Universe Trends

In early 2023 I was running a book that was concentrated in quality-tech, names I'd researched deeply, high-conviction, well-sized. The book was underperforming the S&P by 4 points through February despite every name being fundamentally on-track. I pulled up the Universe Trends card and the problem was immediate and external. On the 3-month window, my sector exposure (tech + communication services) was showing -1.2 on relative strength vs. the S&P; energy was +2.1, industrials +1.5, financials +0.8. On the factor strip, quality was -0.3, momentum was +0.9, and my book's implicit factor tilt was quality-and-not-momentum. I'd built a great book in a rotation that was moving away from me. The names weren't the problem, the style was out of phase. I made two adjustments: trimmed the three largest quality-tech positions by 30% and rotated the proceeds into industrial-cyclical names that had been showing up at the top of the trends leaderboard for 8 consecutive weeks. Over the next four months the industrial allocation contributed +7 points while the residual quality-tech contributed +2. Most of the recovery came from buying what was already working rather than doubling down on what my research said *should* work. That experience changed how I sequence my week: I check Universe Trends before I read any individual ticker research. Getting the sector and factor direction right is usually worth more than getting the name right within a bad direction.

This post is about the Universe Trends card, why rotation-level analysis beats name-level analysis for allocation decisions, and the three patterns that have been the most actionable signals.

TL;DR

• Sector and factor dispersion usually exceeds within-sector dispersion. Getting the sector right matters more than stock-picking within it.
• Leadership change on 3-month windows is often late-stage continuation; historically poor risk-reward to chase.
• Breadth divergence from price leadership warns that a rally is narrow and reversion-prone.
• Style rotation leads regime rotation by 2-4 weeks: often tradeable before the macro narrative catches up.
• Use weekly; don't over-trade. Rotation signals decay slowly.

Why rotation matters more than stock-picking

Over 10-week rolling windows, the median sector return can diverge from the worst-sector return by 12-18 percentage points. That spread is larger than the spread between a good stock and an average stock within a sector over the same period. The implication: if you're in the wrong sector, you're starting with a 10-15-point handicap before stock selection matters. Great stock-picking in a losing sector rarely makes up the difference.

What the trends view gives you that name-level analysis doesn't:

• Where capital is flowing: which sectors are accumulating bid, not just which are up
• Breadth: is the rally broad-based or driven by 3 names?
• Factor leadership: is the working style momentum, value, quality, low-vol, or a rotation between them?
• Geographic leadership: is US leadership being ceded to international or EM?
• Rotation timing: is the leader on 1-month also the leader on 3-month, or is the rotation already mid-cycle?

What the Universe Trends card shows

The Universe Trends card plots rolling relative-strength indices across:

• 11 GICS sectors: on a common scale showing outperformance vs. SPY, not absolute returns
• Major styles: growth, value, quality, low-vol, momentum, size (all six)
• Capitalization: large-cap, mid-cap, small-cap
• Geography: US, international developed, emerging markets
• Sub-sector detail: semis within tech, airlines within industrials, banks within financials

Configurable windows:

• 1m / 3m / 6m / 1y / YTD: swap timeframes to see leadership persistence
• Relative vs. SPY or relative vs. equal-weighted universe
• Log scale toggle for multi-quarter views

Breadth panel:

• % of names making new 52w highs per sector
• % of names above their 50d MA per sector
• New-high / new-low ratio by sub-sector
• Advance-decline line per sector

Three patterns worth acting on

Leadership change on the 3-month window signals late-stage rotation. When a sector that led on 1m underperforms on 3m, the rotation is older than it looks and you're late to the trade. Historically, chasing a sector that has led for 1 month but is fading on the 3m has negative expected return over the following 2 months, distribution is happening. The useful signal is the opposite: a sector leading on 3m that starts leading on 1m is confirming, not reversing, and continuation trades in that sector have historically worked over the following 8-12 weeks.

