Give me prolos. Dm for the discord server

Foreigners are DUMPING South Korean stocks at an UNPRECEDENTED rate: Investors from overseas sold a record -$152 billion in Korean equities so far in 2026. This is equivalent to 3.8% of South Korea’s stock market capitalization. Over the last 8 sessions alone, foreign investors have dumped -$11.6 billion. Meanwhile, SK Hynix has recorded 15 consecutive days of foreign selling, while Samsung has seen 5 straight sessions of outflows, according to Goldman Sachs. These two names account for over 50% of the KOSPI index market cap. Investor pressure on the Korean chip-heavy market is intensifying.
8
45
155
11,826
Valeretn retweeted
Europe is fucked... Gas prices are surging. Germany is Europe's largest gas storage hub and accounts for 25% of EU gas consumption. Storage levels are at just 56%... The lowest ever for this time of year. And look at what Europe has done to itself... - Sanctioned Russia, its largest gas supplier. - Qatar is renewing force majeure notices every month. - Refrained from filling gas storage because prices were too high. First, they switched to renewables while simultaneously reducing baseload power capacity, which is needed when the sun isn't shining and the wind isn't blowing. Oh... And guess when the sun isn't shining and the wind isn't blowing? During winter, when gas consumption is at its highest. In other words, Europe is entering winter with dangerously low gas inventories, disrupted supply routes, and an energy system increasingly dependent on the weather. And now gas prices are surging again... The EU is on its path to economic suicide.
38
175
888
34,499
Every option trade has a dealer on the other side who does not want the directional risk, and hedges it away in the underlying almost immediately. That hedging is not a side detail, in periods of heavy options flow it can move the stock more than the news does. ⤵️
1
1
16
Barbon and Buraschi found this in actual options data, tying large dealer gamma imbalances to intraday momentum when gamma is negative and intraday reversal when it is positive, with the effect strongest in the least liquid names.
1
None of this predicts direction, a dealer's hedge does not know where price should go. It only shapes how big the next move gets once something starts it Public gamma-exposure numbers are estimates of dealer books, not the real thing Educational content, not investment advice.
A portfolio spread across assets with a correlation of 0.3 looks genuinely diversified. In the calm data that number is real. In a crash, the same two assets can trade closer to 0.9, and the diversification you were counting on mostly disappears right when you need it. ⤵️
1
1
19
The key detail: it tracks the market trend, not volatility on its own. Correlation rises in bear markets specifically, not simply whenever swings get bigger. A volatile rally does not carry the same tail-correlation risk as a volatile decline.
1
1
A covariance matrix estimated on calm-period data is not wrong, it is incomplete, it describes the regime you are usually in, not the one that does the most damage to a diversified book. Educational content, not investment advice.
1
A trader who tests a strategy at ten contracts and sees almost no slippage often assumes size will scale for free. Impact does not stay flat, it grows with the square root of size, robust across stocks, futures, options, even Bitcoin. ⤵️
1
1
13
Trade 25 times your test size and the naive flat-rate assumption says nothing changes. The square-root law says your impact rate is about 5 times higher, since the square root of 25 is 5, not 25 times higher, but nowhere near unchanged either.
1
1
Total dollar cost grows even faster than the rate does, since a higher rate is paid on more shares at once, roughly size to the power of 1.5 This is why a strategy's real capacity is usually smaller than a small backtest ever reveals Educational content, not investment advice.
1
Valeretn retweeted
THE BANKS THAT FUNDED THE AI BOOM ARE NOW LOOKING FOR EXIT. Some of Wall Street's biggest banks are now trying to get rid of billions of dollars in AI-related debt. JPMorgan, Morgan Stanley, SMBC, and MUFG are reportedly looking to reduce their exposure to AI data center loans. And what's happening with Oracle is even more worrying. Banks have spent over 6 months trying to sell $38 BILLION in debt tied to Oracle data centers. Some lenders are now even willing to sell parts of these loans at a discount. Meanwhile, Oracle's credit rating is just one step above junk, and the cost of insuring against its default has jumped sharply. This is happening while tech companies are still spending hundreds of billions on AI infrastructure. If banks become less willing to lend, AI companies could find it much harder and more expensive to fund new projects. And if AI spending starts slowing down, it could become a serious problem for the stock market, which has been relying heavily on the AI boom.
108
445
1,754
125,782
Five thousand simulated funds, every one with the exact same true expected return, no skill differences at all. Look only at the ones still open after 15 years and they appear to beat the ones you started with by almost a full point of annual return. ⤵️
1
2
14
Elton, Gruber and Blake (1996) found this in real mutual fund data too, survivors overstating true industry performance by about 0.9% a year. Other studies land anywhere from roughly 0.4% to 1.4% a year, the range depends on the dataset and how long funds are tracked.
1
A backtest built on today's list of tickers, funds, or strategies has the same blind spot, whatever got delisted, liquidated, or quietly shut down along the way is invisible to it by construction, not by accident. Educational content, not investment advice.
1