Recommendations
1 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 |
|---|---|---|---|---|---|---|---|---|---|
| KLAC | KLA Corporation | 183.83 | SELL 30% | 183.83 | 184.00⚠ | 185.21 (+0.8%) | 186.12 (+1.2%) | 180.83 (-1.6%) | We are cutting 30% of KLA because the risk line protecting it has fallen 32% since July while the position size never moved once - a stop walking away from a full-size holding is a risk failure, not patience. The signal agrees: KLA printed its second consecutive closing sell signal at the top of the sell rankings, and its twelve-week outlook of 181 is the only one in the book that sits BELOW today's price. The business is doing well - management raised its 2026 equipment-market outlook to the low-$150bn range, lifted advanced-packaging guidance around 70% and carries a 5bn backlog - but the shares are down 14% in a month and 40% off their 52-week high. Good fundamentals are the reason this is a 30% trim and not an exit; they are not a reason to leave a broken risk line untouched. We keep in the semicap cycle. OPERATIONAL: THREE independent originations - (1) ADR-0020/US-36 Trail Erosion, live stop 184.00 vs Peak Stop 272.00 (2026-07-01) = -32.35% with NO size reduction in 21 sessions; this is still the FIRST fire on this watermark because the 08-25 card was never executed, so THE HELD-NAME SELL VETO STACK DOES NOT APPLY (US-17, US-18/34, US-22, US-30 disapplied by rule); mandatory 30%, never a full exit. (2) US-18/US-34 lone-flip bar SATISFIED - 2nd consecutive SETTLED STRONG SELL - REDUCE (08-25 cv68 -> 08-26 cv69), Sell_Rank 1, 88.52; the intervening intraday AVOID cv52/cv47 prints do not break the settled streak (L4/US-7/US-11). (3) L2 - Est_12W 180.83 vs a 183.83 close = -1.63%, the only negative 12wk in the book. Valuation amplify SUPPRESSED by G2: MoS_FV -4.24 would bump the tier, but Est_12W 180.83 vs DCF_per_share 61.80 = 2.97x the anchor -> unreliable -> NO tier bump, stays 30%. scorecard FailedBreakout_Down_20D|sell +0 (n=29). Model Stop_Price 169.01 is BELOW the standing 184.00 trail - a loosening, REFUSED under US-29; trail holds 184.00. Est_12W exceeds the DCF anchor by 2.9x. Retained; watermark RE-BASES 272.00 -> 184.00 on the observed decrease, and any re-fire from the new watermark is subject to the veto stack again. 4th consecutive unexecuted KLAC trim - US-37 stalemate, needs an explicit Disposition. · news |
| GOOGL | Alphabet, Inc.- Class A | 342.00 | SELL 30% | 342.00 | 333.68 | 342.83 (+0.2%) | 340.99 (-0.3%) | 364.26 (+6.5%) | We are banking 30% of a position that has nearly tripled, because its four-week outlook has gone flat and a real catalyst has arrived: Alphabet fell about 4% after Chief Scientist Jeff Dean announced he is leaving after 27 years to start his own company, alongside a report of Gemini model delays. The signal engine has now gone silent for two straight closes and the shares have slipped below both their 20-day and 50-day averages. The business is not broken - second-quarter revenue rose 24% to 8bn, Google Cloud grew 82% versus 37% at AWS and 43% at Azure, and analysts still rate it Strong Buy - and the twelve-week view is still 6.5% higher. That is exactly why we take a third off the table and keep rather than exiting. OPERATIONAL: ADR-0012 armed + FIRED. ARMED = PROFIT_ARMED, T12_Progress 93.89, and US-26 VERIFIED GENUINE (Gain_ATR +32.74, Unreal_PnL_Pct +178.05% - both positive, not the losing-position artifact). FIRED on tick (d): TWO consecutive SETTLED NO TRADE prints (08-25 cv0, 08-26 cv0) - the tick that was one print short last night, when the 08-24 settled print was STRONG SELL - REDUCE. Ticks NOT fired, stated: (a) PROFIT_ARMED is the only Exit_Warning; (b) conviction, no