Three names today, from a market that is narrowing. The index-level backdrop still reads constructive, but every one of the five strongest groups is decelerating — gaining ground more slowly each week — and only ten of thirty-nine groups are advancing at all. Both of those are cautionary, so the shortlist is capped at three rather than the seven a clean tape would allow, and every name had to carry an independent confirming signal to qualify. Read the caveat on each card before the name. All three gave back their morning gains while this page was being built, and the record below is honest about what this desk's picks have actually done: forty-four of them, 4.32% behind the index.
The market-level reading is constructive — volatility is low and stable, and two of three major indices sit above their medium-term averages. On that alone the desk would be allowed a fuller list. Two separate checks cut it to three. First, all five of the strongest groups are decelerating: still ahead over the week, but adding ground more slowly each time they are measured, which is the signature of a move in its later stages rather than its early ones. Second, only ten of thirty-nine groups are advancing at all — a quarter of the market is doing the work.
Because the market reading sits in the middle band rather than the top one, each name also had to carry an independent confirming signal — it could not qualify on the screen alone. Two of the three available confirmation channels were empty across every candidate today, so that requirement was met entirely by fresh trend-change signals. One otherwise-strong name was dropped purely because it could not be checked.
Leadership is semiconductors (+9.17 momentum, +7.43% on the week), then cloud, software, commodities and AI. Every one of them is decelerating, several sharply — cloud shedding acceleration at −36.4 and semis at −27.0. Energy has collapsed to +0.54 momentum and 10th of 39, negative on relative strength, after weeks as this desk's dominant exposure. Below the top ten the market is uniformly negative: utilities −7.76, solar −7.90.
The tension worth naming: the rotation screen surfaced eleven names today and nine are Healthcare, a group that is negative on momentum. So the screen points at defensives while the tape's leadership sits in technology and fades. That is an internally consistent late-cycle picture, but it means two of today's three names sit outside the leading groups — a caveat carried on their cards rather than smoothed over.
The only name where every leg agrees. Highest relative strength of any candidate at 50.85 and rising (+5.45 on the week), the sole name on both source screens, and the only one with a trend-change buy signal dated today rather than last week — now trading above the price that signal fired at. Fresh analyst initiation this morning at a $318 target, and the only finalist outperforming the index intraday (−0.13% against −0.44%).
Strongest catalyst, weakest screen support. A Barclays upgrade to Overweight with a $515 target, ahead of October Medicare Star ratings that could move several large contracts into bonus status — a dated, identifiable event rather than a vague story. Relative strength 29.4 with a trend-change buy signal dated today, above its signal price. But it appears on only one of the two screens, its relative strength is barely rising (+1.45), and it has round-tripped a +4.6% morning gain to −0.57% after touching 408.19 — the market gave the upgrade back within hours, which deserves more weight than the headline.
Read the caveat first: short-term stretched. Its oscillator block is net negative — stochastic 93.8, Williams %R −1.9, CCI 159.5 — even though RSI is only 64 and all twelve moving averages read Buy. It is also down 1.65% today against the index's 0.44%, the worst intraday of the three. What earns it a place is being the only finalist inside the leading part of the market — technology — with relative strength 31.5 rising +7.49, the best composite score available, and a trend-change buy signal from 09-22 that it still trades above. It is deliberately kept to avoid a three-name all-Healthcare cohort, which is the concentration mistake this ledger has already paid for once.
Both income screens were rebuilt by this run — 983 cash-secured puts across 173 names and 143 bull put spreads — and none of the three finalists appears in either. That is expected rather than broken: those screens want large, high-volatility index constituents, while this funnel produces mid-cap trend names. Two systems sharing a page.
| Cohort | Mean alpha | Median | Beat SPY | Best | Worst |
|---|---|---|---|---|---|
| 2026-08-31 · 7 | −1.52% | −1.17% | 3/7 | CVI +21.8% | GEN −32.4% |
| 2026-09-01 · 7 | −3.89% | −1.44% | 3/7 | CRWD +9.2% | PRGS −16.6% |
| 2026-09-02 · 3 | −13.81% | −12.78% | 0/3 | VET −10.9% | HP −17.5% |
| 2026-09-03 · 3 | −4.93% | −4.88% | 0/3 | PFE −1.0% | ZYME −10.6% |
| 2026-09-08 · 3 | −7.87% | −5.86% | 0/3 | ELPC −3.4% | MTDR −14.2% |
| 2026-09-11 · 3 | −7.96% | −9.46% | 0/3 | TEN −0.0% | DK −13.9% |
| 2026-09-15 · 3 | −10.18% | −7.90% | 0/3 | CVE −5.6% | SM −14.6% |
| 2026-09-16 · 2 | −6.29% | −6.29% | 0/2 | MPC −4.0% | LPG −5.3% |
| 2026-09-17 · 3 | −3.41% | −4.25% | 1/3 | IOVA +0.8% | SMTC −5.8% |
| 2026-09-18 · 2 | −1.73% | −1.73% | 1/2 | ABCL +2.5% | CMBT −4.7% |
| 2026-09-21 · 3 | +4.30% | +3.84% | 2/3 | ALAB +9.3% | RBRK −2.1% |
| 2026-09-23 · 3 | −1.72% | −3.81% | 1/3 | MRNA +3.9% | RNG −5.6% |
| 2026-09-25 · 2 | +0.36% | +0.36% | 1/2 | SN +0.6% | LCUT −1.2% |
| All 44 | −4.32% | −4.57% | 12/44 | trimmed −4.30% · ex top two −5.29% | |
| 2026-09-28 · 3 | — | — | — | flagged today · first measurable reading next session | |
Across forty-four picks the desk is 4.32% behind the index on the mean, 4.57% on the median, with 12 of 44 beating it and −189.9 points of cumulative alpha. Removing the two best names makes it worse, at −5.29%, and the trimmed mean sits at −4.30%, essentially on top of the raw mean. This is not a good book dragged down by outliers; it is a distribution sitting below its benchmark.
The dominant cause is identifiable and it is not the gates: 21 of 44 picks were Energy, four cohorts entirely so, through a period when that group fell apart. Today's list was built with that failure explicitly in mind — which is why a stretched technology name was kept over a cleaner Healthcare one. Two of three are still Healthcare, so the concentration risk is reduced, not solved.