Recommendations
0 to acquire · 2 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 | 1744.16 | SELL 30% | 1744.16 | 1736.09 | 1772.32 (+1.6%) | 1838.02 (+5.4%) | 1758.44 (+0.8%) | Trimming a third of the ASML stake because the model's own twelve-week view on the stock has gone flat: it now projects just +0.82% over that horizon, the fourth consecutive session of deterioration (from +6.68% a week ago), which triggers the profit-take discipline on a position already flagged as having run its course. The business case is intact and improving — Q2 sales of EUR 9.3B, FY26 guidance raised to EUR 43-45B, and Intel now running Panther Lake in high-volume manufacturing on ASML's next-generation High-NA EUV tools — but a twelve-week path worth under one percent against a stop sitting half a percent below the market is not a risk/reward worth carrying at full size. This is a de-risk of a flat forward view, not a call against the company; four shares stay on with a raised trailing stop. ADR-0012 armed (T12_Progress 99.19) + tick (c) confirmed for a 2nd session: Target_12Week 1758.44 vs the 1.02x dead band of 1779.04, -1.16% through it. Ticks (a)/(b)/(d) clean — bare PROFIT_ARMED is structural proof no Tier-1/2/3 code fired; conviction ROSE (+7). Base 30% tier: G2 suppresses the MoS -4.48 amplify (DCF 841.33 = 2.07x price; 12wk = 2.09x the DCF anchor), val_adj 0. Adjusted 49+0+7 = 56 on the FailedBreakdown_Up_20D|sell cell (n=9, hit_1w 0.833, bias +7) — this book's only strongly positive actionable cell. Provenance: 1744.16 IS the anchor snapshot's Entry_Price and equals the settled close exactly; band 1732.43-1755.89. f=0.30 on the 1wk (L5, calib_1w -0.226 for a 16th run). 12wk exceeds the DCF anchor — projection assumes growth beyond priced-in. Trail on the retained 4 raised 1735.76 -> 1736.09 (EMA50); chandelier 1650.30 is looser and never overrides. Retained risk. Reprice if the 08-26 open gaps >~1.5%. · news |
| TSM | Taiwan Semiconductor Manufacturing C | 417.41 | SELL 30% | 417.41 | 404.59 | n/a | n/a | 396.53 (-5.0%) | Continuing to step down the TSMC position because the model has projected the stock lower over the coming twelve weeks for eight straight sessions — it now sees -5.0% from here — and the signal itself has flipped to a bearish label on a name we hold. That verdict sits squarely against the fundamental picture, which is excellent: 73%+ foundry share, Q2 revenue of with EPS of beating, gross margin at 67.7%, a further committed to Arizona, and a Wall Street consensus target of implying 23% upside. We are respecting the risk discipline rather than the narrative, and deliberately not taking the balance — two shares stay on. Worth being plain about the scale: this is a single share and of a book, and the mechanical trim ratchet, not the signal, is now what is driving the cadence. ADR-0012 armed (T12_Progress 105.27) + tick (c) for an 8th session (12wk 396.53 vs the 1.02x band of 425.76). Ticks (a)/(b)/(d) all clean and (d) RESET — the NO TRADE run broke at one print (08-25 NO TRADE -> 08-26 AVOID cv57), so the 2-consecutive requirement will not be met from this sequence; tick (b) again declined per finding 8 is a NoTrade placeholder recovery, not a +57 event). COUNTERWEIGHT RECORDED: the technical genuinely repaired tonight — close +1.78% back above EMA5 and EMA100, Trend recovered Neutral_True->Bullish, consensus Indeterminate->Consolidation_Bullish, VWAP_Distance +1.18% — but ADR-0012 gives the technical no veto over a fired tick, and the optional_trim counterfactual (n=13, -5.14% 1wk / -15.33% 4wk, 'cost') says under-trimming is this book's characteristic failure. G2 suppresses the MoS -13.76 amplify (DCF 45.96 = 9.08x price). Cell moved with the signal to FailedBreakout_Down_20D|sell (n=1, report-only, bias 0) — so unlike the last 7 firings this card is NOT fighting the -6 NoTrade|sell cell. Adjusted 57+0+0 = 57. Provenance: 417.41 IS the anchor Entry_Price and equals the settled close; band 414.56-420.26. 1wk/4wk n/a — the bearish AVOID row prints a BLANK Target_Price, so there is no risk objective to interpolate against (same convention as the NO-TRADE cards); no f applied. 12wk exceeds the DCF anchor. Trail on the retained 2 raised 404.35 -> 404.59 (EMA100); EMA50 416.84 declined as a trail at 0.14% below the close (inside one bar's noise on an 11.39 ATR), chandelier 394.62 looser. Retained risk. Systemic findings 6 (no min trade size) and 7 (terminal-blind ratchet) are the binding constraints on this position, not the signal. · news |
