US Semiconductors — End-of-Day Verdict
0 ACQUIRE, 0 ANTICIPATE, 1 DISPOSE | AMD SELL (100%) @ ~474.32 at the 08-12 open - a RE-ISSUE of the identical 08-10 card, which never executed because it was written execution='market_next_open', a value paper_executor.py silently discards (_EXEC_KINDS={market,gtc_limit,gtc_stop}); today's card uses 'market'. The exit case is unchanged and stronger: close 474.32 still below the 475.00 committed line for a 2nd session, 12wk 457.65 now 3.51% BELOW the close (was -1.95%), 3rd consecutive NO TRADE cv0; realized - (-8.09%) | ACQUIRE NO ACTION: for the SECOND consecutive session the board printed ZERO bullish actionable rows - and this time on a broad RALLY ( up, median +0.96%, KLAC +4.01%, ASML +3.80%), not a rout, which is the more informative version of the same answer | ANTICIPATE slot empty: no name printed BUY - ANTICIPATE in >=2 of the last 3 snapshots (NVDA and LITE each) | TSM HOLD, TRAIL RAISED 389.84 -> 396.58 (4th consecutive ratchet), conviction and 12wk 395.86->403.91 both improved | AVGO cv54 remains the sole near-miss (only 12wk meaningfully above close, +9.4%, Quality PASS), 11 points short of the floor
Recommendations
0 to acquire · 1 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 |
|---|---|---|---|---|---|---|---|---|---|
| AMD | Advanced Micro Devices, Inc | 474.32 | SELL ALL | 474.32 | 475.00⚠ | n/a | n/a | 457.65 (-3.5%) | Selling the whole AMD position. The risk line we set and re-armed four sessions running was decisively broken on 08-10 and the stock has not recovered it - it closed 474.32 today, still below the line, with the 12-week projection (457.65) now 3.5% BELOW the share price and falling. That combination of a broken stop and a negative forward view is our defined exit basis, and the fundamentals are not arguing us out of it: AMD fell 6.9% on the day it reported record quarterly revenue of 54bn (+50% YoY) with data-centre sales doubling to 7bn. When a stock sells off on its own best numbers at ~43x forward earnings - double NVIDIA's ~21x - the price is being set by the multiple, not the business, which is exactly what a mechanical risk line is for. Wall Street disagrees loudly (Citi Buy Bernstein Raymond James) and that is recorded as a genuine counter-argument, but it is standing opinion rather than new news, so it does not override a breached risk line. We book a loss of about (-8.1%) deliberately, at a pre-defined level, rather than leaving it open-ended. | OPERATIONAL: this is a RE-ISSUE of the identical 2026-08-10 card, which never executed. Root cause: that card was written execution='market_next_open', which is NOT in paper_executor.py _EXEC_KINDS={market,gtc_limit,gtc_stop}; _eligibility returned None and the card was dropped SILENTLY - not even reported as skipped (--dry-run today: 0 would fill, 6 skipped, this card absent from the six). All 17 of this book's market_next_open cards have been unfillable. Written execution='market' today so it actually fills. | L2 both legs (2nd session), L3 (execute now - deferred once by machine already), L4 (settled close). Valuation contributes NOTHING: MoS_FV -10.01 amplify SUPPRESSED under G2 (DCF vs a price is a broken anchor). Scorecard NoTrade|sell bias -5 (n=9, hit_1w 0.25) argues AGAINST and is overridden on three grounds: price-line originated not label-originated; L3; optional_trim counterfactual median withheld 4wk -15.33%. ADR-0012 ticks (c)+(d) fire but are NOT the basis - 15th consecutive FALSE arm (T12-progress 103.64% while 8.09% underwater). ADR-0020 erosion 11.98% off the all-time Peak Stop 539.68, under the 15% trigger. Exit range 467.32-481.32 (+/-0.25 ATR 28.0033). Executes at the 2026-08-12 open at market; July CPI prints 08:30 ET that morning (horizon macro, +/-3%) which widens the plausible open - reprice check if it gaps >~1.5% outside 467.2-481.4, but a gap UP through 475.00 does NOT cancel: the escalation was written against the SETTLED CLOSE, now breached twice. 12wk 457.65 exceeds the DCF anchor - annotated, not modified. Model 12wk band -. · news |
| TSM | Taiwan Semiconductor Manufacturing | 422.06 | HOLD | — | 396.58 | n/a | n/a | 403.91 | Holding the 47-share TSMC core and raising the trailing stop again, to 396.58. The business case keeps getting stronger: July revenue of NT58bn was up 44.7% on the year, capital spending is guided to a record -64bn, and TSMC has just agreed a roughly 3bn joint venture with Sony to build next-generation image-sensor chips in Kumamoto. That news was strong enough to lift the equipment makers who supply TSMC - ASML rose 3.8% and KLA 4.0% on the read-through - while TSM itself gained only 0.9%. The market is compounding the business and paying the suppliers for it, which is why these stay a defined core rather than a position we rebuild. Nothing here justifies selling: the price is 6% clear of the stop, conviction improved and the 12-week view rose for a second straight session. The stop does the work from here. | Trail RAISED 389.84 -> 396.58 (2.0xATR chandelier off the 422.06 close, ATR 12.7387), fourth consecutive upward ratchet (382.61 -> 385.57 -> 389.22 -> 389.84 -> 396.58); stops move one way only. No origination class fires: L2 fails leg 1 decisively; label AVOID (bearish family) not SELL - REDUCE; ADR-0020 inapplicable to a monotonically rising trail; PROFIT_ARMED is the 27th CONSECUTIVE FALSE ARM (-3.64% below cost, Gain_ATR -1.25, T12-progress 104.49%). Conviction and 12wk 395.86 -> 403.91 both improved, though the 12wk is still 3.83% under spot. Valuation suppressed under G2 (DCF vs a price). USC-5 continues to describe this name exactly: delivered catalyst, muted price response, de-risk already taken (78% cut 08-04) - hold the core, let the 396.58 line decide. · news |
| AVGO | Broadcom Inc | 416.08 | WATCH | 416.08 | 408.56 | n/a | 446.15 (+7.2%) | 455.03 (+9.4%) | Not buying Broadcom, but keeping it on the watch list as the single most interesting name we are not holding. It is the only stock on our board whose 12-week projection sits meaningfully above today's price (455.03 against a 416.08 close, +9.4%) and one of the few that passes our quality screen. What stops us is that the setup is not yet a buy trigger - the model rates it a WATCH at conviction 54, eleven points below our 65 entry floor - and the stock was the only meaningful decliner today (-1.5%) after a seven-session run of +15.5%, at roughly 68x forward earnings. We would rather buy the confirmation than the hope. | Second consecutive session at cv54 on FailedBreakdown_Up_20D, a family this book is 0-for-4 on (hit_1w 0.0, report-only, n=4) and which USC-3 hardens against as a primary entry basis. Quality_Pass True; MoS_FV -8.12 with val_adj SUPPRESSED under G2 (DCF vs a price). Ledgered direction:none so the counterfactual grader prices what not buying cost or saved. RE-ARM: a BUY label at cv>=65, priced off that session's own Entry_Price - never off today's 416.08. · news |
Outcome & track record
Accuracy at the time of this record.
21 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.