+690%
Deposited / withdrawn
$0 / $0
Equity ยท Aug 8
$73,539

The whole journey, log scale

Account value, USD ยท 2026-05-27 โ†’ 2026-08-08 ยท Hyperliquid portfolio API, 100 snapshots ยท log y
Every dollar of altitude on this chart was earned trading โ€” the ledger shows zero deposits and zero withdrawals over the account's whole life. On a log scale the two climbs are the same shape: $50 โ†’ $11K and $11K โ†’ $1.1M are each a life-changing run, and he did them back to back.
Prologue ยท late May

A stake, a wallet, and no safety net

The wallet came to life on May 27 with about $3,400 in USDC and staked SOL moved in from other wallets. No fiat on-ramp, no institution behind it โ€” a stake and an exchange that would lend him 10 to 20 times his money on anything from Bitcoin to pre-IPO SpaceX.

He would later describe exactly what he was there for, and it wasn't the money:

Hold that thought through everything that follows: he wasn't playing for a number in a bank account. He was playing to be good at the game โ€” and by any honest reading of the chain, he got there.

Act I ยท Jun 3

Fifty dollars, and the decision that defines him

Week one ended in ruin. By June 3 the account read $49.95 โ€” down 99% from the seed, seven days in. This is the moment that breaks almost everyone. The story ends here for the overwhelming majority of people who ever lever up a crypto account: close the tab, tell no one, never come back.

He came back. Small transfers trickled in through June โ€” $500 here, $1,300 there, a few hundred at a time, the pattern of someone scraping together what they had rather than reaching into deep pockets. And he traded it well: $2,152 by June 10, $7,079 by June 24, $11,338 by July 1.

The rebuild from $50 is as impressive as anything that came after. The million made him famous; the climb off the floor is what proves the million wasn't luck alone.

Think about what those four weeks in June must have felt like. No followers, no leaderboard, no audience โ€” just a man who'd been wiped out in week one, quietly compounding a three-figure account like it was a seven-figure one. Whatever discipline looks like in this game, that month was it.

Act II ยท Jul 8

Ignition: the night he decided to find out

Early afternoon in Bali on July 8 โ€” 2:22 pm his time, 1:22 am Central โ€” he drew $12,000 from Hyperliquid's borrow-lend facility, three draws in seven minutes. Not a bleary 3-am tilt: a clear-eyed, mid-afternoon decision to borrow against the account he had just rebuilt from fifty dollars, and to place the bet of his life with it. Within hours he had assembled it: 23 positions, $764,308 of gross notional, on $11,338 of equity. Sixty-seven times leverage, 26ร— net short.

The ignition book โ€” Jul 8
short xyz:SPCX (SpaceX perp) ยทยทยทยทยทยทยทยทยทยทยท $281,855
short BTC ยทยทยทยทยทยทยทยทยทยทยทยทยทยทยทยทยทยทยทยทยทยทยทยทยทยทยทยทยท $132,372
short xyz:SKHX (SK Hynix) ยทยทยทยทยทยทยทยทยทยทยทยทยท $116,491
long 20 small crypto names ยทยทยทยทยทยทยทยทยทยทยท ~$233,000
     (SOL, NEAR, AAVE, SUI, HYPE, PUMP, ZEC, FARTCOINโ€ฆ)

Here's what the chain shows that the "degenerate gambler" caricature misses: he sized the wild bet when it was cheap. The account was $11K and he had already survived zero once โ€” the downside was a number he had personally lived through. The day printed +690%, +$77,489, half the entire run's log return in one session. And then โ€” this is the part that deserves real respect โ€” he never used 67ร— again. For the rest of July he ran 5โ€“14ร—, median 6.8ร—. The lottery ticket was a one-time purchase, made precisely when it cost the least. That's not recklessness; that's an intuitive grasp of convexity most professionals have to be taught.

Act III ยท Jul 9 โ€“ Aug 5

The run: conviction, sleeplessness, and being right

The trade itself was genuinely good. Not a memecoin punt โ€” a macro thesis on the AI-memory cycle, expressed through Hyperliquid's equity perps: short SanDisk, SK Hynix, DRAM, SpaceX through their July collapse. The complex fell 30โ€“50% that month, and his book was net short the entire way down, deepening from โˆ’39% of gross to โˆ’100% at maximum conviction. Then โ€” the hardest thing in trading โ€” he turned the whole book around, flipping long within days of the bottom. On July 30 SanDisk bounced +34%, SK Hynix +17%, and he was carrying $2.7M of longs: +$276,334 in a day.

