Recommendations
0 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 | 223.67 | HOLD | 223.67 | 221.79 | 227.78 (+1.8%) | 237.37 (+6.1%) | 247.03 (+10.4%) | Holding the full 253-share position in Nvidia. The company remains the strongest franchise in the book - last quarter revenue rose 106% to $96bn and profit 126% to $59bn - and the shares still sit above every major moving average with a twelve-week view about 10% higher, so there is no case for selling into a mild pullback. Momentum has clearly cooled, though, and the protective stop is now unusually close beneath the price, so the position is on a short leash rather than being added to. Operational: PROFIT_ARMED printed but is DISCARDED - phantom arm off the T12-progress leg on a position that is -1.38% below cost (Gain_ATR -0.46); learnings advisory (a), same treatment as TSM for 10 sessions. ADR-0012 ticks all negative: no second warning code, conviction -2, not >=20), 12wk +10.4% alive, no NO TRADE. Momentum_Score 26 is the weakest on the board; conviction ->59 is real deterioration but originates nothing. Trail HELD at 221.79 - NOT raised, price fell -0.91% and a trail ratchets on new highs only; headroom = 0.27xATR, and after the ex-div on 09-10. STATED IN ADVANCE: a breach of 221.79 alone does NOT originate a sell (L2 needs stop-break PLUS negative 12wk; 12wk is +10.4%). Catalyst-aware hold: ex-div AND Goldman Sachs Communacopia Vera Rubin update both on 09-10 - hold/trail through, do not pre-trim. f=0.30 (L5, calib_1w -0.226). · news |
| ASML | ASML Holding NV | 1729.52 | HOLD | 1729.52 | 1622.58 | n/a | n/a | 1632.61 (-5.6%) | Holding all six ASML shares and tightening the protective stop rather than selling. The exit case that was proposed eight times and withdrawn last week has not come back: the shares still trade above their short- and long-term averages, and the news is the strongest it has been all year - Intel has now run more than a million wafers through ASML's newest High-NA machines, and both TSMC and Samsung have committed to the same -a-tool platform, with 2026 guidance well ahead of consensus and orders booked into 2028. The position is down about 5% and the signal has weakened, so the stop is raised to keep the loss bounded, but selling a strengthening franchise into good news would be the wrong trade. Operational: PROFIT_ARMED DISCARDED - phantom arm, position -5.33% below cost (Gain_ATR -1.82). Conviction decay = -19, ONE POINT short of ADR-0012 tick (b) >=20 - recorded because it is that close. L2 fails the conjunction: 12wk IS negative (-5.6%) but close is +7.32% above the stop; L2 needs both. ADR-0020 erosion -6.56% vs peak_stop 1736.47, inside 15%. USC-5 governs: technical NOT broken (above EMA10 1718.75/EMA20 1727.59/EMA50 1726.76, far above EMA200 1497.61, Consolidation_Bullish, Trend 62) -> trail, don't trim. Trail RAISED 1611.58 -> 1622.58 (2.0xATR chandelier = 1729.52 - 106.94), the mechanical catcher if decay continues. Row printed an UNUSABLE Stop_Price 1836.46 ABOVE its own close 1729.52 (systemic finding 3, repeat of 09-08's 1877.56) - discarded, chandelier used. No 4wk/1wk: Target_Price blank on the bearish row. Quality gate F does NOT block holds. · news |
| TSM | Taiwan Semiconductor Manufacturing | — | HOLD | 435.36 | 415.38 | 441.48 (+1.4%) | 455.77 (+4.7%) | 441.61 (+1.4%) | Holding the three TSMC shares. This is the healthiest chart in the book - the price sits above every moving average with a bullish trend reading, and the balance sheet is the strongest in the mandate. The twelve-week view is nearly flat, which is worth watching, but there is nothing here that justifies selling, and the position is far above its protective stop. Operational: PROFIT_ARMED DISCARDED for the 10th consecutive session - phantom arm off T12_Progress_Pct 98.58 on a position -0.61% below cost (Gain_ATR -0.26); learnings advisory (a). NOTE: ADR-0012 tick (c) WOULD have fired (12wk 441.61 < 1.02 x 435.36 = 444.07, dead forward view) had the arm been valid - so the phantom-arm defect nearly manufactured a profit-take trim on a losing position tonight. Even if fired, 30% of rounds to 1 share, immaterial vs a book. Signal_Rationale prints [BUY], Trend 78, Consolidation_Bullish. Trail HELD at 415.38 - chandelier computes 414.94, BELOW the existing stop; never lower a trail. Headroom +4.81%. TSM ex-div on 2026-09-16. Monitor reported no_close on TSM (truncated production CSV, 9/20 rows) - outage artifact, stop-checked by hand against the reconstruction. f=0.30 (L5). |
