US Semiconductors — End-of-Day Verdict
0 ACQUIRE, 1 ANTICIPATE (queued), 0 DISPOSE | NVDA QUEUED-ANTICIPATE GTC buy-stop @ 225.76 = (5.81% of a book), stop 206.05 (risk = 0.51%), expires 2026-08-18 - the 08-12 written pre-commitment fires exactly as drafted: ANTICIPATE printed 2 of the last 3 snapshots (cv93 -> cv96), the highest conviction this board has produced since inception, Quality PASS, the mildest MoS_FV on the board (-2.42), and the only quality-passing name besides AVGO with a materially positive 12wk. Priced off tonight's own Entry_Price, never the retired 224.39/225.49 references | ACQUIRE NO ACTION: ASX (the only plain BUY, BuyRank 1) is blocked three ways - Quality_Pass False hard gate, adj 64 one point below the floor, and USC-3 (this book is 3/3 stop-first on FailedBreakdown_Up_20D, and that cell crossed n=5 tonight to go ACTIONABLE at -5, exactly as USC-3 predicted) | ASML's reserved add-after-print tranche DECLINED one session: AMAT beat and raised AMC on 08-13 but fell ~2.5% AH, the anchor close predates that reaction, and the row is 1-of-3 on ANTICIPATE | DISPOSE NO ACTION: TSM's PROFIT_ARMED is a FALSE ARM for a 29th session (T12 leg 100.82% while 1.72% underwater, Gain_ATR -0.66) - only tick (c) fires and the profit-take is withheld; ADR-0020 cannot fire on a trail that has only ratcheted (erosion 0%) | The 08-12 blocks were VALIDATED IN ONE SESSION: COHR -7.99% (quality gate + pre-print run-up) and LITE -5.58% (L6 post-earnings gap). The ASML buy-stop FILLED on strength at 1826.98 and marks +1.15% - USC-4 now 2-for-2 | Trails: TSM 405.26 -> 413.27 (6th ratchet, chandelier 1.5 ATR, correcting the 08-12 use of the looser model stop); ASML 1776.26 STANDS
Recommendations
1 to acquire · 0 to dispose
| Ticker | Name | Close | Action | Action Price Where to act — a buy-limit on a pullback or a buy-stop above resistance, not the current price. Sanity-check vs Close: on a held name, an action price stranded above the close means the queued level was set when the stock was higher — stale, awaiting reprice. | Stop The risk-defined exit, set below the entry for a long (~1.2–1.8×ATR). A held long whose Stop sits at/above the Close is already below its stop — a stale box awaiting reprice. | 1wk Model estimate of the likely price at this horizon — a forward projection, not a guarantee, and separate from the trade's Action Price or Stop. Where the ML estimator is enabled these are per-ticker calibrated forecasts (~80% coverage band) that are refreshed periodically and can drift between refreshes; otherwise they're interpolated from the signal's swing target. Confirm against the settled close. | 4wk Model estimate of the likely price at this horizon — a forward projection, not a guarantee, and separate from the trade's Action Price or Stop. Where the ML estimator is enabled these are per-ticker calibrated forecasts (~80% coverage band) that are refreshed periodically and can drift between refreshes; otherwise they're interpolated from the signal's swing target. Confirm against the settled close. | 12wk Model estimate of the likely price at this horizon — a forward projection, not a guarantee, and separate from the trade's Action Price or Stop. Where the ML estimator is enabled these are per-ticker calibrated forecasts (~80% coverage band) that are refreshed periodically and can drift between refreshes; otherwise they're interpolated from the signal's swing target. Confirm against the settled close. | Rationale |
|---|---|---|---|---|---|---|---|---|---|
| NVDA | NVIDIA Corporation | 225.30 | BUY-STOP — QUEUED @225.76 | 225.76 | 206.05 | 237.58 (+5.2%) | 265.16 (+17.5%) | 252.32 (+11.8%) | Nvidia is the strongest setup this book has produced since inception: the highest conviction on the board (96), the only name pairing a passing quality screen with a meaningfully positive twelve-week view, and the mildest valuation stretch of any semiconductor in the mandate. It is bought on a stop above the market rather than at the market, so capital is committed only if the stock proves itself through the pivot; the AI capital-spending cycle behind it is accelerating (Applied Materials just guided next