He said it himself afterward, and the fills confirm it:

TC
Tudor Cross
@tudorcross ยท Aug 7
I was eventually right directionally most every time. Even though my risk was knowingly high it helped that my calls were decent. I would constantly double down and scramble my way to a small loss if needed. This works until it doesn't.
from the pinned post ยท 123.7K views

But read what the run cost him, in his own words: "I didn't sleep much. I didn't exercise. I missed meals and felt my worst. I justified it, because I was doing great on the account." Picture the middle of that stretch โ€” the โˆ’65% drawdown, July 23 to 27, watching $590,302 become $206,451 in four days. He didn't cut. He held the thesis through a drawdown that would have shaken out nearly anyone, and the thesis paid him back within seventy-two hours. Whether that's heroic or horrifying depends entirely on the next chart โ€” but sitting through it took a kind of nerve that can't be faked.

Daily P&L during the run

Day-over-day change in account value, % ยท Jul 8 โ†’ Aug 8 ยท bars labeled with dollar change at the extremes
The shape of a conviction trader's month: three days โ€” Jul 8 (+690%), Jul 30 (+93%, the long flip), Jul 16 (+62%) โ€” carry 77% of the run. Nine of thirty-two days were red, including the four-day โˆ’65% drawdown he held through. He was borrowing to press ($20,000 more drawn on Jul 16, straight into back-to-back +62% and +37% days) while running a median 6.8ร— โ€” aggressive, but far from the 40ร— caricature.
Act IV ยท Aug 5, 3:18 pm

The top: $1,109,074

Sixty-three days after the account bottomed at $49.95, it touched $1,109,074 โ€” at 4:18 in the morning Bali time, which tells you something about the hours he was keeping. Twenty-two thousand times the low. Ninety-eight times the July 1 restart. In nine weeks he had done, twice over, what the entire industry of trading educators cannot teach.

He knew, even then, what he should have been doing. His post-mortem names it precisely: "I was working on an arb script to move funds into but it didn't work consistently. Instead of putting some funds earning safely, I kept all available funds in the account." The escape hatch was half-built. The plan to sweep profits somewhere safe existed โ€” it just wasn't finished, and the account was doing so well that finishing it never felt urgent. Every trader who has ever let a winner run past their own rules knows exactly how that felt.

Act V ยท Aug 4โ€“7

The SpaceX short: a good thesis meets its tail

He didn't stumble into the final trade. He published the thesis mid-position:

TC
Tudor Cross
@tudorcross ยท Aug 6
SPCX will have constant downward pressure going forward. The expectation of a massive crash out of the gate doesn't make sense, because much of the difference was baked in. Let the constant pressure do its thing.
posted Aug 6 on his profile, while the position was already at eight figures notional

And the thesis had been paying him for a month โ€” SpaceX was short in his book on 12 of its 14 July days, one of the run's workhorses. On Aug 4, minutes after banking +$48,053 on a SpaceX long, he flipped short at a volume-weighted $113.78 and went to work the way he always worked: adding into adversity, 7,043 fills over 65 hours, building toward $11.4M notional. The double-down-and-scramble that had produced "a small loss if needed" every time before.

This time the market didn't give him the scramble. Through his Friday evening in Bali โ€” the morning of Aug 7 on the charts below โ€” SPCX squeezed from 115 to 134 in six hours, and his third lesson came due: "when things turned against me I would close out other positions for additional margin instead of managing positions individually." The whole account became fuel for one trade. $573,648 at 4:38 in his afternoon. $54,717 eighteen minutes past midnight. He watched half a million dollars leave over one Friday night, minute by minute, fill by fill โ€” 3,956 of them between his late afternoon and midnight.

