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 | 1883.12 | BUY-STOP — QUEUED @1879.54 | 1879.54 | 1704.10 | 1926.56 (+2.5%) | 2230.42 (+18.7%) | 2088.75 (+11.1%) | ASML is the one lithography monopoly in the AI buildout and it has now raised full-year 2026 guidance twice, taking revenue to EUR 43-45bn from EUR 36-40bn on accelerating AI-infrastructure orders, with the interim dividend up 17% and consensus still carrying roughly 20% upside. The chart agrees: a confirmed anticipation setup at conviction 81, printed in 2 of the last 3 sessions, and unlike every other name in its family it passes the quality screen outright (F-Score 7). It is also the least stretched of the group at 10% above VWAP - we are buying a base, not a spike. The stock is priced well beyond any conservative intrinsic anchor, so we take it small and with a defined stop rather than pass on a genuine trend. OPERATIONAL: slot-eligible (2-of-3 ANTICIPATE prints, cv81 >= 55 floor); quality gate PASSES (Quality_Pass True, F 7) - the differentiator vs LITE/SNDK/NBIS. val_adj -10 (MoS -4.92, capped) and 0.5x size tier; x0.5 anticipate-slot factor -> / = 2.43% of, defined risk = 0.227% of. CAVEAT: Entry_Price 1879.54 sits 0.19% BELOW the 1883.12 close (EDA -0.06), so this buy-stop is already through and will fill at/near the open - it is effectively market-on-open, NOT the self-limiting strength filter the slot design assumes. Not a chase (no-chase bar is -0.25 ATR). L8: reprice if the open gaps >1.5%. Scorecard bias 0 - AnticipationUp_VCP|buy is n=1 and UNGRADED; the anticipate counterfactual (n=4, median 4wk -8.81%) argues against upsizing, hence minimum slot size. WARNING: 4wk target 2230.42 sits ABOVE the 12wk projection 2088.75, and 12wk exceeds the DCF anchor 836.35 - read 12wk as the defensible number, 4wk as an upper bound. Withdraw if conviction < 65 or by 2026-08-20. · news |
| DELL | Dell Technologies Inc | 479.81 | HOLD | 479.81 | 455.32 | n/a | 538.49 (+12.2%) | 683.73 (+42.5%) | Dell fell 2.2% on a patent scare it is not actually part of: Netlist filed an ITC complaint over DDR5 memory-module patents naming Super Micro, Micron, HPE and Lenovo - Dell is not a respondent. The semiconductor index rose 2% the same day and Micron, an actual defendant, rallied 4%, which tells you this was a sympathy sale rather than a change in the business. The AI-server franchise is intact, the model still projects roughly 42% upside over twelve weeks, and on our own valuation work Dell is the only name in this book flagged as a genuine bargain on fundamentals (F-Score 8). We hold the full position and raise the protective stop instead of selling into someone else's headline. OPERATIONAL: no sell origination class fires. Conviction (-5, vs the -20 collapse tick); Exit_Warning blank; NOT profit-armed (Gain_ATR 0.63 < 3.0, T12_Progress 70.18% < 90) so ADR-0012 is unavailable; L2 fails both legs (close +5.5% above stop, 12wk +42.5%); G3 forward-view gate would block a trim regardless. Valuation is amplify-only and here argues a trim DOWN a tier (Bargain True) - there is no trim to scale. TRAIL-RAISE 454.65 -> 455.32 = chandelier 2.0 x ATR (29.34) below the 514.00 ATH; raise comes from ATR contracting 29.67->29.34. Engine Stop_Price 465.14 REJECTED as 0.50 ATR too tight on a 6.1%-ATR name (today's low 469.05 was above it). ADR-0020: new stop exceeds prior peak 454.65 so watermark REBASES to 455.32, erosion stays 0.00%. FORWARD: Dell FQ2 ~2026-09-03, just beyond the horizon window - flag at the next refresh, do not pre-position. · news |
| AAOI | Applied Optoelectronics, Inc. | 154.89 | HOLD | 154.89 | 111.93 | n/a | 238.26 (+53.8%) | 168.66 (+8.9%) | Our optics order filled exactly at the limit and is already up 3.2% on day one. The business case is the AI data-centre transceiver ramp - roughly 1bn of revenue this year, a stated target of over 170% revenue growth in 2027, and a multi-year Amazon supply agreement that gives real visibility. Against that, the stock is expensive by any conservative measure and one published analysis puts fair value near so the thesis leans on a small number of large customers and an unproven 800G ramp. Both sides were known when we sized this at half weight; nothing has changed, so we hold and put a protective stop underneath it for the first time. OPERATIONAL: entry FILLED at 