quarter 6% above consensus) and the analyst median target sits about a third higher. The open risk is the circular-financing critique of Nvidia's $500bn buyer- vehicle, which is why the position is half-sized and stopped 3 ATR below entry. || Ops: fires the 08-12 written pre-commitment - ANTICIPATE printed 2 of last 3 (cv93 -> cv96) at the >=55 slot floor. Priced off tonight's own Entry_Price 225.76 (provenance band 224.12-227.40); the retired 224.39 (08-10, withdrawn) and 225.49 (08-12) references are NOT carried. Sizing 77.68% x x 20% =, x0.50 anticipate slot =, x0.75 size modulator (MoS -2.42) =, /225.76 = = (5.81% of book). Risk = 0.51% of book. val_adj suppressed per note B/G2 (DCF vs close, Rev_Growth +94%). Scorecard cell AnticipationUp_VCP|buy n=1 report-only - no track record either way; review at n>=3 not n>=5. f=0.30 mandatory (L5, calib_1w -0.548). WARNING: Target_12Week 252.32 sits ABOVE its own Est_High 241.10 and BELOW the 4wk Target_Price 265.16 - systemic #1/#2. AUTO-WITHDRAW if conviction <65 at any EOD, if Entry_Price stops printing, or on expiry 08-18; the queue deliberately expires BEFORE the 08-26 earnings print and must be re-decided if still open. USC-4 vehicle (2-for-2). · news |
| ASML | ASML Holding NV | 1847.90 | NO ACTION | 1860.13 | 1776.26 | n/a | 2261.08 (+21.6%) | 1960.14 (+5.4%) | The half-position bought yesterday is working (+1.2% on day one) and the lithography thesis is intact - ASML has raised its 2026 revenue outlook toward $51bn and is expanding EUV capacity 30% a year for two years. The second half of the intended position is held back one more session: Applied Materials' results landed after last night's close and, although they beat and raised, the shares sold off on the news, so the market's verdict on the equipment cycle is not yet visible in ASML's own price. Adding into an unpriced reaction is the mistake this book has a written rule against. || Ops: the 08-12 card reserved half a starter as the add-after-print tranche. DECLINED one session on three grounds: (i) BUY - ANTICIPATE printed 1 of last 3, below the >=2-of-3 slot bar; (ii) the single anticipate slot goes to NVDA on merit (cv96 vs cv81); (iii) the anchor close 1847.90 predates the AMAT reaction, so adding is the L6 shape not the add-after-print shape. RE-ARM IN WRITING: if ASML prints BUY - ANTICIPATE at the 08-14 settled close (2-of-3), OR prints a plain BUY at conviction >=65 with the gate passing, the reserved tranche competes at THAT session's printed Entry_Price - never at tonight's 1860.13 and never at the 1826.98 already paid. Fill reconciled: ussemicon-20260814-0001, @ 1826.98, marked 1847.90 = +/+1.15%; USC-4 buy-stop filled on strength (+2.09% session), the vehicle is now 2-for-2. STOP 1776.26 STANDS - chandelier 1.5 ATR 1747.66 and model 1659.66 are both LOOSER and rejected; a stop never widens. val_adj suppressed under G2 (DCF vs). · news |
| TSM | Taiwan Semiconductor Manufacturing Co | 430.49 | NO ACTION | 434.43 | 413.27 | n/a | 503.29 (+15.9%) | 426.97 (-1.7%) | TSMC's business is running as hot as it ever has - July sales up 45% year on year, full-year growth guided above 40%, capex lifted to -64bn with a fresh $100bn Arizona build - and the shares have barely responded, which says the bar is already high rather than that anything is wrong. The holding is small, a fraction under water, and its signal turned constructive again, so the position is kept and the protective stop is ratcheted up for a sixth straight session rather than trimmed. The stop, not a valuation opinion, decides from here. || Ops: profit-take WITHHELD for a 29th session on a FALSE ARM - PROFIT_ARMED fires solely via the T12_Progress leg (100.82%) while the position is 1.72% UNDERWATER (Avg_Cost 438.02, close 430.49, Gain_ATR -0.66); a profit-take on a loss is not a mechanism this framework has (systemic #3, learnings advisory (a), 8th consecutive decision turning on the defect). Tick (c) fires alone: Target_12Week 426.97 < 1.02 x close 439.10 (a -0.82% view, flat not falling). Ticks (a)/(b)/(d) all fail: no second Exit_Warning code, conviction exactly unchanged, label HOLD not NO TRADE. L2 fails leg 1 - close