When the Market Changes Character

There is a moment, late in the trading day, when a market headline can make the whole market look simple. The index is green. The financial headlines are calm. A few large companies are doing enough to make the whole screen look healthy.

Then someone asks a less comfortable question: how many stocks are actually participating?

The answer is not always what the headline suggests. The index can be moving higher while the market underneath it is quietly losing support. Or the index can look damaged while more individual stocks are beginning to recover. The visible market and the lived-in market—the one made up of hundreds of separate companies—can tell two different stories at the same time.

That disagreement is where the idea of a market regime becomes useful. Not because it solves the mystery of what comes next. It does something more modest, and more difficult: it describes what kind of market is in front of us before the explanation becomes obvious.

The market beneath the market

Start with a simple count. tradingtwohundred measures the percentage of tracked S&P 500 stocks whose latest price is above its 200-day moving average. Each stock gets one vote. A company with a trillion-dollar market value and a company with a much smaller one both count as one stock.

This is what breadth measures: participation. It is different from the market-cap-weighted return of the index, which can be carried by a small group of very large companies. If you want the longer introduction, see what market breadth measures and what the 200-day moving average represents.

The counterintuitive part is that a market does not need to look dramatic to be changing. Participation can narrow by degrees. A recovery can begin while the index still feels heavy. A strong reading can become less convincing without turning negative. The important change is often not the headline number, but the relationship between the number and its recent history.

So the classifier adds two questions. How much has breadth changed over the past 20 sessions? And how much have its daily changes varied during that window? The first gives the move a direction. The second gives it a temperament: settled, or unsettled.

The interruption comes first

Most classification systems want to tell you whether something is going up or down. This one begins by asking whether the market is moving too quickly for that answer to be comfortable.

If 20-session breadth volatility is at least 4.5 points, or if the absolute 20-session breadth change is at least 14 points, the label is Transition. That check comes first. It can override what might otherwise look like a broad expansion or contraction because the speed or instability of the move is itself the most important fact.

This is the part that is easy to misunderstand. Transition does not mean that prices are about to rise or fall. It means the market's character is moving. The evidence is changing fast enough—or disagreeing enough—that a stable directional story deserves caution.

Think of it as the moment in a conversation when the subject changes. You may not know what the new subject will be. You only know that the old one no longer explains everything happening in the room.

Once the noise settles, the rules speak

If the Transition condition is not met, the classifier becomes easier to follow. Breadth supplies the level. The 20-session change supplies the direction. Together they sort the market into a small set of descriptions.

At least 60% breadth with a 20-session change of at least 0 is Broad expansion. At most 40% breadth with a 20-session change of at most 0 is Broad contraction. These labels do not make a claim about valuation, the economy, or future returns. They say that participation is either broadly established or broadly weak, and whether it is holding or moving in the same direction.

The middle of the range is more interesting. If breadth is below 50% but has improved by at least 4 points over 20 sessions, the label is Recovery. If breadth is at least 50% but has declined by at least 4 points, the label is Narrowing advance.

In one case, participation is coming off a weak base. In the other, a majority of stocks may still be above their long-term averages, but the support is becoming less widespread. The surface can look similar—an index that has not done much—while the structure underneath is changing in opposite directions.

Everything left after those checks is Range-bound. That last category is not a failure to decide. It is a decision not to force a directional story onto mixed evidence.

A label needs a setting

A single word can be useful, but a single word can also become too persuasive. That is why the Market Regimes dashboard places the current inputs beside a historical timeline. The timeline gives the reading a past. It shows whether today's label is an isolated interruption or part of a longer stretch of similar conditions.

The dashboard also shows sector participation and stock dispersion. Sector participation is the percentage of sector ETFs above their 200-day averages. Stock dispersion describes how spread out individual stocks are relative to their own long-term averages. These measures can make the setting more legible: is a broad reading shared across sectors, or are the individual outcomes unusually scattered?

But context is not the same thing as classification. The regime label comes from breadth, the 20-session breadth change, and 20-session breadth volatility. A longer 60-session breadth change may appear in the current snapshot, but it does not replace the 20-session rule used to assign the label.

That separation is a quiet kind of honesty. It tells you which facts changed the answer and which facts merely helped explain it.

Where the story stops

The thresholds are boundaries, not natural laws. A reading just above one can receive a different label from a reading just below it even when the market feels nearly identical. That is the price of a reproducible classifier: the rule is visible, and so is the edge case. The full definitions are on the data and methodology page.

The 200-day moving average is intentionally slow, so breadth can change after some price movement has already happened. A regime describes measured participation; it does not predict the next session, identify a perfect entry or exit, or establish that a historical pattern will repeat. It is evidence about the present, not a promise about the future.

There are technical boundaries as well. The historical timeline requires matching date and breadth arrays with at least 21 observations, and its labels begin once the 20-session history needed for the calculation exists. The tracked universe is not every security in the market. The underlying data can be delayed, revised, incomplete, or affected by a provider or processing error. Historical labels can change if that underlying data changes.

And none of this accounts for an individual person's goals, taxes, time horizon, or risk tolerance. A regime can help a reader ask a better question of the data. It cannot make a decision personal by itself.

The useful question

In the end, a market regime is a compression device. It takes three observations—how broad participation is, how it has changed over 20 sessions, and how steadily it has changed—and turns them into a phrase a person can carry from one screen to the next.

The phrase is not the conclusion. It is the beginning of the inspection. Read Transition with the volatility and breadth move that triggered it. Read Recovery with the weak base it is recovering from. Read Narrowing advance with the participation that is quietly disappearing. Let the timeline and the context metrics show what the label cannot.

The useful question is not, “What will the market do?” It is, “What is the evidence saying about the market's character right now?” A good description does not remove uncertainty. It gives uncertainty a shape you can see.