| — | Lumentum Holdings | — | NO ACTION | — | — | n/a | n/a | n/a | Passing on Lumentum for a third session despite the strongest buy signal this book has ever recorded. The business is genuinely booming - full-year revenue up 83% to 01bn, fourth-quarter revenue doubled, the stock has more than doubled this year, and twenty of twenty-five analysts rate it a buy on the AI optical-networking build-out. But it fails our balance-sheet and cash-generation screen, and that screen is a hard stop on opening a new position no matter how good the chart looks. Free cash flow yield is barely a third of a percent against a stock that has already run over 100% this year - a great story we are not willing to pay for on a weak balance sheet. Operational: 2-of-3 ANTICIPATE precondition NOW MET (WATCH 57 -> BUY-ANTIC 100 -> BUY-ANTIC 93), exactly as the 09-08 re-arm predicted - the quality gate is the SOLE remaining blocker. Quality_Pass == False: F_Score 5, Quality_Rank 14/20, FCF_Yield 0.35%. USC-3 forbids overriding the gate to admit a chase. val_adj SUPPRESSED to 0 per note B - the anchor swung MoS -34.07 -> -7.75, FV/sh 24.78 -> 112.98, EPV -> -625.57 on a single 7-day refresh; a 4x move in a week is instability, not information (systemic finding 6). Entry would be a GTC buy-stop 1006.49 (+1.77%, EDATR +0.28, self-limiting), stop 816.44. 12wk +21.9% and 1wk 1120.52 both carry the DCF-anchor warning (T12 1205.16 > DCF 112.98). Mostly_Aligned consensus, Trend 91, Momentum 68 - the only Mostly_Aligned row on the board. RE-ARM: unblocks only if Quality_Pass turns True. |
| — | ASE Technology Holding | — | NO ACTION | — | — | n/a | n/a | n/a | Passing on ASE Technology. The chart is the strongest trend reading in the book and the AI advanced-packaging demand is real - second-quarter packaging revenue rose 36% and management expects that business to double again next year. The problem is that ASE is burning cash to get there: free cash flow is deeply negative as it adds another $2bn of capital spending to this year's plan, and the stock has already risen 169% this year, leaving little margin if the build-out slips. We are being asked to pay a fully-priced multiple for a business that is currently consuming cash rather than generating it. Operational: 2-of-3 ANTICIPATE precondition NOW MET (AVOID 53 -> BUY-ANTIC 83 -> BUY-ANTIC 88) - the re-arm fired as predicted 09-08. BLOCKED on Quality_Pass == False: F_Score 6, Quality_Rank 16/20, FCF TWD -43.77bn, FCF_Yield -14.2%. The gate is reading a REAL cash burn, not an artifact - the news corroborates it directly. Standing SA thesis titled 'When The Future Is Already Fully Priced In' corroborates MoS -10.85 -> val_adj -10. 12wk 43.35 is now POSITIVE (+5.2%), having been BELOW the close on 09-08. ASX is one of this book's three realized FailedBreakdown stop-outs (2026-07-10, -8.65%). Entry would be a GTC buy-stop 40.59, stop 35.81. USC-3: a gate override is not a licence to chase. |
| — | Micron Technology, Inc. | — | NO ACTION | — | — | n/a | n/a | n/a | Not buying Micron tonight - and this is the closest and most frustrating pass of the session, because on the merits it is the best idea on the board. Micron has sold out its entire 2026 high-bandwidth-memory supply at locked-in prices, is guiding to roughly $50bn of quarterly revenue at an 86% gross margin, carries almost no debt, and the average Wall Street target of about sits nearly 50% above where we would buy. It also passes our balance-sheet screen comfortably - the second-highest quality score in the mandate. It is held back purely by a timing rule: our anticipation slot requires the setup to persist for two of the last three sessions and this is only its first such print. It re-arms on the very next one. Operational: BLOCKED SOLELY on the ADR-0012 anticipate precondition - MU printed BUY 76 -> BUY 68 -> BUY-ANTICIPATE 83 = ANTICIPATE prints (though BULLISH ACTIONABLE prints, conviction strengthening). The persistence argument is real but the rule is honoured AS WRITTEN - bending a precondition to admit a trade is how this book got hurt (USC-3/USC-4). Quality_Pass == TRUE: F_Score 8, Quality_Rank 2/20, FCF 2bn, FCF_Yield 2.36%, D/E 0.06, Current_Ratio 3.42, Rev_Growth +259%. adj = 83 - 7 (val_adj, MoS -1.38) = 76, well clear of the 65 floor. Trend 91, above EMA10/20/50/200. Entry would be a GTC buy-stop 1045.66 (+1.74%, EDATR +0.39), stop 907.83; 12wk +16.4% carries the DCF-anchor warning (T12 1196.59 > DCF 772.94). Counter-note: management flagged 'meaningful moderation in the rate of price increases' in FQ4; MU FQ4 reports 2026-09-30. Contrast with LITE: MU's Street target sits ABOVE the model entry, LITE's sat BELOW it. RE-ARM LIVE: a second BUY - ANTICIPATE print on the 09-10 close makes it - and MU, unlike LITE and ASX, has NO GATE LEFT TO FAIL. Ledgered direction:none so the counterfactual grader prices the wait. AnticipationUp_VCP|buy n=1 hit_1w 0.00 report-only - stated, not applied. f=0.30 (L5). |
