US Semiconductors — End-of-Day Verdict
1 ACQUIRE (queued), 0 ANTICIPATE, 0 DISPOSE | ASML QUEUED GTC buy-stop @ 1826.98 (= Entry_Price 1810.07 + 0.25 ATR, top of the provenance band), expires 2026-08-17 - a USC-3 half-starter placed as a USC-4 buy-stop rather than a market fill because Applied Materials reports FQ3 AMC on 08-13 (ecosystem +/-5%, horizon stance add-after-print) and the 1776.26 stop is only -1.87% below the close, so a down-gap would blow through it; a buy-stop cannot fill on a down-gap at all. The other half of the starter is explicitly reserved as the add-after-print tranche | DISPOSE NO ACTION: AMD (the only live sell case) FILLED at the 08-12 open at 488.80, realized - / -5.29%, which is BETTER than the card's -projection at its 474.32 reference because the open gapped +3.05% through the line - the card had pre-committed in writing that a gap UP does not cancel the exit, and honouring that was correct. The 08-11 executor fix (execution='market' replacing the silently-discarded 'market_next_open') is now VALIDATED IN PRODUCTION | TSM HOLD, TRAIL RAISED 396.58 -> 405.26 (5th consecutive ratchet), conviction and the signal family flipped bearish (FailedBreakout_Down_20D/AVOID) to bullish (AnticipationUp_VCP/HOLD); the armed ADR-0012 tick (c) trim is WITHHELD and ledgered as optional | ANTICIPATE slot empty for a 2nd session: NVDA printed BUY - ANTICIPATE at cv93 - the highest conviction this board has recorded, Quality PASS, mildest MoS_FV (-2.42), 12wk +10.1% - but only 1 of the last 3 snapshots vs a >=2 requirement | Board melt-up: up, median +2.58%, four bullish actionable rows after two sessions of zero, and THREE of the four blocked by different pre-existing rules (COHR quality gate + same-day AMC earnings, LITE post-earnings gap + quality + slot, NVDA slot count)
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 |
|---|---|---|---|---|---|---|---|---|---|
| ASML | ASML Holding NV | 1810.07 | BUY — QUEUED @1826.98 | 1826.98 | 1776.26 | 1862.48 (+1.9%) | 1945.32 (+6.5%) | 1860.20 (+1.8%) | ASML is the only quality-passing actionable buy on a board that just produced four bullish signals, and it triggered after its move had digested rather than into it - the stock rose just 0.6% on a day the sector averaged +2.6%, with RSI at 57 and the entry sitting exactly at the model's level, so nothing here is being chased. The lithography monopoly is compounding: 2026 sales guidance lifted to -45bn at a 54% gross margin, Q3 guided to -12bn, the interim dividend raised 17%, and management saying outright that chip demand is outpacing supply while Intel runs High-NA EUV in volume production. Against that, the shares trade at ~49.5x forward and the model's own 12-week view is only +1.8% above the trigger, so this is deliberately a starter, not a full position. Order is a GTC buy-stop at 1826.98 (= Entry_Price 1810.07 + 0.25 ATR, top of the provenance band), =, expires 2026-08-17. USC-4 vehicle chosen over a market-open fill because Applied Materials reports FQ3 after the close on 08-13 (ecosystem +/-5%, horizon stance add-after-print) and the 1776.26 stop is only -1.87% below the close - a down-gap would blow through it, and a buy-stop cannot fill on a down-gap at all. Size is the 0.5x valuation modulator (MoS_FV -4.52 <= -3), which coincides exactly with the USC-3 half-starter prescription for this cell; the remaining half is reserved as the add-after-print tranche. val_adj 0 (G2-suppressed: DCF anchor vs a price is a broken artifact). Scorecard cell FailedBreakdown_Up_20D|buy is n=4 report-only (hit_1w 0.25, hit_4w 0.0) and per USC-2/USC-3 reads as stop-prone, not neutral - this book's realized record on the cell is 3/3 stop-first and two of those three passed the quality gate, so Quality PASS is not being claimed as an escape. 1wk uses f=0.30 (L5, calib_1w -0.548 degenerate). WARNING: 12wk (1860.20) prints BELOW the 1wk interpolation (1862.48) - inverted term structure, logged as systemic finding #1. 12wk exceeds the DCF anchor - projection assumes growth beyond priced-in. Auto-withdraw if conviction < 65 at any EOD before expiry, or on expiry unfilled. · news |
| TSM | Taiwan Semiconductor Manufacturing C | 429.15 | NO ACTION | 430.13 | 405.26 | n/a | 504.75 (+17.3%) | 419.03 (-2.6%) | Holding the 47-share core and raising the trailing stop to 405.26 rather than trimming, because the signal just improved decisively: conviction jumped and the setup flipped from a bearish FailedBreakout_Down pattern to a bullish AnticipationUp_VCP, with price sitting 5.9% clear of the new stop. The business case is getting louder, not quieter - July revenue of NT58bn was up 44.7% year-on-year, year-to-date revenue is up 37%, Bernstein raised its target to NTarguing CPUs become TSMC's next growth engine as agentic AI needs CPUs alongside accelerators, and a 32bn Sony image-sensor joint venture in Kumamoto adds to the pipeline. The shares are undeniably rich - roughly a third above conventional fair-value estimates - but a richly priced name that is trending is not a sell case, and the trailing stop is where that risk gets managed. WITHHELD TRIM (USC-6, ledgered not prose): the row prints PROFIT_ARMED for a 28th consecutive session and ADR-0012 tick (c) does fire - Target_12Week 