# Gooaye EP716: On the Day You Can't Raise Cash, the Number on the Statement Doesn't Count Source: Realpha Blog (blog.getrealpha.com) Original article and charts: https://blog.getrealpha.com/en/blog/gooaye-2026-11-21-ep716/ > Notes on Gooaye EP716 (21 November 2026): the host frames this Taiwan sell-off as a liquidity problem — OTC turnover under NT$100bn, a heavily traded stock adding twenty lots in two hours, his own drawdown past 30% — plus what happened to him in April 2025 when he could not produce a few tens of millions. Educational listening notes only: no investment advice, no price targets, and the judgements here may be wrong. Published: 2026-11-22 Locale: en Tags: Gooaye, podcast notes, liquidity, risk control, leverage TL;DR: The host puts this sell-off down to liquidity: forced sellers cannot get out, and the market needs one high-volume day to clear it. The homework he leaves listeners is risk control — a lesson most people only learn after taking the hit themselves. ![A fishing harbour at low tide, a large boat sitting tilted on the mud, mooring lines slack on the ground, a tide mark on the sea wall well above the hull, the channel narrowed to a thin stream running into the distance](/covers/gooaye-2026-11-21-ep716-cover.png) > Empty, it tilts; half full, it stands upright; full, it topples. … Ah! What is there that is full and does not topple? > > —— Xunzi, "You Zuo" (Warring States; my own translation) In Gooaye EP716, recorded on 21 November 2026, the host Hsieh Meng-kung frames this week's Taiwan sell-off as a liquidity problem. Turnover on the OTC market fell below NT$100 billion; he went for a 14km run during Friday's session, came home, and found that a stock which normally trades heavily had added twenty lots in two hours. He says his own drawdown is past 30% and may break 40% next week — and that is with cash still on hand. For his script to work, next week has to produce one high-volume day so that the margin-called and the merely frightened can get out. If it doesn't, his read is that even people running just over 1x leverage get liquidated. ## Is this the AI bubble popping? He leans against that reading. What the market is short of, in his view, is money. Chasing the succession story, the regulator's inquiry into chip revenue recognition, or politics is beside the point now, he says — those were catalysts. The signs he lists all point one way: people with cash won't step in, lenders are pulling umbrellas in the rain, and circular financing and share-backed loans are both sized off market value, so when value shrinks fast, forced selling follows. In Taiwan he notes something rare: a number of convertible bonds trading below their issue price. He also mentions that the cost of taking a convertible apart has gone from two or three percent to around eight. Money got expensive. He saw Oracle post after the close, getting ahead of any coming coverage about financing and data centre buildouts — his read is that the coverage is still to come, and that of everyone in AI, SoftBank runs the most leverage and is the most exposed right now. As for the depth of the fall, he puts it down to mean reversion: things that tripled or quintupled on the way up fall hard on the way back, and a few supply-chain missteps plus an event or two got the whole correction done at once. He and his circle had been penciling in the first half of 2027, some of them 2028. Nobody had late 2026. ## I'm not running much leverage — why did this hit me too? Because the threshold is lower than most people assume. What used to pass for a safe level, 1.6x, gets taken out in this one, he says. The vessel in the epigraph only stands upright while it is half full; fill it and it tips. His 1.6x is what it looks like to think you are still at the halfway mark. He retells April 2025 at more length than before. The reciprocal tariff headline landed, the hot names went limit-down three days running. He had leverage, but his book was spread out, so the plan was simple: if something broke he'd sell the large caps and have cash. Instead everything fell together, and when he needed cash he could not get it. He went round the banks and even leasing companies to open every credit line he could, then went back to the friends he invests in businesses with each year and asked whether they could hold off a few days, a few months, before opening the company. He calls that day his national day of shame: a net worth that was fine on paper, and he could not produce a few tens of millions. His child had just been born around then, with complications that had him going up to the intensive care floor at NTU Hospital and waiting outside for the doors to open. Family, business and portfolio all burning at once. From that he draws the distinction. A person's net worth measures what their equity is worth. When people quote Elon Musk's wealth, they are quoting his stake in SpaceX and Tesla, not cash sitting in a bank. In calm weather anything converts; in an extreme, the rich person's trapped money is the money nobody can rescue. An ordinary household with a hundred thousand stuck can go home and ask their mother. When the amount is large enough that the bank is worried about you going under, there is no next hand. I thought about my own account here. I track position sizes, and I have rarely asked how many days it would take me to produce a given sum if I needed it tomorrow. ## Parts are still short — why did the share prices break first? Because a different hand is setting the price. He describes it as fighting himself: from the industry side, memory and a range of components are still scarce. A Japanese supplier told him one input is short enough that a line could go down — they may not be able to build. That supplier's stock, he says, has fallen past recognition. The old pattern was that shortages and rising revenue kept a stock going up to some plateau, the price broke first, and revenue followed six months later. His view is that the era of flooding capital rewrote that sequence: AI amplifies and accelerates, so the way up and the way down are both fast. He won't accept "the market is always right," and he is blunt about why: the market gets things wrong often, and that is where the money is. If it were always right you would have no information gap to trade. Their industry read is a shallow downgrade. Visibility into 2027 holds up in the