150.03 with ZERO slippage (txn ustech-20260818-0001) vs NVDA +0.46% and DELL +2.59% gap-throughs. Row prints WATCH cv51 (down from the BUY cv74 that generated the entry) - a conviction fade on a filled entry, not an exit signal. NOT armed, hand-derived because the scan/executor race blanks the position columns (CSV shows Held=N): Gain_ATR = (154.89-150.03)/16.57 = 0.29 vs 3.0 floor; 12wk progress ~26% vs 90. L2 fails (close +38.4% above stop; 12wk 168.66 = +8.9%). Valuation cannot originate: Val_Signal NA arises from a NEGATIVE FV/share (-3.38), F 4 - already expressed as the 0.5x size at entry. STOP SET 106.06 -> 111.93 (engine chandelier_trail, 2.59 ATR / -27.7% below close); closes the monitor no_stop warning. Risk from cost (150.03-111.93) x 342 = = 1.29% of - wide by design on a 10.7%-ATR name, carryable only at half size. IMPROVEMENT vs entry: the 12wk projection has lifted from 148.34 (below the 150.03 entry) to 168.66, resolving the contradiction flagged on the 08-15 card. · news |
| NVDA | NVIDIA Corporation | 225.01 | HOLD | 225.01 | 206.05 | n/a | n/a | 255.91 (+13.7%) | We hold Nvidia into its 26 August quarter. The set-up is unusually reasonable: the shares trade near 25 times forward earnings against a more typical 35-plus at this point in the year, the company has guided to $91bn of revenue against 9bn of consensus, and last quarter it guided $78bn and delivered close to $82bn. The signal engine has gone quiet - trend still positive, momentum rolling over - which is what a stock does while it waits for news, not what it does when it breaks. The position is a fraction of a percent underwater and well above its stop, so there is nothing to act on before the print. OPERATIONAL: second consecutive NO TRADE cv0 ('Consensus conflict + low conviction'), pillars split Trend Bullish vs Momentum Bearish. Fire tick (d) - 2 consecutive NO TRADE - IS satisfied and fires NOTHING because ticks require an ARMED position: Gain_ATR -0.34 (underwater), T12_Progress 87.93% (< 90 floor) and UT-9 rules an arm on a losing position inert. L2 fails: close +9.2% above the 206.05 stop, 12wk 255.91 = +13.7%. STOP HELD at 206.05, DELIBERATELY NOT TIGHTENED: a mechanical chandelier off the 227.92 high (2.0 x ATR 5.29) gives 217.33 = only 3.4% below the close, inside the horizon's +-6.75% earnings band on a 3-session-old position - a near-certain shakeout on the exact event the position exists to hold through. 206.05 = 3.58 ATR below close. WATCH: a third NO TRADE still does not fire while unarmed; the 08-26 print is what decides this line. · news |
| MU | Micron Technology, Inc. | 1011.75 | BUY-STOP | 1017.79 | 837.13 | n/a | 1379.12 (+35.5%) | 1125.03 (+10.5%) | Micron is the closest thing to a buy we did not take. It rose 4% on the very day it was named in an ITC patent complaint over DDR5 memory - the market judged that a memory maker faces far less import risk than the server builders who ship finished systems - and it is the highest-quality name on our board, comfortably passing the fundamental screen. What held us back is process, not doubt: this was its first anticipation print, and our rule requires the set-up to repeat before we commit capital to a pattern with no track record in this book. One more confirming session and it becomes the front-runner. OPERATIONAL: adj 74 = highest QUALITY-PASSING adjusted conviction on the board. Fails ONLY the anticipate-slot eligibility test - 1 of the last 3 snapshots printed BUY - ANTICIPATE (NO TRADE -> BUY cv69 -> ANTICIPATE cv81), rule requires >=. Quality_Pass True, F-Score 8; val_adj only -7 (MoS -1.35), the mildest haircut in the buy-relevant set. Overlay's standing 'add only on signal' override now HAS its signal. Ledgered direction:none so the counterfactual grader prices what standing down cost or saved. Re-arm trigger: a second consecutive BUY - ANTICIPATE print makes it slot-eligible at adj 74 and it would clear the quality gate, unlike LITE/SNDK. · news |