change (the 08-22 transition was already declined as the US-29 NoTrade artifact and is not re-used); (c) Est_12W 364.26 > 1.02 x 342.00 = 348.84, forward alive, AND US-33 blocks the (c) limb outright on a T12-armed name. ONE tick -> profit_take_1 = 30%; never 100% from a profit-take. US-19: this is the FIRST fire of the soft tick under the current arm - execute it as the mandatory 30% card (US-13 bans optional-trim prose), do NOT escalate across subsequent EODs. US-17 is satisfied, not violated: this is not a bare-NO-TRADE sell but an ADR-0012 armed+fired profit-take, one of exactly three valid held-name sell bases; NoTrade|sell reads +0 (n=46,.477) and US-38 already stripped that cell's veto force. G1/G2/G3 govern the valuation-elevate path only and do not block a fired profit-take; G1 is moot (row is NoTrade cv0, not a fresh BreakoutUp/MomentumContinuation at cv>=80). Valuation: MoS_FV -0.69 is above the -3 threshold - no amplify either way, size x1.0. Est_12W exceeds the DCF anchor 145.29 by 2.5x. Overlay: the 2026-07-05 'invalidated below 345' override is 53 days stale and has LOST origination force under US-31 - noted only for completeness that the 342.00 close is below that line. Ex-dividend 09-04 (~6bps, immaterial). Retained, trail 333.68 (the NoTrade row prints no stop; close +2.49% above); watermark 358.00 -> RE-BASES to 333.68 on the observed decrease. ADR-0020 erosion was -6.79%, below threshold. · news |
| SAP | SAP SE | 211.68 | BUY — QUEUED @204.03 | 204.03 | 197.68 | 212.70 (+4.2%) | 221.50 (+8.6%) | 226.18 (+10.9%) | We are placing a limit order to buy SAP at 204.03 - about 3.6% below where it closed - rather than paying up today. SAP has fallen 30% this year and 40% over twelve months, and dropped another 4% on the day UBS downgraded it for having only 17 AI agents ready for broad use. But UBS RAISED its price target while cutting the rating, second-quarter earnings beat with revenue up 9% and cloud backlog up 26%, and management left full-year cloud guidance of EUR 25.8-26.2bn (23-25% growth) untouched. That is an expectations reset, not a broken business - and the shares are now above their 20-day, 50-day and 200-day averages, which is a base breaking out rather than a downtrend continuing. This is also the first genuinely high-quality new buy candidate the screen has admitted in over three weeks. We buy the pullback, not the bounce. OPERATIONAL: fusion 71 raw - 7 (GROWTH_PRICED_IN, MoS_FV -1.33 -> round(-6.65)) + 1 (PullbackBuy_VWAP|buy, n=24) = adj 65, EXACTLY at the floor with no margin. QUALITY GATE PASSES: Quality_Pass True, F-Score 8, Quality_Rank 25/52 - the first quality-passing new BUY to clear the floor since VIST on 08-01, ending a 15-session US-20 outage. Stability: FRESH, ONE settled print (NO TRADE cv0 on 08-22 and 08-25, BUY cv57 on both 08-26 intraday prints, cv71 at the settled close) - stated as fresh, not stable. No-chase: close 211.68, Entry_Distance_ATR -1.39, i.e. 3.61%/1.39 ATR above the entry, far past the -0.25 ACT-NOW bar -> GTC buy-LIMIT, not a fill. Provenance: 204.03 is the anchor snapshot's own printed Entry_Price, not a carried level. Entry range 202.66-205.40 (204.03 +/- 0.25 x 5.49 ATR). expires_on 2026-08-31 (anchor + 3 sessions). Sizing: 0.4924 x x 20% buy_pct_of_portfolio = x0.75 size modulator (-3 < MoS -1.33 <= -1) =; max_position_pct 60% x and both non-binding; / 204.03 = = if filled. Est_12W 226.18 exceeds the DCF anchor 168.64. AUTO-WITHDRAW, pre-committed in writing (US-8/US-9): cancels without re-litigation if (i) the family leaves PullbackBuy_VWAP, (ii) conviction < 65, (iii) Quality_Pass flips False, (iv) a further downgrade or guidance cut lands, or (v) 08-31 passes unfilled. IF SAP GAPS BELOW 204.03, DO NOT FILL AT THE GAP - reprice to the live Entry_Price at the next EOD (US-9). Advisory only, not blended: counterfactuals.no_chase_queue n=26 reads +0.95 at 1wk [cost] but -1.93 at 4wk [saved]. · news |
