Perfect math, wrong problem.

I'm a bit of a maverick when it comes to investing, and honestly when it comes to the way I think in general. So here's something I believe that gets me funny looks: just because someone has a Nobel Prize and their math is flawless and beautiful, that doesn't mean what they wrote is actually useful.

You can have a perfect formula where one plus one always equals two, everything adds up on paper, and it still doesn't comport to reality, because you're describing the wrong things in the wrong context.

Take Harry Markowitz. His 1952 paper, "Portfolio Selection," is the foundation of modern portfolio theory, and it eventually won him a Nobel Prize (well, the economics version of one, in 1990). But read the opening paragraphs. He says right up front that picking a portfolio has two stages. Stage one is forming your beliefs about how every security is going to perform in the future. Stage two is turning those beliefs into a portfolio. And then he tells you his paper is only about stage two.

So the entire theory sits on top of inputs it doesn't provide. Expected returns, variances, how everything moves together. You supply those. If your guesses are right, his math will build you a lovely portfolio. If your guesses are wrong, his math will build you a lovely portfolio out of wrong guesses, and it'll look just as elegant on paper. Seventy years later people are still arguing about which assumptions to feed the machine, and that argument is the whole ballgame. The math was never the hard part.

Here's how I think they actually work. Public markets run on liquidity. How many shares are available for sale, and how many are demanded for purchase, at a given price at a given time. The value of the company is in there somewhere, sure. But how do you even value a company? There are different methodologies, different standards, different estimates, and every one of them still has to pass through supply and demand before it becomes a price.

We just got a live demonstration. In July 2026, a hedge fund run by a 20-something former OpenAI researcher, a fund that had reportedly returned more than 1,000% since it launched in 2024 and was sitting on something like $45 billion, got caught levered about 4 to 1 in AI infrastructure stocks. The Nasdaq 100 fell roughly 10% that month. His core holdings fell 27% to 54%. At 4x leverage, a 30% drop in the book is about a 120% hit to your equity, and the prime brokers (Goldman, JPMorgan, Bank of America) started making margin calls. On July 30 he sold something like $20 billion of stock to Citadel in a single block before the market opened.

Into the forced sale Day the forced seller was gone

CoreWeave

Nebius

SanDisk

SK Hynix

Micron

−60 −40 −20 0 +20 +40

Same companies. Same revenue. Same CEO. What changed was who had to sell.

Then the interesting part. The day the forced seller disappeared, those same stocks jumped 20% to 27%. CoreWeave, Nebius, SanDisk. Same companies as the day before. Same revenue. Same CEO. What changed was who had to sell.

That is a liquidity story. Not a valuation story, not a modern portfolio theory story, not a my-formula-is-perfect story, not an I-have-a-PhD story. Somebody had to dump a lot of shares, and there weren't enough buyers on the other side to hold the price up.

It happens on a smaller scale every December. Lots of retirees with lots of money in IRAs get a stack of notices from their custodian: take your required minimum distribution by December 31 or eat a penalty. So they call in and say, "I need to pull money out. Sell." Does that selling, concentrated into a few weeks because of a tax calendar, have anything to do with the value of the underlying companies, or GDP growth, or anything like that? No. It's human behavior on a deadline.

So you've got fear and greed, you've got supply and demand, and then, oh yeah, there's the various ways of estimating what a company is worth. And that last piece is the one everyone harps on. The efficient market crowd says information travels fast and everyone can react to it instantly, so prices are always about right, and that's why we punish insider trading, to keep the market efficient.

The pieces are true. Prices move fast, data routes instantly, access is broad. But a lot of the people making decisions are making them on very little of that information. Plenty of retail traders are buying because "line go up." They're already late to the party. There's the old joke that when your cab driver starts giving you stock tips, it's time to sell. It's a joke, but there's some reality in it. Retail demand shows up, institutions unload into it, and they do it just gently enough to keep the line going up until they've exited what they wanted to exit.

So the day-to-day game is volume. How much can you move at a certain price. That's the day trader game, the market-maker game, the algorithms splitting orders into pieces so they don't move the price against themselves while they're building or unwinding a position. Not everyone behaves that way. Not everyone uses all available information. Not everyone prices things the same way. And even though one plus one equals two, that doesn't tell you whether you're better off adopting a cat or a dog.

Long-Term Capital Management, back in 1998, nearly took the global financial system down with it. They were running about 25 to 1 on the balance sheet, with over a trillion dollars of derivatives notional stacked on top, and when their trades went the wrong way the New York Fed had to herd 14 banks into a $3.6 billion rescue. Two of the partners were Myron Scholes and Robert Merton. They'd won the Nobel the year before, for the option pricing model with Scholes' name on it. Brilliant idea, egregiously misapplied, with a big dose of hubris, and it cost billions.

So markets go up and down, sometimes violently, sometimes for no reason that has any rational or mathematical basis beyond liquidity and behavior. Occasionally a few rational players walk in, look at the pricing, decide it's off, and buy or sell to close the gap. That's the readjustment. The rest is people. And people with PhDs, Silicon Valley geniuses, even multiple Nobel laureates in one room, screw up quite often.

Which brings me to Tzur Magen

That's why I built Tzur Magen differently. I didn't want it to rest on a single theory or a single idea that has to stand on its own. I didn't want it constrained to a single market. Money moves between markets. So should your portfolio.

See what this looks like on a Tuesday →

If you want to look at a different way to manage your portfolio, let's have a conversation.