UMAC gave back a big chunk of its parabolic run — RSI cooled from 72.3 to 54.7 on a -19.24% weekly return — but it's still up 39.45% on the month and 160.99% over the trailing year, with price 64.9% above its 200-day: a sharp pullback, not a reversal. AVAV and KTOS are the opposite case — both are genuinely negative over the trailing year (-31.84% and -11.75%) and trading well below their 200-day averages, so their downtrend tags are earned, not noise. The collapse pocket grew from five names to six as KRKNF round-tripped back into it one cycle after clearing out.

Data as-of the 2026-08-21 close for the full 22-name book — no staleness this cycle. No 2026-08-21-dated scan exists for this lane; the comparison baseline below is the August 14 scan.

This cycle's headline is UMAC's giveback: after a +25.04% single-day move and +83.91% monthly print last cycle, its RSI cooled from 72.3 to 54.7 on a -19.24% weekly return. Read on its own, that would look like a reversal. Read against the trailing year, it doesn't: UMAC is still up 39.45% on the month and 160.99% over the year, with price 64.9% above its 200-day average — that combination is a sharp pullback inside an intact uptrend, and its regime tag still reads pullback rather than anything worse. RCAT is a more genuine caution case buried in the same-looking chart: its 30-day return is still a healthy +22.86%, but its trailing year is essentially flat at -1.23% with a negative alpha1y of -23.05 — RCAT hasn't actually shown real one-year strength despite the recent burst, and its downtrend regime tag is arguably the more honest read of the two windows.

AVAV and KTOS are the clearest case this cycle for taking a regime tag at face value rather than assuming a short-window read is missing something. Both are tagged downtrend with a death-cross, and unlike several other names in this book, the trailing-year numbers agree rather than argue against it: AVAV is down 31.84% over the year and trades 27.6% below its 200-day; KTOS is down 11.75% over the year and 21.6% below its 200-day. Their RSI readings (44.3 and 51.6, both cooling sharply from 65.6 and 70.9 last cycle) simply confirm what the longer window already said. DRSHF is similarly unambiguous — down 41.33% over the year, 37.9% below its 200-day, RSI at 36.4 — a name genuinely weak on every horizon, not a data-window artifact.

Against that, MSI sits at the top of the book on RSI (66.1, still uptrend-tagged) but its own trailing year is nearly flat at +3.9% with a negative alpha1y (-17.91) — strong near-term structure, unremarkable annual performance. AXON shows the opposite mismatch in the other direction: its trailing year is negative (-17.78%), but it now trades 23.9% above its 200-day and its 30-day return is +27.72% — a real recovery in progress even though the year-over-year lens still reads red. PL's thesis is explicitly rated broken in the underlying data, and its technicals agree: collapse regime, 16.2% below its 200-day, despite a startling +243.98% trailing-year print that likely reflects an early-year base effect rather than a durable trend.

The collapse pocket grew from five names to six this cycle. KRKNF, which had just been upgraded out of collapse to downtrend last scan, round-tripped straight back into collapse — a fast reversal similar to the whipsaws other books in this set have shown. DPRO, AMPX, PL, LASR, and DRSHF all held their collapse tags. LASR's own thesis is still rated intact despite the collapse tag and a -34.58% monthly loss — it's still up 75.19% over the trailing year, even as it trades 16.6% below its 200-day, so the pullback is real but the longer trend hasn't been marked broken.

• UMAC: RSI 72.3 → 54.7, the sharpest cooling in the book; 7-day flipped to -19.24% (was +30.45%); 30-day narrowed to +39.45% (was +83.91%); still +160.99% over the trailing year and 64.9% above its 200-day.
• AVAV: RSI 65.6 → 44.3; 7-day flipped to -16.86% (was +3.26%); regime held downtrend, and the trailing year (-31.84%) now confirms it rather than arguing against it.
• KTOS: RSI 70.9 → 51.6; 7-day flipped to -11.47% (was +6.27%); regime held downtrend, trailing year -11.75%.
• KRKNF: regime downgraded downtrend → collapse, one cycle after clearing out of it; RSI 46.8 (was 51.9, roughly flat); 30-day held near flat at +2.61%.
• RCAT: RSI 66.1 → 51.7; 30-day narrowed to +22.86% (was +35.07%); trailing year essentially flat at -1.23% with negative alpha.
• MSI: RSI 66.5 → 66.1, still the highest in the book; 30-day climbed to +18.83% (was +14.57%); trailing year nearly flat at +3.9%.
• AXON: RSI 58.7 → 58.3, roughly flat; 30-day climbed to +27.72% (was +13.25%); still negative on the trailing year (-17.78%) but well above its 200-day.
• ESLT: RSI 46.3 → 38.0, continuing last cycle's sharp reversal; 30-day widened to -6.48% (was +7.16%).
• EOPSF: RSI 56.1 → 59.2, now current (no longer one session stale); 7-day narrowed to +12.3% (was +9.96%); 30-day climbed to +20.92% (was -0.17%).
• LASR: RSI 39.8 → 36.5; 30-day loss widened further to -34.58% (was -22.44%); thesis still rated intact; still +75.19% over the trailing year.
• DFEN: RSI 63.8 → 38.9, one of the sharpest RSI drops in the book; 7-day widened to -18.27% (was +2.68%).

Regime labels describe chart structure, not a buy/sell call, and this cycle is a genuine mixed bag rather than a book-wide read in either direction. UMAC's giveback and AXON's recovery both look like short-window noise sitting on top of intact or improving longer trends — chasing either one's regime tag alone would give the wrong steer. AVAV, KTOS, and DRSHF are the opposite case: their downtrend and collapse tags are corroborated rather than contradicted by the trailing-year data, so there's no long-horizon argument to soften the near-term weakness there. RCAT sits in between — decent monthly numbers but a flat-to-negative trailing year with negative alpha, which argues its regime tag (downtrend) is closer to the truth than its recent burst would suggest. KRKNF's round trip back into collapse after just one cycle out is worth tracking on its own; a tag that flips that fast hasn't found a stable read yet.

Tier A — RSI 55+, confirmed structure (4): MSI (uptrend, flat 1-year), EOPSF (pullback, now current), AXON (pullback, negative 1-year but well above SMA200), UMAC (pullback, still up 160.99% on the year despite the giveback).

Tier B — Mid-book, RSI 45-55 (7): RCAT (downtrend, flat 1-year — the mismatch case), KTOS (downtrend, negative 1-year), ONDS (downtrend), DPRO (collapse), KRKNF (fresh re-collapse), AVEX (too-young), NOC (pullback, negative 1-year).

Tier C — Cooling, RSI 40-45 (5): LMT, RTX (uptrend, positive 1-year), AVAV (downtrend, negative 1-year and below SMA200 — a genuine long-horizon confirm), AMPX (collapse), PL (collapse, thesis broken).

Tier D — Structural avoid, RSI <40 (6): ITA, TDY, DFEN, ESLT, LASR (collapse but thesis intact, still +75.19% 1-year), DRSHF (weakest on every horizon).

• Tape data (price/RSI/SMA/returns/1-year, SMA200): full-scan pre-computed summaries as of the 2026-08-21 close — drone-defense.