| NVDA | NVIDIA Corporation | 209.66 | HOLD | 209.66 | 217.07⚠ | 210.38 (+0.3%) | 212.80 (+1.5%) | 222.92 (+6.3%) | Nvidia reported after today's close and delivered a large beat and a much larger raise - revenue 2bn against a 3bn estimate, adjusted earnings versus, and next-quarter guidance of -110.1bn, with Jensen Huang saying AI has reached its inflection point. The shares nonetheless fell over 2% after hours, the fifth straight time the stock has declined on a beat, because this name trades the expectations bar rather than the number. The twelve-week view remains about 6% higher and the trend is intact, so we hold - but we have written down in advance exactly what would change our mind. OPERATIONAL: close 209.66 is -3.41% below the 217.07 trail, yet every dispose limb fails. The 08-25 pre-commitment required a settled close below 217.07 that is ALSO below the POST-PRINT session's low, or any settled close below 201.33 - the print landed AFTER this close, so the post-print session is 08-27 and has not traded: UNTRIGGERED BY CONSTRUCTION. L2's negative-12wk limb fails (Est_12W 222.92 = +6.32%). US-18/34 finds no SELL print at all - the label is HOLD cv52. ADR-0012: armed via T12_Progress 94.05 but US-26 CANNOT BE COMPLETED (Avg_Cost blank), and no tick fires - (a) PROFIT_ARMED alone; (b) = -7, well inside the 20-pt bar; (c) 222.92 > 213.85 and US-33 blocks the limb; (d) label is HOLD, not NO TRADE. ADR-0020 erosion -0.53%, the tightest risk line in the book after MU. Expect an open near ~205, roughly 5.6% below the trail. PRE-COMMITTED TRIGGER (US-35), re-stated and sharpened: a settled 08-27 close below 217.07 that is ALSO below the post-print session's low, OR any settled close below 201.33, fires SELL (30%) at the next EOD without re-litigation. A GAP-DOWN OPEN ALONE DOES NOT FIRE IT (L4 - the settled close decides). Read-through from the raise is a direct positive for MU (HBM content), KLAC (process-control intensity), NBIS (neocloud capex) and GOOGL (TPU vs merchant-GPU); MRVL prints AMC 08-27 as the next check. · news |
| MU | Micron Technology, Inc. | 938.40 | HOLD | 938.40 | 929.07 | 951.34 (+1.4%) | 1009.43 (+7.6%) | 1147.73 (+22.3%) | Micron rose again to 938.40 and sits comfortably above its risk line, with the strongest forward view in the book - about 7.6% higher in four weeks and 22% higher in twelve. It is also the top-ranked name in the portfolio on quality and is up 9% on our cost. Nvidia's guidance for next quarter of -110.1bn reads directly through to the high-bandwidth memory Micron supplies. Nothing here asks us to do anything but hold. OPERATIONAL: HOLD cv66, +0.58%, +1.00% clear of the 929.07 trail. Quality_Rank 1/52, F=8, +9.18% over cost, no Exit_Warning. Model Stop_Price 914.71 is a LOOSENING - refused under US-29; trail holds 929.07. ADR-0020 erosion 0.00% - the cleanest risk line in the book (peak_stop = live stop). MRVL AMC 08-27 is the next ecosystem check; AVGO 09-02 and DELL 09-03 follow. · news |
