Rotation or De-Risking? Reading Breadth When Both Sides Fall - Part 2

If you haven't read yesterday's blog post yet, please do so before continuing here.

What the Band structure distinguishes

There's a structural version of this same question, and it's one readers can run on charts they already have open. Relative outperformance during a broad decline tends to sit differently within a Band structure than outperformance during accumulation. Two conditions are worth separating. In the first, both indexes occupy comparable positions within their respective ranges, and the outperformer is simply descending at a lower rate. Position within the structure is deteriorating on both sides. The numerical gap between them widens, but neither index is doing anything a Band read would describe as constructive. In the second, the outperformer holds or improves its position relative to the Intermediate Band while the laggard loses ground within its own range. The two are separating structurally rather than just numerically, and that separation tends to persist rather than reverse the moment selling pressure lifts.

The question that distinguishes them is short enough to carry around: is the outperformer holding its position in the range, or falling slower inside the same part of it? A percentage spread between two indexes tells you the gap widened. It doesn't tell you whether the leader is being accumulated or simply being sold less aggressively. Band position separates those conditions where a raw relative number can't. The same logic applies below the index level. A holding that's outperforming its sector invites the identical check, and the answer is frequently less flattering than the relative number on its own would suggest.

The semiconductor variable

There's a measurement problem sitting underneath all of this, and it's large enough to deserve its own treatment. The Philadelphia Semiconductor Index closed last week 20.2% below its record high from June 22. It fell more than 18% over the course of July. It's still up close to 65% for 2026, which gives some sense of how far and how fast the complex traveled before it turned. When a group that size reprices that quickly, mechanical effects follow. Index level breadth and index level relative strength both absorb the distortion. Cap weighted indexes carry the damage directly because the affected names sit near the top of them. Small cap indexes carry very little of it, which means some portion of last week's 2.4 point spread is arithmetic rather than allocation. None of that requires a view on where semiconductors go from here. It's a reason to be precise about what the spread is actually measuring. A single sector unwinding at that speed can produce something that resembles rotation without any corresponding demand appearing anywhere else on the tape.

What would resolve it

The distinction is observable going forward, and a handful of conditions separate the two readings more cleanly than last week's numbers do. The most useful is whether small cap relative strength persists on days the Nasdaq is higher. Outperformance during declines is ambiguous by construction. Outperformance during advances isn't, because it requires something to actually be bought.

Beyond that: whether the advance-decline line turns positive on up days rather than merely less negative on down days. Whether sector participation widens past a single energy or defensive pocket. Whether equal weight continues to lead once the semiconductor unwind stops dominating the arithmetic. There's a near term catalyst as well. More than 80 S&P 500 companies report this week, with AI spending plans drawing particular scrutiny. Events of that type tend to force the question rather than let it hang.

The discipline

The useful skill here isn't calling rotation or de-risking correctly in advance. It's recognizing which one the available evidence currently supports, and sizing to that rather than to the more comfortable interpretation. Relative strength will always find you a leader. In a declining tape, that leader may be nothing more than the thing that fell least. Breadth is what keeps a relative reading honest, and structure is what tells you whether the separation means anything at all.

This material is published by Arbitrage Trade for informational and educational purposes only. It does not constitute investment advice, an offer or solicitation to buy or sell any security, or a recommendation of any investment strategy. All references to specific indexes, sectors, or instruments are illustrative and are intended to describe observed market conditions and patterns rather than to direct any course of action.

Market conditions change. Patterns identified in historical or current data do not indicate future results, and no analytical framework identifies outcomes in advance. Readers are responsible for their own analysis and should consult a qualified professional regarding their individual circumstances before making any investment decision.