Dylan Parker @dylanquantt

michael burry was short $8.4 billion of subprime at the peak. the most he could lose on it was under $100 million. he still nearly did not survive the trade. those two numbers are the best risk engineering of the last fifty years. the thing that almost broke him is not in them, he said all of this himself, on a stage at ucla in 2012, in the only long public account he has ever given of it. the structure first: "i put my money where my mouth was. at its peak i was short $8.4 billion worth of subprime mortgages and certain financial companies. the most we could lose was less than $100 million thanks to credit derivatives." read that ratio again. $8.4 billion of exposure against a defined maximum loss of under $100 million. he did not size up and hope. he bought a bounded downside and an unbounded upside, which is the only thing anyone should actually want from a position. and the trade still almost killed the firm. here is why, in his words: "and at first we did lose. it was a negative carry trade. investors, business partners and even employees questioned the strategy. lawsuits were threatened. our distress was reported in the press. and wall street looked to squeeze our shorts." go through that list slowly, because every item on it is a separate person who could end the position, and none of them is the market. his investors could withdraw. his partners could walk. his own employees stopped believing in it. lawyers were circling. journalists were writing that the fund was in trouble. and the banks on the other side of the trade could see his position and had an incentive to push it. the maximum loss on the instrument was under $100 million. the maximum loss on the structure around the instrument was everything. then the human cost, in one of the flattest sentences i have ever heard a billionaire use: "besides all this i had to stomach what i knew was coming, a tragic end to the follies. this was not fun. i did not tap dance to work." he was right, by the way. he says it earlier in the same speech: "the financial meltdown was both predictable and preventable. this was no black swan." and afterwards, having been right: "all six of my defunct funds were audited. the congressional financial crisis inquiry commission demanded all my emails and lists of people with whom i had conversed going back to 2003. and a little later the fbi showed up. a million in legal and accounting costs and thousands of hours of time wasted, all because i asked questions." so here is the thing to take, and it is not about subprime. you can engineer the asymmetry of a trade. defined risk, open upside. that part is solvable, and most people never bother to solve it. you cannot engineer the asymmetry of being doubted. the people around a position have options you do not have. they can leave, you have to hold. the size you can actually carry is set by the weakest person attached to your capital, not by your account balance and not by your conviction. before you size anything, ask who can force you out of it. a partner. a drawdown limit at a prop firm. a spouse. a monthly withdrawal. a deadline you invented yourself. that list is your real stop loss, and unlike a price stop you did not choose where it sits. bookmark this and comment "SCION" and i'll send you the strategy i actually run, plus the one page i fill in before any position that needs time: who can end it besides the market, how long each of them will wait, and the size that fits the shortest one.

Shevaxgod @shevaxgod

a month ago

x.com/i/article/2098…

9 7 113 834K 187

@shevaxgod worth being precise: he also said the long term value sits with the model builders, not the chip suppliers. so he's not even claiming the house captures the value, only the spending.

in november 2000 jim chanos started shorting enron. his entire edge was a document that anyone on earth could download for free. he wrote down exactly how he did it, and the place he wrote it is itself a document almost nobody has read, enron was near $90 and was one of the most admired companies in america. chanos had no source inside it. he had no leak. here is his own account, from a written statement he filed with the sec. how he even heard about it: a friend mentioned an article in the texas regional edition of the wall street journal, by a reporter named jonathan weil, about accounting practices at energy trading firms. so a regional edition. of a newspaper. that everybody already subscribed to. then the actual work: "the first enron document my firm analyzed was its 1999 form 10-k filing, which it had filed with the sec." and what was sitting in it: "enron's return on capital, a widely used measure of profitability, was a paltry 7 percent before taxes. that is, for every dollar in outside capital that enron employed, it earned about seven cents." seven cents on the dollar, in a company the market was pricing as a technology business. that number was not hidden. it was arithmetic on published figures. then the line that i think is the whole post: "we read the footnotes in enron's financial statements about these transactions over and over again." over and over again. not "we found a document nobody had". not "we had a contact". they read the footnotes repeatedly until the structure made sense, and the thing that made the structure make sense was that it did not make sense. they shorted in november 2000. enron filed for bankruptcy on 2 december 2001, the largest corporate bankruptcy in american history at the time. so here is the uncomfortable question for everyone reading this on a timeline. you spend your day consuming analysis of things. threads about a company. videos about a protocol. someone's breakdown of a chart. all of it is a secondhand account of a primary document that is sitting in public, that is free, and that you have never opened. the asymmetry is not between people with information and people without it. almost everything is public. the asymmetry is between people who read the source and people who read the summary of the source. and the second half matters as much as the first. chanos did not read it once. nobody finds anything on the first read, because on the first read you are still learning the vocabulary. pick one thing you already have a position in. find its actual primary document, the filing, the whitepaper, the audited statement, the exchange rulebook, whatever it is in your market. read the part everyone skips. then read that part again. the edge is almost never access. it is attention, applied twice. bookmark this and comment "10K" and i'll send you the strategy i actually run, plus the checklist i start every asset with: which primary documents i pull, in what order, and the five things i look for in the footnotes before i look at a single chart.

Shevaxgod @shevaxgod

a month ago

x.com/i/article/2098…

9 7 113 834K 187

@shevaxgod that is why i built the post around it. not a leak, not a contact. repetition on a public document until the structure stopped making sense.

@shevaxgod the government broke up his company and handed him a bigger fortune. that should be in every finance course.