419.03 < 1.02 x close 437.73 - which by the letter mandates a 30% trim. Withheld on two grounds: (1) the arm is FALSE - the position is 2.03% underwater (Avg_Cost 438.02 vs 429.15 close, Gain_ATR -0.71) and arms only via the T12_Progress leg at 102.42%, which reads 102% because the target collapsed toward the, not because the position gained; a profit-take on a loss is not a mechanism this framework has. (2) The single tick is contradicted by the same print - conviction ROSE 19 points and the signal family flipped bearish to bullish; ticks (a), (b) and (d) all fail. G3's principle points the same way. Trail ratchet 396.58 -> 405.26 is the 5th consecutive raise (382.61 -> 385.57 -> 389.22 -> 389.84 -> 396.58 -> 405.26); ADR-0020 cannot fire - Peak Stop IS the live stop, erosion 0%. val_adj 0 (G2-suppressed: DCF vs). Valuation cannot originate a sell per CLAUDE.md section 9 and cannot amplify one on a broken anchor. The 405.26 line decides from here. · news |
| NVDA | NVIDIA Corporation | 224.09 | BUY-STOP | 225.49 | 203.50 | n/a | 269.47 (+19.5%) | 246.81 (+9.5%) | Nvidia printed the highest conviction score this book has ever recorded (93) with the cleanest valuation on the board and a properly risk-defined entry, and it is still being left alone today - purely because the anticipation setup has only appeared in one of the last three readings, and the rule requires two. That is a deliberate patience gate, not a view on the company: the setup is a buy-stop at 225.49 that only fills on strength, stop 203.50, with the model's 12-week view at 246.81 (+10.1%), one of only two names in the book with a genuinely positive forward projection. The news is genuinely two-sided - Nvidia is anchoring a >$500bn third-party AI-infrastructure financing effort with Apollo, BlackRock, Blackstone, Brookfield, Goldman and KKR, which could unlock a far larger customer base, while Mark Cuban has warned that same financing makes the market fragile and Microsoft's in-house Maia 300 chip is a real competitive threat. Quality gate PASS, MoS_FV -2.42 is the mildest on the board. RE-ARM IS MECHANICAL: a second consecutive BUY - ANTICIPATE print at the 08-13 close makes it 2-of-3 and eligible at tomorrow's EOD, priced off that session's own Entry_Price - never off today's 225.49, and never off the retired 224.39 trigger from the 08-07 queue withdrawn on 08-10 (note NVDA has closed at 224.09, back within 0.13% of that dead level; provenance forbids reviving it). Ledgered direction:none / optional:true so the counterfactual grader prices what the wait cost. FQ2 FY27 earnings land 08-26. · news |
| COHR | Coherent Corp. | 355.64 | NO ACTION | 355.64 | 336.67 | n/a | 431.54 (+21.3%) | 346.38 (-2.6%) | Coherent looks like the second-best buy on the board and is being passed on for two independent reasons, either of which is sufficient. First, it fails the fundamental quality screen, which is a hard block on opening a new position - the AI-optics story is real (a multi-year Nvidia partnership, datacentre revenue growing over 40% year-on-year, the stock up 195% in twelve months) but the balance-sheet and cash-flow screen says do not open here, and overriding a quality gate to chase a strong-looking chart is precisely the mistake this book has already paid for. Second, and decisively: Coherent reported its fiscal Q4 after the close on the same day as this reading, so the 8.2% gain being scored is a pre-earnings run-up and the signal contains none of the actual result. Buying at tomorrow's open would mean buying an earnings report nobody in this process has seen. Street expected EPS (+62% YoY) on revenue 99bn (+30%). This is the same COHR/LITE pair that round-tripped -14.24% / -8.61% on 08-10 after STRONG BUY prints - USC-3 (a gate override is not a licence to chase) and L6 in one cluster for the second time in four sessions. Ledgered direction:none / optional:true so the counterfactual grader prices the pass. Re-arm requires a passing quality gate AND a post-print settled close, priced off that session's own Entry_Price. · news |
| AVGO | Broadcom Inc | 416.05 | NO ACTION | 416.05 | 408.85 | n/a | 444.86 (+6.9%) | 453.95 (+9.1%) | Broadcom stays on the watch list for a third session rather than entering the book, 11 points short of the conviction floor on a signal that is still only a WATCH. It remains the most interesting name not owned: it is the only holding candidate whose 12-week projection sits meaningfully above the current price (+9.1%) while also clearing the fundamental quality screen, a combination nothing else on this board offers. What holds it back is the tape - on a day the sector averaged +2.6%, Broadcom was the only name that did not participate at all (-0.01%), a second straight session of relative weakness following a +15.5% seven-session run at a rich multiple. Trend is intact but momentum has stalled, and paying up for a stalled leader is not this book's edge. val_adj 0 (G2-suppressed, DCF). Re-arm unchanged: a BUY label at conviction >= 65, priced off that session's own Entry_Price. Ledgered direction:none / optional:true so the counterfactual grader prices the continued wait. |
Outcome & track record
Accuracy at the time of this record.
22 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.