MP's monthly return nearly doubled to +31.46% and its RSI stayed elevated at 65.2, but its trailing year is actually negative (-12.05%) with the steepest negative alpha in the book — this looks more like a bounce than a resumed uptrend. AMKR is the clearer casualty this cycle, downgrading fresh into collapse as its RSI fell from 49.5 to 39.5 and its monthly loss deepened to -24.95%. CRS, the book's three-time regime whipsaw, cooled further but stayed above its 200-day and up over 100% on the year — a deep pullback, not the breakdown its RSI alone would suggest.

Data as-of the 2026-08-21 close for the full 16-name book — no staleness this cycle, though AMVMF again printed zero volume. No 2026-08-21-dated scan exists for this lane; the comparison baseline below is the August 14 scan.

MP is the name most likely to be misread this cycle if the trailing-year check is skipped. Its RSI held at 65.2 (down only slightly from 67.1) and its 30-day return nearly doubled to +31.46% from +18.76% — on the surface, a clean continuation of last cycle's repair out of collapse. But MP's trailing year is actually negative at -12.05%, with an alpha1y of -33.87, and it trades only 4% above its 200-day despite carrying a death-cross on its longer moving averages. That combination argues this is more likely a sharp bounce inside a name that's still lagging its own year-ago level than the resumption of a durable uptrend — the short window looks strong, but the long window disagrees.

AMKR is the clean casualty this cycle: its regime downgraded fresh into collapse, RSI fell from 49.5 to 39.5, and its monthly loss deepened sharply to -24.95% from -12.79%. Its trailing year is still positive (+116.54%), but it now trades 9.3% below its 200-day — a trend genuinely under test rather than a false short-window alarm. CRS, the name that's whipsawed between uptrend and pullback three times in three cycles, cooled further this cycle — RSI down to 35.6 from 47.4, its monthly loss widening to -16.21% — but its longer-horizon numbers are still solidly positive: up 104.03% over the trailing year and 16% above its 200-day. Read against that, CRS's rough week looks like a deep pullback inside an intact multi-year trend rather than confirmation that the whipsaw is finally resolving to the downside.

Against those two, the strength cohort held up well: ALM posted the highest RSI in the book at 66.9 with a 30-day gain of 29.53%, ELMT climbed to 65.0 with a 30.26% monthly gain, and both AMVMF (62.2) and SDVKY (62.0) extended further. TER gave back most of its prior cycle's surge — RSI cooled from 60.6 to 48.2 and its 30-day return fell from +22.41% to +1.7% — but it remains up 243.77% over the trailing year and 22.2% above its 200-day, another case where the sharp short-term give-back sits on top of an intact longer trend. WOLF's own trailing-year number (+1,780.29%) is likely distorted by a prior severe decline that set an unusually low base a year ago; the more reliable tell for this name is that it trades 9.4% below its 200-day inside a fresh collapse tag.

AMVMF again printed zero shares traded this cycle, the same caveat this name carried before — its RSI (62.2) and 30-day return (+20.55%) should be treated as unconfirmed until a real print lands.

• AMKR: regime downgraded pullback → collapse; RSI 49.5 → 39.5; 30-day loss deepened to -24.95% (was -12.79%), now trading below its 200-day.
• MP: RSI held near-flat at 65.2 (was 67.1); 30-day nearly doubled to +31.46% (was +18.76%); trailing year is negative (-12.05%) despite the strong short-term print.
• CRS: RSI 47.4 → 35.6; 30-day widened to -16.21% (was -5.63%); still +104.03% over the trailing year and above its 200-day.
• TER: RSI 60.6 → 48.2; 30-day fell sharply to +1.7% (was +22.41%) — most of last cycle's surge round-tripped, but still +243.77% over the trailing year.
• KLIC: RSI 51.3 → 39.8; 30-day loss widened to -18.8% (was -7.75%); thesis still rated intact-but-priced.
• WOLF: regime held collapse; 7-day widened to -18.97% (was -3.29%); 30-day loss deepened to -11.99% (was -9.43%).
• ALM: RSI 66.9, still the highest in the book (was 55.8); 30-day climbed to +29.53% (was +7.71%).
• ELMT: RSI 65.0 (was 67.5, roughly flat); 30-day held strong at +30.26% (was +25.55%); still tagged too-young.
• SDVKY: RSI 62.0 (was 47.8); regime upgraded to uptrend; 30-day climbed to +12.09% (was -6.11%).
• KMT: RSI 40.4 → 36.5, still one of the lowest in the book; regime now downtrend with a death-cross; 30-day widened to -13.73% (was -9.61%).

Regime labels describe chart structure, not a buy/sell call, and this cycle's most important read is separating MP's short-term bounce from a real trend resumption. MP's RSI and 30-day return both look strong, but its trailing year and alpha are both negative — that mismatch is worth taking seriously rather than chasing the recent print. AMKR's fresh collapse tag is the more straightforward story: its trailing year is still positive, but it's now trading under its 200-day for the first time in this dataset, so the near-term weakness has real longer-horizon backing. CRS and TER both look worse on the week than they are on the year — both remain comfortably above their 200-day and up well over 100% on the trailing year, which argues their sharp pullbacks are digestion inside intact trends rather than trend breaks. WOLF's outsized 1-year print likely reflects a distorted base from a prior severe decline rather than durable strength; its position below the 200-day is the more trustworthy signal there.

Tier A — Confirmed strength (6): ALM, MP (short-term bounce, negative on the trailing year — flagged above), ELMT, AMVMF (data caveat), SDVKY, ATEYY.

Tier B — Mid-book, cooling (5): TER (sharp giveback, still strong 1-year), ASX, BESIY (thesis intact), WOLF (collapse), KEYS.

Tier C — Weakest, RSI <40 (5): KLIC, AMKR (fresh collapse), TDY, KMT (downtrend, death-cross), CRS (three-cycle whipsaw, still positive 1-year).

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