Breadth divergence from price leadership warns of narrow rallies. A sector making new highs on the cap-weighted index while fewer than 40% of its names are above their 50d MA is telling you the rally is concentrated in 2-5 stocks. Historically these narrow rallies resolve with the index mean-reverting rather than breadth expanding. When tech is up 8% but breadth is 35%, you're in a 3-name rally; fading the tape or rotating to broader sectors usually pays off within 6-10 weeks. Broad rallies (> 65% breadth) persist; narrow ones tend not to.

Style rotation leads regime rotation by 2-4 weeks. Growth-to-value rotations have historically preceded macro regime shifts by roughly a month. The market's sub-surface often reprices the trade before the narrative catches up. Watching the factor strip and seeing value moving from -0.5 to +0.3 on the 1-month window, while growth moves from +0.4 to -0.2, is a rotation signal you can act on before the "value is back" articles show up on financial media. The directional call is usually easier than the exact timing; being right on direction and approximately right on timing is enough.

Example: the Q1 2023 rotation

My Universe Trends snapshot from late February 2023, when I made the sector rotation:

The picture was: tech-and-quality fading, energy-industrials-momentum leading on both 1m and 3m, with industrial breadth at 68% (broad rally, not narrow). Rotating 30% of my quality-tech exposure into industrial-cyclical names at the tail end of February turned out to be well-timed. The continuation trade worked for another 4 months before breadth started rolling over. I didn't catch the exact top, I trimmed 70% of the industrial exposure in late June when breadth fell to 52% and 1m RS flipped negative. Over the full arc, the rotation contributed +7 points and the retained quality-tech contributed +2, roughly 3/4 of the recovery came from moving the book to where the wind was blowing rather than hoping the wind would change.

What the trends card can miss

• Single-day shocks. A 4% move in a sector on one day can distort RS readings; eyeball the chart, don't just read the number.
• Sector composition changes. GICS reclassifications occasionally shuffle names between sectors; RS is computed on the current classification, so historical sector RS can be slightly misleading at reclassification boundaries.
• Size vs. sector effects. Small-cap energy can behave very differently from large-cap energy; check the size cross-filter for important sectors.
• Correlation with macro. Energy RS is highly correlated with oil prices; the trend card shows the effect, not the cause. Pair with macro data for interpretation.
• Factor definition drift. "Quality" as Russell defines it differs from MSCI's definition; the card uses one framework throughout, but cross-source comparisons need attention.

Common mistakes

• Over-trading on 1-month-only signals. 1m RS is noisy; use 3m as the filter and 1m as the timing.
• Ignoring breadth. A sector up 6% with 30% breadth is fragile; same sector up 6% with 70% breadth is durable.
• Chasing late-stage leadership. When everyone is talking about the leading sector, the rotation is often already mid-cycle.
• Not pairing with regime. Sector trends in a trending regime behave very differently than in a correction regime.
• Letting stock-picking override allocation. Great names in weak sectors usually underperform average names in strong ones.

Where it fits

Pair Universe Trends with Universe Browser for name-level drill-down once you've identified the leading sector, Universe Heatmap for a visual snapshot of where winners and losers cluster, Universe Analytics for the cross-section correlation view, and the Regime Overview for the macro context around any rotation.

FAQ

How often does the card update?

Relative-strength series update end-of-day; breadth computes intraday.

Can I customize the benchmark?

Yes, swap SPY for any named index or a custom basket.

What smoothing is applied?

RS lines are lightly smoothed (3-day EMA) to reduce day-to-day noise; raw data is available as a toggle.

Does it handle sub-sectors well?

Yes, GICS level-3 (industry group) is available with the same RS and breadth treatment.

How should I think about the 1m vs 3m comparison?

3m is the context (is the leader persistent?); 1m is the timing (is the move still in motion?). Use both.

Related reading

• Universe Portfolios, Model Compare
• Stock Universe Explained
• Sector Rotation Cycle

Open the Universe Trends card → /app/universe