Recommendations
0 to acquire · 3 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 | 220.78 | SELL 30% | 217.55 | 211.56 | 223.18 (+2.6%) | 236.31 (+8.6%) | 229.51 (+5.5%) | Selling roughly a third of our Nvidia stake at and keeping the rest. Two sessions ago the stock jumped 8.7% on a blowout quarter; it has now given the entire gap back, and our signal engine's confidence in it collapsed by 27 points in a single day - from its strongest reading on record to a middling one. On top of the market-wide hit from the Fed chair's hawkish Jackson Hole speech, Nvidia specifically paused its AI Compute Partnership, the financing programme it launched two months ago to help smaller cloud providers buy its hardware, after partners objected to Nvidia wanting a say over their end customers and staff flagged antitrust risk. That is the vendor-financing engine behind a chunk of the demand story being throttled by its own participants. We keep - the core - because the business case is intact and the four-week objective is still 8.6% higher; we take a third off because the momentum that justified the full size is gone. OPERATIONAL: ADR-0012 armed (PROFIT_ARMED, T12_Progress 94.79) + tick (b) conviction = -27; ticks (a)/(c)/(d) do not fire; one tick, no Tier-2/3 code -> profit_take_1 30% tier =. G1 softening inapplicable (family FailedBreakdown_Up_20D, cv 63 < 80) and tick (b) fires regardless per ADR-0012. FINDING 10 CAVEAT: the arming leg is the broken T12_Progress one on a position now -4.08% underwater (Gain_ATR -0.99), so the "profit-take" label is false - the card rests on the deterioration conjunction instead (signal AnticipationUp_VCP [STRONG BUY] -> FailedBreakdown_Up_20D; Trend pillar, Momentum; fresh company-specific negative catalyst). TRAIL HELD at 211.56 - the printed Stop_Price 198.79 is lower and a trail cannot be lowered; clearance is only 0.64 x ATR, inside finding 9's proposed 1.0x floor, so one more session like this becomes an L2 exit case, not a trim. val_adj 0 (MoS -2.28 not <= -3; G2 suppresses anyway, DCF 163.98 = 1.33x). Bias +3 from FailedBreakdown_Up_20D|sell (n=11, hit_1w.667). Reissue of the withdrawn 08-27 card at the L8-repriced level; same, same intent. Executes at the Mon 2026-08-31 open - note the MSCI August review is effective at that session's close, so use the closing auction if the open gaps wide. Retained risk 177 x =. · news |
| ASML | ASML Holding NV | 1696.01 | SELL ALL | 1696.16 | 1736.47⚠ | n/a | n/a | 1642.47 (-3.2%) | Closing the ASML position entirely at, taking a 7.2% loss. This is an excellent business that we are simply not being paid to own on our horizon. ASML just raised its 2026 revenue outlook to roughly -51bn from -47bn, lifted its dividend 17%, and has EUV order books stretching into 2028 - and the shares have still fallen five sessions running, straight through the stop level we set, on volume. Our model now projects the price 3.2% LOWER twelve weeks out, and the signal engine has withdrawn entirely, printing NO TRADE with zero conviction where it read a bullish hold a week ago. When the risk line is broken and the forward view has turned negative, the story stops mattering; we take the loss and redeploy. OPERATIONAL: origination is L2 (decisive close below recorded stop + negative 12wk), NOT ADR-0012. Breach 1,696.16 vs trail 1,736.47 = -2.32% = 0.95 x ATR (vs 0.035 x ATR of noise on 08-27 - twenty-seven times larger, now decisive). Target_12Week 1,642.47 vs 1.02x band 1,730.08 = -3.17%, fifth consecutive decay (+3.55 -> +0.82 -> +0.09 -> -0.89 -> -3.17). Conviction; Action HOLD-> NO TRADE ("Consensus conflict + low conviction"); Pillar_Consensus Consolidation_Bullish->Indeterminate; second