| — | Marvell Technology | — | NO ACTION | — | — | n/a | n/a | n/a | Declining Marvell despite it ranking first on tonight's buy list. The 4% jump that produced the signal came from the chief executive telling CNBC that the company has 'earned a lot of trust' with its customers - a television appearance, not a new contract - and it lands just after Marvell lost a major Amazon chip programme to Qualcomm. The stock is still down more than 20% over three months, it fails our balance-sheet screen, and the model's own twelve-week view offers barely 1% above the entry. Buying a one-day pop on a talking point, into a name that just lost a flagship customer, is the pattern that has cost this book most. Operational: BLOCKED on three independent grounds. (i) Quality_Pass == False - F_Score 6, Quality_Rank 13/20, FCF_Yield 0.87%; hard gate on new entries. (ii) One-session flip AVOID 64 -> AVOID 64 -> BUY 76 after +4.26% = textbook US-1. (iii) News is a talking-head pop post-Amazon-loss, with a live SA 'Reiterating Strong Sell' thesis. adj = 76 + 0 (val_adj SUPPRESSED per note B - DCF 63.95 vs a 235.01 price is a broken anchor) - 10 (scorecard) = 66, clears the 65 floor by ONE POINT, which is why the gate matters. SCORECARD EVENT: FailedBreakdown_Up_20D|buy crossed n=5 and is now ACTIONABLE at bias -2, but hit_1w 0.40 / hit_4w 0.00 and this book's realized record is 3/3 stop-first - USC-3 pre-committed to the -10 FLOOR on exactly this crossing, so -10 applied, not -2. At -2 MRVL would have scored 74 and only the gate would have stopped it (systemic finding 5). R:R 2.41:1, 12wk +1.3% only. 12wk carries the DCF-anchor warning. |
| — | Sandisk Corporation | — | NO ACTION | — | — | n/a | n/a | n/a | Not buying Sandisk, though it is the highest-quality name in the mandate and the only one our valuation work now calls fairly priced rather than expensive. Three straight sessions of buy signals, a clean balance sheet, and a twelve-week view about 18% higher all argue for it. What rules it out is that the shares have already run 13.5% in four sessions and sit 12% above their twenty-day average - we would be paying up for a move that has largely happened, in a signal family whose recent record in this book is poor. Worth revisiting on a pullback toward the average. Operational: adj = 69 - 7 (val_adj, MoS -1.38) - 10 (USC-3 floor) = 52, thirteen short of the 65 floor. VALUATION UPGRADE: re-rated GROWTH_PRICED_IN -> FAIR_VALUE on tonight's refresh, DCF 1005.14 -> 2625.87, so T12 2087.60 now sits BELOW the DCF anchor and takes NO warning annotation - the only clean forward view in the mandate. F_Score 8, Quality_Rank 1/20, FCF_Yield 4.60%, D/E 0.011 - Quality_Pass TRUE. Three consecutive BUY prints -> 69). R:R 4.00:1. Extended: +13.5% in 4 sessions, VWAP_Distance +11.57%, close +12.1% over EMA20, RSI 62.2. Re-arms on a PullbackBuy_VWAP print (L1 - the family this book should prefer). f=0.30 (L5). |
| — | Advanced Micro Devices, Inc | — | NO ACTION | — | — | n/a | n/a | n/a | Passing on AMD. The signal is bullish and the balance sheet is sound, but the model's own twelve-week projection sits slightly below today's price - it is asking us to buy strength into a forward view that goes nowhere. The setup is also only one session old. No trade. Operational: BLOCKED on two grounds. (i) Anticipate precondition (AVOID 86 -> AVOID 67 -> BUY-ANTIC 76). (ii) 12wk 517.24 sits BELOW the 521.09 close (-0.7%) - a STRONG BUY rationale with a negative forward projection is internally inconsistent (systemic finding 4). val_adj SUPPRESSED to 0 per note B (FV/sh 42.64 vs a 521.09 price, DCF 171.74 - broken anchor; the -11.22 MoS must not originate a -10 haircut off it). Quality_Pass TRUE (F_Score 8, Quality_Rank 6/20). Entry would be a GTC buy-stop 519.46, stop 456.07. Trend 87, Consolidation_Bullish. |
Outcome & track record
Accuracy at the time of this record.
45 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 →
-0.226 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.