| LITE | Lumentum Holdings | 968.90 | BUY-STOP | 975.06 | 721.18 | n/a | 1482.83 (+52.1%) | 1169.12 (+19.9%) | Lumentum was the strongest signal on the board and we still passed. It fails on two counts: the set-up appeared only once in the last three sessions rather than the two we require, and the company does not clear our fundamental quality screen. On valuation it is the most stretched name we track by a wide margin - the shares trade at roughly forty times any defensible estimate of intrinsic worth. We already own this exact theme through our optics holding, so there is no need to reach for a lower- quality expression of it. OPERATIONAL: cv 83 = highest raw conviction on the board, adj 73. FAILS TWICE - (1) anticipate-slot eligibility 1-of-3 (WATCH cv61 -> BUY cv76 -> ANTICIPATE cv83; a FRESH anticipation print, not a confirmed one); (2) Quality_Pass False. MoS_FV -37.67, FV/share 24.78 vs a 968.90 close (uncapped val_adj -188, capped -10). Note: one more ANTICIPATE print makes it slot-eligible and the quality gate would STILL block it - do not re-raise this next session. Theme already owned via AAOI. |
| SNDK | Sandisk Corporation | 1786.85 | BUY-STOP | 1712.05 | 1241.62 | n/a | 2652.91 (+55.0%) | 1724.87 (+0.7%) | SanDisk was the day's biggest gainer, up 8.9%, and it has a real catalyst ahead - it is the largest addition to the MSCI World index at the end-August review, which typically pulls in forced buying worth a few percent. We are still standing aside. The company does not pass our fundamental quality screen, and the shares now sit almost 32% above their average traded price, the most stretched reading in the book, two sessions after our own model was telling us to avoid them. Buying that far above the crowd's, into a name we would not own on fundamentals, is the kind of chase this book is built to refuse. We will watch the index event rather than front-run it. OPERATIONAL: adj 72. FAILS TWICE - (1) anticipate eligibility 1-of-3 (AVOID -> AVOID -> ANTICIPATE, a same-session reversal off two AVOID prints); (2) Quality_Pass False, F-Score 5. VWAP_Distance +31.87% = most extended row in the book. Horizon carries a genuine catalyst: MSCI August Index Review effective 2026-08-31, SNDK largest MSCI World addition, band +1.5/+4.0%, stance 'queue-below' - but the quality gate blocks a new entry regardless of catalyst (L1/UT-6 no-chase). Ledgered direction:none to price the gate's cost (learnings backlog item d-iii). Re-arm: quality gate must clear first; catalyst alone is not a basis. |
| MSFT | Microsoft Corp | 480.35 | WATCH | 454.71 | 442.12 | n/a | 473.41 (+4.1%) | 657.19 (+44.5%) | Microsoft screened as our top-ranked buy candidate and we passed on it. It falls just short of our conviction bar before valuation is even considered, it does not clear the fundamental quality screen, and the price has already run more than 5% past the level our model would want to pay - so buying here means chasing. It was also one of the weakest names in the book on the day, down 3%, as money rotated out of software and into hardware. Nothing about the entry is attractive at this price. OPERATIONAL: Buy_Rank 1 and the L1-favoured PullbackBuy_VWAP family, yet rejected on arithmetic not on family. cv 64 misses the 65 raw floor by 1 BEFORE valuation; val_adj -8 (MoS -1.65) -> adj 56. Independently blocked by Quality_Pass False (F-Score 6). Entry_Distance_ATR -2.04 = close is 2.04 ATR past the 454.71 model entry, textbook no-chase; VWAP +5.64%. Down -3.04% on the session. Ledgered direction:none for gate-cost measurement. |
| AMZN | Amazon.com Inc | 261.31 | WATCH | 258.65 | 251.66 | n/a | 269.03 (+4.0%) | 287.86 (+11.3%) | Amazon is the best-priced entry on our board - the share price is sitting almost exactly where our model would want to buy, rather than well above it - but the setup is weakening rather than building: our conviction score has fallen for two straight sessions while the stock has gone nowhere. It also fails our fundamental quality screen and is expensive against intrinsic value. A good entry price on a fading signal is not a reason to commit capital. OPERATIONAL: cv across two sessions on a flat tape = deteriorating, not accumulating. val_adj -10 (MoS -3.01) -> adj 49, far under the 65 floor; Quality_Pass False (F-Score 5). Genuinely the LEAST extended name in the book (VWAP +1.03%, EDA -0.38) and the only candidate near its model entry - if the gate ever clears, this is the L1-preferred expression. Ledgered direction:none for gate-cost measurement. |
Outcome & track record
Accuracy at the time of this record.
16 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 →
0.134 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.