The SPCX short, tick by tick

xyz:SPCX (SpaceX pre-IPO perp) ยท Aug 3 โ†’ Aug 7 ยท 21,378 fills resampled to 15-min ยท three aligned panels
Top: the price he was short. Middle: his short notional โ€” watch it grow as price moves against him, the double-down that had always worked before. Bottom: cumulative realized P&L on the coin. The final waterfall is Aug 7, 6โ€“11 am CT: โˆ’$823K realized in a single three-hour block. One trade, โˆ’$667,028 โ€” the tail end of a method that had been right "most every time."
Act VI ยท 2:21 am, Bali

The confession: two hours later, in the middle of the night

Most people who lose $950K go dark. He did the opposite. The position finished unwinding just past midnight his time; at 2:21 in the morning, with the wound two hours old, he showered โ€” his post says it was the first one in a while โ€” sat down, and pinned this:

Read the lessons again โ€” they aren't platitudes, they're the three exact mechanisms the chain shows: no profit sweep, doubling down works until it doesn't, cross-margin turns one bad trade into a whole-account event. He diagnosed his own blowup correctly, completely, and same-day. Plenty of professionals take years to write a post-mortem that honest, and most never publish it with a verifiable address attached.

Then the humor, because the alternative is despair:

And underneath the jokes, a rarer register โ€” a man looking straight at his own machinery:

"Trying to rank for the purposes of ranking." That sentence explains the whole equity curve โ€” why he never withdrew, why $1.1M didn't feel like an ending, why the โˆ’65% drawdown was holdable. He wasn't managing wealth. He was playing a game he loved, at the highest difficulty setting, and keeping score in public.

The mechanics

What "good at the game" looks like on tape

The fill record shows real craft, not button-mashing. 34,106 fills over the last 5.6 days collapse to just sixteen round trips โ€” huge swing positions, each worked with a ladder of small limit orders, median 60 adds and 92 reduces per position, 54% of volume posted as maker (earning the spread, not paying it). Median hold 9.2 hours. One trade under five minutes in the entire window. This is a patient position trader operating at the fill-cadence of a market maker.

The distribution beneath it: 56% of trades won, the median trade made +$33, and a handful of monsters โ€” best five +$58,170, worst five โˆ’$752,382 โ€” decided everything. That's the shape of every conviction-driven book: the entries were competent, the sizing wrote both halves of the legend.

MarketOpened (CT)SideFillsAddsReducesHoldTotal notionalNet P&L
Epilogue ยท Aug 8 and on

"I'll be back" โ€” and the chain says he means it

The morning after, the wallet held $73,539 โ€” still 22ร— his original seed and about 1,470ร— the June bottom, a fact that gets lost entirely in the shadow of the peak. He was already trading again: short BTC, short HYPE, posting positioning analysis of the leaderboard whales within a day. The desk called him back, just as the pinned post promised.

Here's the honest summary of what this ten weeks was. A trader with a genuinely good macro read, real tape craft, and an intuitive feel for when to swing big took a three-figure account to seven figures โ€” through one full wipeout and one 65% drawdown โ€” and then paid the full price of the one discipline he hadn't built yet: taking chips off the table. He knows it. He wrote the lesson himself, better than any outsider could: "Instead of putting some funds earning safely, I kept all available funds in the account."

The market charged him $950K for a risk-management education he'll never forget โ€” and he walked out of the classroom with $73K, the lesson written in his own hand, and an audience that watched him earn it honestly.

Fifty dollars to a million is the resume. The same-day post-mortem is the character reference. If he really is back โ€” and the wallet is right there in his bio for anyone to check โ€” the second run starts from a far better place than the first: same eye for the trade, plus the one lesson that was missing. That's a trader worth watching.

Provenance. All account data pulled 2026-08-08 from Hyperliquid's public info API for wallet 0x6807โ€ฆ1f8D (linked from the @tudorcross X bio): full account-value history (portfolio endpoint, 100 snapshots), complete bridge/transfer ledger (30 entries), funding history, and every fill the API retains โ€” 34,106 fills covering the final 5.57 days, the window all trade-mechanics figures come from. Equity reconciles against reported P&L to within $1,366 over the life of the account; the ledger shows $0 deposited and $0 withdrawn. Quoted posts are from x.com/tudorcross (pinned post of Aug 7 18:21 UTC, and posts of Aug 6โ€“9), retrieved Aug 8โ€“9; ellipses mark trimmed passages, nothing is paraphrased. Chart axes and timestamps are US Central; where the narrative describes his experience, times are converted to Bali (UTC+8, per his profile location), 13 hours ahead of Central. Daily-bar percentages derive from day-close snapshots, so the +690% attributed to Jul 8 covers the Jul 1โ†’8 gap in early sparse data; intraday history confirms the jump occurred on Jul 8 itself. Lifetime venue costs โ€” ~$86K fees (measured 1.14 bps effective ร— $759M volume) and $26,185 funding โ€” are stated for completeness, not judgment.