The Weather Map

For most of 2022, a strange and unwelcome thing kept happening to millions of retirement accounts. Stocks fell. That was expected—the Federal Reserve was raising interest rates at the fastest pace in four decades, and stocks don't like that. But bonds fell too. At the same time. By a lot.

This shouldn't have happened, or at least not according to the playbook every financial advisor had been using since the early 2000s. The classic 60/40 portfolio—60% stocks, 40% bonds—was built on a simple idea: when stocks fall, bonds usually rise, because investors flee to safety. The two assets were supposed to move in opposite directions, cushioning each other. Diversification, in a word.

In 2022, that relationship broke. A standard 60/40 blend lost somewhere around 16-17% for the year—by some measures its worst calendar year since the 1930s. Investors who thought they were diversified discovered, in the most expensive way possible, that they weren't.

What actually happened is that the correlation between stocks and bonds—reliably negative for two decades—flipped positive. That's exactly what a correlation matrix is built to catch.

What the Matrix Actually Measures

A correlation matrix answers one narrow question for every pair of assets you feed it: over some recent stretch of time, when one moved, did the other tend to move with it, against it, or independently?

The math is a coefficient between -1 and +1. A reading near +1 means the two assets have been moving together almost in lockstep. A reading near -1 means they've been moving in opposite directions—when one rises, the other tends to fall. A reading near 0 means their day-to-day movements have been essentially unrelated, which is the closest thing markets offer to true independence.

Our correlation matrix computes this across sector ETFs, major indices, commodities, and crypto, using trailing daily returns rather than raw prices. That distinction matters more than it sounds. Two assets can both be rising steadily and still have almost nothing to do with each other—what correlation captures is whether their day-to-day wiggles rhyme, not whether they share a long-term direction.

There's a subtler wrinkle, too: not every asset trades on the same calendar. Crypto markets never close. Stock and sector ETFs only trade on business days. Comparing the two properly means lining up the actual calendar dates they share, not just the last N data points of each—otherwise you're accidentally comparing a Tuesday to a Saturday and calling it a relationship.

Correlations Go to One

There's an old trading-desk saying that sounds cynical until you've lived through it: in a crisis, correlations go to one. It means that in calm markets, different assets each dance to their own music—energy stocks respond to oil supply, tech stocks respond to interest rate expectations, gold responds to real yields. But when real fear arrives, all of that nuance evaporates. Everything gets sold at once, for the same reason: somebody, somewhere, needs cash.

This is the mechanism behind March 2020, when stocks, corporate bonds, and even gold—historically a safe haven—all fell together in the same violent week. It's the same mechanism behind 2022, just slower and less dramatic: persistent inflation and an aggressive Fed created one dominant risk factor—the path of interest rates—that pushed nearly everything down at once, regardless of an asset's usual behavior.

Daniel Kahneman and Amos Tversky's research on anchoring found something durable about how people handle new information: once a belief is fixed, people adjust away from it far too slowly, even when the evidence in front of them has changed. An investor who learned in 2010 that “bonds go up when stocks go down” anchored on that rule and carried it into 2022 as an assumption rather than a hypothesis. The market had already moved on. The rule hadn't been checked in years.

A correlation matrix is, in a sense, a discipline against that kind of drift. It doesn't ask what used to be true. It shows you what has actually been happening lately, whether or not it matches the story you've been telling yourself.

Where the Strength Is Rotating

Correlation tells you what moves together. It doesn't tell you what's winning. For that, you need the other half of the picture: relative strength.

Relative strength doesn't ask whether an asset is up or down in isolation. It asks how an asset's trend compares to everything else in its universe—other sectors, other indices, other asset classes. An energy sector up 8% sounds good until you notice every other sector is up 15%. It's not strong. It's lagging.

This matters because markets rotate. Capital doesn't sit still—it moves from expensive, exhausted trends into cheaper, accelerating ones. The dot-com bust is the clearest example: starting in 2000, money rotated out of richly priced growth stocks and into unglamorous value names, and over the following several years that same shift extended into commodities as demand from a rapidly industrializing China took hold. The 2020-2021 stretch ran the same rotation in fast motion and in both directions—a sharp swing back into value during the vaccine-driven reopening trade in early 2021, then back into a handful of mega-cap growth names for the rest of that run. None of these shifts announced themselves with a headline. Each showed up first as a change in relative strength, before it was obvious in anyone's portfolio statement.

Our relative strength leaderboard ranks sectors, indices, commodities, and crypto against each other by trend versus their own 200-day moving average—not against individual stocks, but against the broader field of macro assets. It's the same underlying idea as sector rotation, just applied across the whole cross-asset landscape rather than one corner of the equity market.

What This Doesn't Tell You

A correlation reading is backward-looking by construction—it describes a trailing window of days that have already happened, not a law of physics. The 60-day correlation between two sectors can be meaningfully different from their 6-month or 3-year correlation, and neither one is more “correct” than the other. They're describing different regimes.

This is also why a correlation matrix should be read as a weather map, not a forecast. A weather map doesn't tell you it will rain next month. It tells you what the atmosphere is doing right now, so you can dress appropriately today. A correlation matrix doesn't predict that stocks and bonds will keep moving together—it tells you that, lately, they have been, which is useful information for anyone assuming their portfolio is diversified when it might not currently be.

The same caution applies to relative strength. A sector or asset at the top of the leaderboard has been strong recently. It hasn't been certified to stay that way. Momentum is a real, well-documented phenomenon in markets, but it reverses without warning, and the leaderboard will only show you the reversal after it's begun—same as every other trend-following tool, including the 200-day moving average this entire site is built around.

How to Use It

Before you assume a position is diversified, check whether it actually is. If you're holding two assets specifically because you expect them to offset each other, look at their recent correlation. If it has drifted toward +1, the offset you're counting on may no longer be there—not because your logic was wrong, but because the regime that made it true has moved on.

When you're deciding where to look for opportunity, start with the leaderboard rather than a single chart. A sector sitting at the top, well above its 200-day MA and ahead of its peers, is where capital has already been flowing. That doesn't make it a buy on its own, but it tells you where the market's attention currently is—information worth having before you dig into any single name.

And when nearly everything in the matrix turns the same shade of red or green at once, pay attention. That's not noise. That's the signature of a single dominant factor—a rate shock, a liquidity crunch, a risk-on surge—overwhelming the usual differences between assets. It's the market telling you, in the clearest way it knows how, that this is not a normal week.

The Deeper Lesson

Diversification was never a fixed property of a portfolio. It was always a bet on a relationship holding—stocks and bonds, growth and value, dollars and gold. Those relationships are real, and they hold for long stretches. They are not permanent.

The investors who were hurt worst in 2022 weren't wrong to build a 60/40 portfolio. They were wrong to stop checking whether the assumption underneath it was still true. The correlation had quietly shifted months before the losses showed up in a statement, visible to anyone who knew to look.

You can't control which way the wind is blowing. But you can look at the sky before you leave the house. That, in the end, is what a cross-asset view is for—not predicting the next storm, but noticing, a little earlier than most, that the weather has already changed.