| NVDA | NVIDIA Corporation | 213.05 | HOLD | 213.05 | 207.19 | 216.57 (+1.7%) | 224.77 (+5.5%) | 219.56 (+3.1%) | Holding all of Nvidia through tonight's results and raising the protective stop instead of trimming. Every leg of the case improved on the day: the seven-session losing streak — the longest since 2022 — ended with a +2.19% close, conviction rose again, and the model's twelve-week view widened back to +3.06%, so nothing in the signal argues for cutting. The quarter lands after tonight's close with consensus at roughly of revenue and of earnings against company guidance of, and the stock at 25.1x forward earnings; the market's attention has already moved past the print to the October guide, which is the real event. Our discipline is explicit that a position is not pre-trimmed into its own catalyst — it is held with a tighter stop — and that is what we are doing, accepting that a sharp gap would clear the stop outright. Armed (T12_Progress 97.03) with ZERO ticks: (a) bare PROFIT_ARMED; (b) conviction, a +10 GAIN and the 3rd rise in 4 sessions; (c) 12wk 219.56 vs the 1.02x band of 217.31, alive at +3.06% and WIDENING from +1.92%; (d) HOLD, no NO-TRADE run. Armed-without-a-tick mandates a trail RAISE where legal. TRAIL ARBITRATION (stated judgement call): chandelier 201.33 and EMA100 205.70 are both BELOW the standing 206.32 so neither is a raise; EMA50 210.56 is now below the market and qualifies on last night's stated test, but is DECLINED at 0.43 ATR below the close on the session of a print the Event Horizon bands at -8%/+9% — a stop inside half an ATR is a market order with extra steps and placing one into the catalyst is a pre-trim in disguise. Adopted instead: a 1.0x ATR clearance line, 213.05 - 5.86 = 207.19 (+, first movement in 4 sessions). ADR-0012 has NO minimum-clearance rule for the raise — logged as systemic finding 9; this is the stated assumption in its absence. The 207.19 stop sits 5.7% INSIDE the lower half of the 196.01-232.22 implied band, so it protects against drift, not the event; risk 213.05->207.19 across = = 0.152% of book. Catalyst-aware hold ACTIVE: trim only on a confirmed lower high or a decisive stop break — neither has occurred. NVDA is 82.2% of capital going in, 87.3% after tonight's two trims execute at the same open, BEFORE the print. L6: no queue may live into 08-26 and none exists. Structure: above EMA100/200 and back above EMA50, still below EMA5/10/20 — a partial reclaim. f=0.30 on the 1wk (L5). 12wk exceeds the DCF anchor 163.98 (1.34x). MATERIAL CAVEAT (systemic finding 10): this position's PROFIT_ARMED flag is itself unsound — it arms off T12_Progress 97.03 while the position is -6.06% / -2.35 Gain_ATR, i.e. the target fell to meet the price rather than the price advancing. · news |
| LITE | Lumentum Holdings | 885.57 | WATCH | 885.57 | 841.10 | n/a | 1036.02 (+17.0%) | 965.87 (+9.1%) | Passing on Lumentum despite it printing the strongest buy signal on the board, because the name fails our quality screen and the setup is exactly the kind this book has lost money on before. The move is real and news-driven — a Barclays upgrade to Overweight with a target on top of a quarter that grew revenue 109% year-on-year to about with earnings of against expected — but the stock closed up 6.67% and sits nearly 6% above its volume-weighted average price, so buying here is chasing a gap rather than entering on a defined edge. The valuation screen is the most stretched in the entire mandate. Recorded as a deliberate no-action so the cost of standing aside gets measured. Quality gate F = HARD BLOCK on a new entry. Signal family FailedBreakdown_Up_20D is USC-3's 3/3 stop-first record in this book. MoS_FV -32.51 -> val_adj -10 (floor). Arithmetic for the record: 79-10-2 = adj 67 on the live cell, or adj 59 on USC-3's mandated -10 read — clears the 65 floor on one reading and fails on the other; the gate settles it either way. USC-3 is explicit that a gate override is NEVER a licence to admit a chase-family buy. VWAP_Distance +5.69%, close 885.57 vs EMA20 847.42. RE-ARM TRIGGER: a PullbackBuy_VWAP confirmation, or the quality gate flipping to pass, at a non-extended entry. · news |
| MRVL | Marvell Technology | 240.38 | WATCH | 240.38 | 231.32 | n/a | 276.63 (+15.1%) | 272.80 (+13.5%) | Passing on Marvell on three independent grounds. It falls short of our conviction bar once the stretched valuation is accounted for; it fails the quality screen; and it reports earnings tomorrow after the close, which our rules forbid buying into. The stock closed up 4.84% and trades roughly 11% above its volume-weighted average price — the most extended reading in the mandate — so there is no non-chasing entry available even setting the other two blocks aside. Adj 68-10-2 = 56 vs the 65 floor, so it fails on CONVICTION BEFORE the gate is reached. Quality gate F. MoS_FV -7.07 -> val_adj -10. Q2 FY27 prints 2026-08-27, one session after the execution open — L6 forbids queueing an add into a post-earnings gap and the Event Horizon carries it as a direct-scope event. VWAP_Distance +11.10%. RE-ARM: reassess after the print, on a pullback to the model entry, with the gate passing. |
Outcome & track record
Accuracy at the time of this record.
32 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.