close below EMA50 1,734.83. Full not sliced per findings 5/7 (a 30% remainder is / sub-lot-scale) and because finding 11's reaffirm-don't-stack exception has lapsed on fresh deterioration. PROFIT_ARMED prints (T12_Progress 103.27) on a -7.16% position (Gain_ATR -3.08) - finding 10 again; calling this a profit-take would be false labelling. G2 suppresses the MoS -4.48 amplify (DCF 841.33 = 2.02x); val_adj 0. SCORECARD CAUTION STATED AND OVERRULED: cell is now NoTrade|sell, n=11, hit_1w 0.182, bias -6 - this book's worst actionable cell, arguing the exit is early; overruled because the origination is L2, not the conviction floor. 1wk/4wk are n/a: the NO TRADE row prints no Target_Price so the f=0.30 interpolation has no anchor (finding 15); model two-sided band 1,662.24-1,879.22. Reissue of the withdrawn 08-27 full-exit card at the L8-repriced level. · news |
| TSM | Taiwan Semiconductor Manufacturing C | 415.32 | SELL 33.3% | 417.52 | 405.91 | 424.07 (+1.6%) | 439.35 (+5.2%) | 402.81 (-3.5%) | Selling one of our three TSMC shares at and holding the other two. The business is performing superbly - record quarterly profit up 77%, July revenue up 45% year on year, and capital spending raised to -64bn as customers scramble for leading-edge capacity - but our model has projected the price LOWER twelve weeks out for eleven consecutive sessions, and yesterday's bullish momentum reading evaporated in a single day as the whole chip complex sold off. We are not being paid for the holding period we opened this for, so we take a small amount off and keep a core. The position is tiny either way - three shares is barely a tenth of a percent of the book. OPERATIONAL: armed (PROFIT_ARMED, T12_Progress 103.65); TWO ticks fire - (b) conviction = -25, and (c) Target_12Week 402.81 vs 1.02x band 425.87 = -5.42%, eleventh consecutive negative and its worst reading. Two ticks maps to profit_take_2 = 50% of 3 = 1.; the escalation is DELIBERATELY ROUNDED DOWN to under the terminal-tranche rule (findings 5/7) because would strand a single share, and under the 08-27 pre-commitment that TSM is exit-or-hold. AFTER THIS FILL TSM IS AND FORMALLY EXIT-OR-HOLD - no further trim will be written on this name. No full exit available: L2 does not fire (close +2.86% ABOVE the 405.91 trail) and ADR-0012 caps at 50%; ADR-0020 does not fire either (peak stop 405.91 vs live printed 395.69 = -2.52%, far short of the 15% Trail Erosion bar). TRAIL HELD at 405.91 - printed Stop_Price 395.69 is lower and not adoptable; clearance 1.06 x ATR clears finding 9's proposed 1.0x floor. EMA50 417.29 is now at the close and EMA100 405.58 sits on the trail - that is the line to watch. G2 suppresses the MoS -13.76 amplify (DCF 45.96 = 9.08x, broken-anchor class per note B); val_adj 0. Bias +3. Reissue of the withdrawn 08-27 card at the L8-repriced level; second reprice on this name. Momentum pillar collapsed. Retained risk 2 x =. · news |
| NVDA | NVIDIA Corporation | 220.78 | HOLD | 217.55 | 211.56 | 223.18 (+2.6%) | 236.31 (+8.6%) | 229.51 (+5.5%) | Keeping 177 of the 253 Nvidia shares after trimming a third. The stock has round-tripped its entire post-earnings jump in two sessions and the financing-programme pause is a genuine negative, but the position still sits above every major moving average, the model's four-week objective is 8.6% above here, and the twelve-week view remains positive. We hold the core with a stop at and let the trend decide. OPERATIONAL: trail held at 211.56 (printed Stop_Price 198.79 is lower, not adoptable); clearance 0.64 x ATR is inside finding 9's proposed 1.0x floor - flagged as the next risk line in the book. A close below 211.56 with the 12wk still positive is a trail break; a close below it WITH a negative 12wk converts this to an L2 full-exit case. AVGO FQ3 on 09-02 is the cleanest independent check on the datacentre guide; MSCI review effective at the 08-31 close. · news |