consensus; 2028 is where the uncertainty starts, especially until the succession question is settled — here he passes on Jeff Bezos's argument that a machine this complex is best handed to an insider who understands it deeply, and that bringing in an outsider makes success remote. Then he points at TSMC through 2022 and 2023: with hindsight earnings did come in below expectations, but not by much — a shallow dip, and the stock went on up afterwards. That shallow dip still cut the share price in half. Earnings estimates and the multiple both came down, and the product of the two is where the damage comes from. ## Is there any signal that tells you it's close? He lists a few, and says plainly that one of them failed this time. Convertibles trading below issue price is one. The pattern is rare, and when it shows up the market is usually in extreme fear; looking back, that area is often some kind of low — though he adds that how many more down days come after it, nobody knows. Another is news of settlement defaults: the old hands' rule is that once that surfaces, it's about done. This time it surfaced and the market fell for another four or five sessions, with perhaps two or three more lows to go. What he is actually waiting for is one big-volume day: the margin-called and the people who want out all get filled. Once the liquidity knot loosens, the bounce tends to be violent, with several names limit-up. He also spends a stretch on market structure, the most concrete part of the episode for me. After the stress test in July 2026, the call auction interval for flagged stocks was cut from five minutes or more to two, but he thinks that locking people up while turnover has dried out still causes problems. His suggestion runs the other way: require cash up front to buy a flagged name, the way US settlement works, rather than letting someone wire the money two days later — and leave ordinary stocks on two-day settlement. His reasoning: tightening during a melt-up is fine, but a falling market is when liquidity matters most, and pulling the lifeline then also keeps out the people willing to bet this is the bottom. ## So what can I do with this? You are probably sitting with one of two questions: whether to cut or hold what you're stuck in, or whether to buy if you still have money. He answers neither, saying strategies differ and when to buy or sell is a personal call — and noting how many people turn up at moments like this to tell others what to do. But he does leave three things you can check yourself. First, the order of operations: count the cash before you argue about cheap. He had cash, and he still says he bought index futures too early and bought the US dip too boldly; he thought hiding in the satellite names would be safe, and they came down too. Second, what a real low looks like. In all these years, he says, the bottom has never arrived while the friends around him were saying buy: they either want to buy and have no money, or they're stuck up the mountain, or they've lost conviction. None of those three states has anything to do with valuation. So when a pile of names falls under seven or eight times earnings, into places on the river chart that rarely get visited, the rational move is to go hunting — and most people go hold the steadiest thing instead. He admits he gets scared too, scared of touching the great white. Third, how the risk-control lesson actually lands. He's direct about it: you don't pick it up while the wind is behind you, you pick it up after eating a mouthful, and the next helping is always bigger than the last. Friends still tell him his own risk control isn't tight enough. He quotes Munger on where clever people end up — liquor, ladies, leverage — and says leverage is the main reason he has watched clever people blow up: they may be right, and may be proven right in the end, but with a position that can't absorb uncertainty they get forced into doing the thing they didn't want to do, and the outcome is unrecoverable. Slower is faster, he concludes, because you cut the odds of a mistake and the odds of a forced sale, which is what lets your actual judgement show up in your returns. My two favourite moments in the episode have nothing to do with the tape. One is the 14km run during Friday's session — he says anyone watching would think he'd lost it, but with liquidity frozen nothing he did would have mattered. The other is him recommending the newly released GTA, where he has taken to picking up a basketball and throwing it at pedestrians, and finds watching a scooter rider fly off hilarious — not that funny on a normal day, he says, but it releases something when the pressure is high. The tone matches the risk-control talk: there are stretches where effort and reward are disconnected, so protect the mood first, protect the family relationships first, and win the money back later. ## Worth a look - Gooaye EP716 (21 November 2026), the source for these notes - Daily turnover and margin balance statistics from the TWSE and the Taipei Exchange, for liquidity and leverage levels - The TWSE rules on disposition measures and call auction intervals, to check his description of the interval and prepayment - Secondary market quotes and issue prospectuses for convertible bonds, to understand the below-issue-price signal - TSMC's quarterly reports and share price range across 2022–2023, to test for yourself how a shallow downgrade halves a price ## The one thing to take with you **What you have and what you can produce today are two separate ledgers.** That is the step worth paying for in this episode. The number on the statement is a market price times a quantity, and the price only holds while someone is on the other side. On the day those people leave together, the multiplication breaks, and what remains is whatever you can actually produce. Most of the time the two ledgers read the same, so people only keep one. Here's something I've tried that works outside markets too: pick three resources you assume you can call on any time — not savings; the kind where "if something happens, they'll help" — someone who promised you something, a credit line you've never drawn, a position you assume you could go back to. Write one line each: if I needed it today, how many days from asking to having it. The ones where you can write a number are in your hands. The ones where you can't, you only know exist.