is +6.22% ABOVE the line. ADR-0020 cannot fire: trail has only ever ratcheted (382.61 -> 385.57 -> 389.22 -> 389.84 -> 396.58 -> 405.26 -> 413.27), Peak Stop IS the live stop, erosion 0%. Valuation MoS -14.26 cannot originate (CLAUDE.md 9) and cannot amplify (G2: DCF vs = 9.5x artifact). TRAIL RAISE 405.26 -> 413.27 = chandelier 1.5 ATR (430.49 - 1.5 x 11.478); model Stop_Price 411.47 is looser and rejected - this CORRECTS the 08-12 run, which took the model's line over the tighter chandelier. New line -4.00% below close, -5.65% below cost. Ledgered as a real card per USC-6, never prose; counterfactual optional_trim n=8 median withheld 4wk -15.33% runs AGAINST this decision and this entry is how that gets counted. · news |
| ASX | ASE Technology Holding | 39.64 | WATCH | 39.64 | 38.70 | n/a | 43.35 (+9.4%) | 41.43 (+4.5%) | ASE Technology is the highest-ranked buy candidate on the board and the business news is genuinely good - advanced-packaging revenue guided above 5bn for the year after a 10% raise, second-quarter revenue up 15.5% and earnings nearly tripled - but it fails this book's quality screen outright, and the setup type behind the signal has lost money every single time this book has traded it. A good story on a failing screen in a losing pattern is not a trade. || Ops: BLOCKED THREE WAYS. (i) Quality_Pass FALSE = hard gate on new entries (F 6, no computable DCF, FV/sh vs a price = 12x artifact). (ii) adj 64 = 69 raw + 0 val_adj - 5 bias, ONE POINT below the 65 floor. The FailedBreakdown_Up_20D|buy cell crossed n=5 tonight and went ACTIONABLE at -5 (hit_1w 0.25, hit_4w 0.0) - USC-3's predicted hardening ARRIVED. (iii) USC-3 verbatim: a FailedBreakdown BUY needs a PullbackBuy_VWAP confirmation or a passing gate before size, and a gate override is not a licence to chase. This book is 3/3 stop-first on this family (TSM 07-07 -6.69%, AMD 07-07 -3.20%, ASX 07-10 -8.65%) - ASX is itself on that list. val_adj: MoS -10.41 would haircut a full -10 to adj 54, SUPPRESSED to 0 per note B on a broken anchor - the generous reading, and it still fails. NOTE systemic #5: USC-3 prescribes the -10 FLOOR at n>=5, the scorer produced -5; under USC-3's own wording this is 59, not 64. RE-ARM IN WRITING: a PullbackBuy_VWAP print at conviction >=75, or a Quality_Pass True refresh, re-opens this name - a FailedBreakdown label alone will not, at any conviction. · news |
| AVGO | Broadcom Inc | 417.82 | WATCH | 417.82 | 410.71 | n/a | 446.27 (+6.8%) | 456.20 (+9.2%) | Broadcom keeps screening as the most interesting name the book cannot yet buy: it passes the quality screen and its twelve-week view sits about 9% above the current price, one of only two names on the board where that is true. But the signal is still only a watch flag, not an actionable buy, so there is nothing to act on - the setup has to trigger first. || Ops: NEAR-MISS, 4th consecutive session, now 9 short of the floor - adj 56 = 61 raw + 0 val_adj (suppressed, DCF) - 5 (FailedBreakdown_Up_20D|buy now ACTIONABLE at n=5). Conviction line finally moved after three flat sessions. Still non-actionable WATCH, Quality_Pass True, Target_12Week 456.20 vs 417.82 close = +9.19%. RE-ARM RAISED from cv >=65 to cv >=70 to absorb the new -5 cell bias - at cv65 it would now adjust to 60 and fail anyway, so the stated trigger is brought in line with the arithmetic. Ledgered direction:none so the counterfactual grader prices the wait. |
Outcome & track record
Accuracy at the time of this record.
23 Graded decisions Past recommendations old enough to score against what actually happened (target hit / stopped out / closed). This is the evidence base — the bigger the number, the more history behind the calls. Click to inspect every one. See the decisions →
-0.548 1-week calibration How much of the eventual 4-week move typically shows up in the first week (0 = almost none, 1 = most of it). Use it to read the 1-week estimates: a low number means prices drift toward the target slowly, so don't over-read a soft week-1.
53 Published records How many dated end-of-day verdicts exist for this book — the length of the public track record you can browse in the Archive below.