| TSM | Taiwan Semiconductor Manufacturing C | 415.32 | HOLD | 417.52 | 405.91 | 424.07 (+1.6%) | 439.35 (+5.2%) | 402.81 (-3.5%) | Holding the remaining two TSMC shares with a stop at. The business is excellent and the position is trivially small; the only bearish input is a twelve-week model view that has been negative for eleven sessions. OPERATIONAL: TSM is now formally exit-or-hold under the terminal-tranche rule - no further trim will be written. Trail held at 405.91 (printed stop 395.69 is lower); clearance 1.06 x ATR. EMA100 405.58 sits on the trail and EMA50 417.29 is at the close - a decisive close below 405.91 with the 12wk still negative is a full-exit case under L2. · news |
| LITE | Lumentum Holdings | 914.76 | WATCH | 895.00 | 860.74 | 936.12 (+4.6%) | 1032.06 (+15.3%) | 1054.53 (+17.8%) | Still not buying Lumentum, and the refusal has now paid. The book declined this name for five straight sessions while it ran +6.7%, +6.0% and +1.8% on blowout results; it fell 6.4% today, giving back a third of that run in a session. The signal itself has downgraded from an outright buy to a watch. The business case remains real - revenue more than doubled year on year - but the company fails our balance-sheet and cash-flow quality screen, and the shares are up around 700% in twelve months. OPERATIONAL: Action BUY cv76 -> WATCH cv54. Quality gate F = hard block on a new entry. MoS_FV -32.51 -> val_adj -10 (floor). USC-3 mandates reading FailedBreakdown_Up_20D|buy at the -10 floor rather than the live -2: 54 - 10 - 10 = adj 34 vs the 65 bar. VWAP_Distance compressed +10.06% -> +2.09%; Momentum pillar 26. Direct vindication of L1/USC-3 and a favourable input to the no_chase_queue counterfactual. |
| MRVL | Marvell Technology | 211.66 | WATCH | 216.62 | 207.72 | 227.31 (+4.9%) | 252.24 (+16.4%) | 238.06 (+9.9%) | Not buying Marvell, and the last two sessions are the reason we do not chase earnings. The company reported a strong quarter with good guidance on Wednesday; the stock fell 1.5% on the print and then a further 10.3% today - the worst decline anywhere in the book. It also fails our quality screen. OPERATIONAL: worst print in the mandate at -10.28%, through EMA5/10/20 and now below EMA50 221.16. Target_12Week cratered 282.59 -> 238.06. Gate F; MoS_FV -7.07 -> val_adj -10; USC-3 -10 floor -> adj 43 vs the 65 bar. Second consecutive vindication of L6 (do not chase post-earnings moves). |
| MU | Micron Technology, Inc. | 958.73 | WATCH | 932.86 | 908.98 | 961.52 (+3.1%) | 1028.38 (+10.2%) | 980.25 (+5.1%) | Not buying Micron, but it is the one genuinely encouraging name on the board. On a day when the median semiconductor stock in this mandate fell 2.9% and nineteen of twenty declined, Micron fell 0.3%. It is also the only holding candidate that combines a passing quality screen with a sensible valuation anchor. The signal is simply not strong enough yet. OPERATIONAL: WATCH cv54 (-7); the mandate's most resilient print at -0.27%. Quality_Pass True and DCF_per_share 772.94 vs 932.86 = 1.21x - a sound anchor, so note B does not suppress the haircut: val_adj -6 (degree-scaled from MoS -1.107), USC-3 bias -10 -> adj 38 vs the 65 bar. RE-ARM TRIGGER unchanged: raw conviction >= 82 on a bullish-actionable print, or a PullbackBuy_VWAP confirmation. |
Outcome & track record
Accuracy at the time of this record.
35 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.