@Ruzveltxbt The scary part is how normal he looks here

in december 1983 richard dennis took 23 people with no trading background, taught them for two weeks, and gave them his own money. over the next five years they made about $175 million. the rules have been public for decades. the reason almost nobody can run them is said out loud in the first forty seconds of this interview, jerry parker was one of those 23. he later founded chesapeake capital. in 2019 he sat down with real vision, and before anything else he told a story he heard paul tudor jones give at uva in 1988. jones had just made a fortune short japan. everyone wanted the trade. this is what parker says he told the room: "what people don't want to hear is that i was short 5 times before that and took small losses each time." that sentence is the entire distance between a published rule and a realised return. the trade that pays for your year is almost never the first attempt. it is the sixth. and the five before it are small, boring, and every one of them looks like proof the idea does not work. every one of them hands you a reason to adjust the rule slightly. then parker says the second half, which is harder: "we risk a small amount, but if the volatility gets higher and higher as the trade becomes very, very profitable, we don't have a tendency to get out of any of it. we just sort of let it go." sit with what that asks of a person. the position is now big. it swings more in a day than you used to risk in a month. every instinct says take something off. and the system says do nothing. so one strategy makes two opposite demands. be calm about being wrong five times in a row, and be calm about a position that is uncomfortable precisely because it is working. dennis won his bet. he proved the rules are teachable. he did not prove the behaviour is, and parker is blunt about that: "you can't be a marine by reading a book." here is the version you can run this week, and it costs nothing. go back through your last twenty trades and count how many were small losses taken exactly as planned. not stopped out late. not widened. not talked out of. planned. if that number is low, you do not have a problem with your losers. you have a sample problem. you have never given any edge enough attempts to show up, which means you have no idea whether you have one. a system is not what you know. it is what you are willing to repeat while it is not working. bookmark this and comment "TURTLE" and i'll send you the strategy i actually trade, plus the sheet i count attempts on: how many planned losses a setup is allowed before i am permitted to change anything, and the rule that stops me trimming a winner the moment it starts moving loudly.

Shevaxgod @shevaxgod

a month ago

x.com/i/article/2098…

9 7 113 834K 187

@codewhatyouwant real deep backtest its about a 100+ trades

@recketino sometimes much more than 6

@Vladic_ETH if that $8/$30/$100/$200 list is real the multipliers say a lot. that's $8 per unit at Premium, $7.50 at Plus, about $6.70 at Super and $4 at Ultra. top tier is half price per unit, classic way to pull heavy users all the way up

@0xbobaaa the number i'd want next to 45% is how often the real humans in the same test got called human. without that baseline you can't tell if 45% is a pass or just people guessing on a video call

@shevaxgod parker opens the interview with it. jones had just made a fortune short japan and the first thing he told the room at uva was that he had been short five times before and taken small losses each time.

@shevaxgod "every position you own sits on top of a rule" - that's the line.

on 14 september 2011 a ubs trader emailed his own manager and confessed. not one risk system had found him. the bill was £1.4 billion. the regulator later fined the bank £29.7 million, and the reason they gave is not the one you are expecting, kweku adoboli was 31, a director on the etf desk in london, global synthetic equities. he was arrested the next morning. convicted 20 november 2012, seven years. here is what the fsa actually wrote when it fined ubs: "ubs failed to question the increasing revenue of the desk and failed to ensure that there was a corresponding increase in the controls in place over the desk." read that twice. the finding is not that they missed a loss. the finding is that nobody asked about a profit. they also found the desk "breached the risk limits set for their desk without being disciplined for doing so", and that ubs failed to investigate "the underlying reasons for the substantial increase in profitability." so the pnl was going up and no one asked why. that is the whole mechanism, and it is not a swiss problem. a losing trader gets questions. a winning trader gets left alone. which means supervision switches off at exactly the moment the position is becoming large enough to end you. in 2016, out of prison, he gave his first broadcast interview to the bbc. they asked him whether anything had changed: "the young people i've spoken to, former colleagues i've spoken to, are still struggling with the same issues, the same conflicts, the same pressures to achieve no matter what. and we know where the conflict comes, is where people fall into this grey zone. so i think it can absolutely happen again." "no matter what" is the whole sentence. everything else in that quote is decoration. now the part that is about you, because you do not work at ubs and you still have this exact structure. you set a number you have to make this month. you think of it as a goal. it is not a goal, it is a risk parameter, and it is the thing that will decide what you do on the day your process says stand down. and you have the same blind spot the bank had. you review your trades after losing weeks. after winning weeks you do not review anything, you just feel competent. so every audit you have ever run on yourself was run on the wrong sample. an unexplained profit is not a win. it is a position you have not measured yet. bookmark this and comment "AUDIT" and i'll send you the strategy i actually run, plus the review i do after a green month: the four questions i ask about every winning trade, and the one number that tells me whether the month was skill or size.

Shevaxgod @shevaxgod

a month ago

x.com/i/article/2098…

9 7 113 834K 187

@limalemonnn underrated part is the "stay out of the way". BYD went through some brutal drawdowns after Berkshire bought in and most people would've pulled the money at least once. picking the right person is the easy half

@arjunquants word for word from the notice: the desk breached the risk limits set for their desk without being disciplined for doing so. it is not a swiss problem, it is a human one.

@shevaxgod you check size, not outcome. did the winners come from the same setup or from bigger positions, and what was the worst drawdown inside the winning trades. bookmark this and comment AUDIT