| NBIS | Nebius Group, N.V. | 213.93 | HOLD | 213.93 | 213.00 | 217.42 (+1.6%) | 229.87 (+7.5%) | 240.06 (+12.2%) | Nebius gave back most of yesterday's 5% gain, closing 3.6% lower at 213.93, but it is still above its risk line and the forward view is intact - roughly 7% higher in four weeks and 12% in twelve. A down day that stays above the line is not a break. We hold, and we have written down the condition that would change that. OPERATIONAL: HOLD cv40, close +0.44% ABOVE the 213.00 trail, so US-22's decisive-break test is not even reached. No Exit_Warning, +1.10% over cost, ADR-0020 erosion -9.63% (below the 15% threshold). Model Stop_Price 205.21 is a LOOSENING - refused under US-29. The overlay's 'keep the core, no adds until the Meta question clears' is 53 days stale (US-31) and is honoured as RESTRAINT only, never origination - NBIS is excluded from any NVDA-print read-through add. WATCH: Event Horizon flags the NVDA ~9.3% NBIS stake lock-up window opening around 09-08 (exact date unresolved) - a structural overhang, NOT a sell basis. PRE-COMMITTED TRIGGER (US-35): TWO consecutive settled closes below 213.00 fires SELL (30%). · news |
| LLY | Eli Lilly & Co | 1189.41 | HOLD | 1189.41 | 1168.46 | 1197.23 (+0.7%) | 1215.43 (+2.2%) | 1213.99 (+2.1%) | Lilly fell 3.6% today, and that pullback is what turned the signal bullish - it printed a buy-the-dip ADD signal for the first time in four sessions, right into its 20-day and 50-day averages with momentum neutral. It narrowly missed our buy threshold once we haircut it for being expensive relative to intrinsic value, so we are not adding tonight. We hold the existing, which sit comfortably above the risk line. OPERATIONAL: BUY - ADD cv73, Buy_Rank 2, out of three straight settled NO TRADE cv0 prints. Fusion 73 - 10 (MoS_FV -2.98 -> round(-14.88) clamped) + 1 (PullbackBuy_VWAP|buy) = adj 64, ONE POINT under the 65 floor. Quality gate irrelevant - F=8, Quality_Rank 5/52, Quality_Pass True - and would not block an ADD to a held name anyway. PROFIT-TAKE DECLINED: armed via T12_Progress 97.98 but US-26 CANNOT BE COMPLETED (Avg_Cost blank), and no tick fires - (a) PROFIT_ARMED alone; (b) conviction, it ROSE; (c) Est_12W 1213.99 vs 1.02 x 1189.41 = 1213.20, alive by (thinnest margin in the book) and US-33 blocks the limb; (d) label is BUY - ADD. Model Stop_Price 1118.53 is a LOOSENING - refused under US-29; trail holds 1168.46 (close +1.79% above). PRE-COMMITTED RE-ARM (US-35): a SECOND consecutive settled BUY - ADD on PullbackBuy_VWAP at raw conviction >= 76 (clearing the -10 haircut to adj >= 67) queues the ADD at the printed Entry_Price in the next EOD's ACQUIRE slot without re-discovery. · news |
| SPCX | Space Exploration Tech Corp | — | HOLD | — | 129.00 | n/a | n/a | n/a | SpaceX has no public quote, so we judge it on news rather than signals. The staggered share-unlock keeps being absorbed better than feared: the first and largest tranche on 6 August freed more shares than the entire IPO and the stock rose 35% in five sessions, adding roughly $500bn of market value and reclaiming its IPO price; the 20 August tranche released a further 319m shares and the stock eased only about 5%. Morgan Stanley reads fundamentals as largely unchanged and current levels as an attractive entry. The next real tests are a 1.3bn-share tranche around third-quarter earnings in early November and the 180-day expiry in December. We hold on the manual stop. OPERATIONAL: muted on signals (permanent no_close) - NO Action, conviction, entry or target is synthesized. Classification: WATCH, not concern - unchanged. The schedule releases ~88% of the 13bn shares through 2027; Musk's 6.4bn stay locked to June 2027. Event Horizon also flags MSCI effective 08-31 and the S&P DJI September review announcement 09-04 (SPCX inclusion is the book's open structural question). Stop 129.00 is a live line, not a formality. No trim, no adds. AUTO-UNMUTES the moment it prints a non-null Close. · news |
Outcome & track record
Accuracy at the time of this record.
164 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 →
-0.191 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.
52 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.