Recommendations
1 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 |
|---|---|---|---|---|---|---|---|---|---|
| KLAC | KLA Corporation | 188.98 | SELL 30% | 188.98 | 174.01 | 188.44 (-0.3%) | 195.38 (+3.4%) | 192.10 (+1.7%) | KLA's business is fine - a fifth straight earnings beat, a raised outlook and a 5bn backlog - but the risk line protecting our position has walked 37.7% away from where it once sat (272.00 in July to 169.53 today) while we never once reduced the position. We bank roughly a third at and keep in the uptrend with a tightened stop. This is risk discipline, not a call on the company. OPERATIONAL: ADR-0020 Trail Erosion, 37.67% vs the 272.00 Peak Stop (2026-07-01), zero size reduction ever, FIRST FIRE on this watermark, therefore VETO-IMMUNE (US-36) - US-17/US-18/US-34/US-22/US-30 do not apply. Pinned at the 30% floor; the MoS -3.93 valuation amplify is deliberately WITHHELD because Step 6(b) scales a signal-originated trim and there is none (the 09-04 STRONG SELL-REDUCE cv76 did not repeat - it lapsed as a US-34 lone flip and today prints AVOID cv64). Trim-discipline gates G1/G2/G3 and the catalyst-aware hold govern the valuation-elevate path only and do not gate an ADR-0020 fire; the 09-11 Goldman Communacopia appearance is a conference, not an earnings/index catalyst. Core trail RAISED 169.53 -> 174.01 (PROFIT_ARMED chandelier, 2.0xATR). Model Stop_Price 203.95 sits ABOVE the close - US-29 bearish-family artifact, ignored. 6th consecutive session this card has been written without a Disposition (US-37 STALEMATE / ADR-0019). On execution the Peak Stop re-bases and future fires lose veto-immunity. · news |
| LLY | Eli Lilly & Co | 1123.91 | SELL 30% | 1123.91 | 1061.36 | 1123.69 (-0.0%) | 1123.47 (-0.0%) | 1111.64 (-1.1%) | Lilly has now spent five straight sessions below the level we said we would defend, closed 2.21% lower at, and is down 6.7% on the month while the sector rose - and the model's own twelve-week view has turned negative. An FDA decision on the oral obesity pill plus intensifying Novo Nordisk competition are the pressure. We take a third off at and hold into the 29-Oct earnings. OPERATIONAL: this is the PRE-COMMITTED trigger written on the 2026-09-04 card firing exactly as specified - a settled close below 1142.00 (1123.91) AND a 5th session below the 1168.46 stop of record with the 12wk still negative (1111.64 = -1.09%). L2 satisfied on both legs; L4/US-22 honoured (settled close, not the intraday). US-26 DECLINES the profit-take path - the PROFIT_ARMED flag here is the known artifact (no Avg_Cost, T12_Progress 101.1 because Target_12Week collapsed BELOW the close), so ADR-0012 is NOT the origination; L2 is. US-17 not implicated (the NO TRADE label is not the basis). Valuation MoS -2.70 is above the -3.0 bar so NO tier bump - 30%, not 50%. Core stop RE-BASED 1168.46 -> 1061.36 (2.0xATR chandelier); the re-base is legitimate only because the position is being cut - the old line had been traded through by 3.81% for five sessions and was no longer a risk control. · news |
| GOOGL | Alphabet, Inc.- Class A | 338.36 | SELL 30% | 338.36 | 335.34 | 337.56 (-0.2%) | 340.63 (+0.7%) | 382.80 (+13.1%) | We are up 175% on Alphabet from a cost and the engine has now gone silent on it for four straight sessions - no buy case, no sell case, just no conviction either way. Meanwhile 2026 capex guidance was lifted to -205bn, which has pushed free cash flow close to zero even as Cloud grows 82%. We bank a third of a very large gain at and keep, because the twelve-week view is still +13%. OPERATIONAL: ADR-0012 armed + fired. Arm FULLY US-26-VERIFIED - Avg_Cost 123.00, Unreal_PnL +175.09%, Gain_ATR 31.26, both legs positive, and the arm comes from the genuine Gain_ATR limb (31.26 >= 3.0), NOT the T12_Progress limb, so US-33's single-metric objection does not apply. Fire tick (d): four consecutive NO TRADE cv0 (09-02/03/04/08) against a 2-print bar. Tick (c) does NOT fire - 12wk 382.80 vs 1.02xClose 345.13, forward view alive. One tick -> profit_take_1 = 30%, never a full exit. US-19 respected: this is a RE-CARD of the unexecuted 09-03/09-04 card at the SAME 30%, not a day-by-day escalation. Valuation MoS -0.63 gives no tier bump in either direction. Core trail HOLDS at 335.34 (chandelier 324.58 is lower; ratchet). Watch 09-16: redaction-motion deadline on the sealed ad-tech remedies memorandum, 6 sessions out. · news |
| VIST | Vista Energy, S.A.B. de | 75.86 | BUY | 75.86 | 74.33 | 76.25 (+0.5%) | 77.45 (+2.1%) | 81.10 (+6.9%) | Vista is the first stock in weeks that is both cheap on our own intrinsic model and passes the quality screen - a rare combination in this market. Production grew 32% year-on-year to 156,061 barrels of oil equivalent per day, management guides to 170,000 by Q4, the Bandurria Sur and Bajo del Toro acquisitions have closed, and Peter Thiel has just taken a roughly 1% stake for about. Eight of nine analysts rate it Strong Buy with an average target near We buy at with a stop at. OPERATIONAL: Buy_Rank 2, Entry_Distance_ATR -0.0 (sitting exactly on the model entry, no chase). Quality gate PASS (F-Score 7) - the ONLY Quality_Pass + UNDERVALUED name in the entire buy pool, which is precisely the drought US-20/US-35 flagged. val_adj 0 (MoS +0.0014 rounds to nothing). Scorecard -3 on FailedBreakdown_Up_20D|buy (n=12, hit_1w.364); raw 76 clears the US-10/US-16 chase-family bar of 75. Size modulator held at 1.0x rather than the permitted 1.25x - UNDERVALUED here means a 0.14% margin of safety, which is fair value, not a bargain. Sizing 49.24% x x 20% = -> =, well inside the position cap and fully covered by. 12wk 81.10 sits BELOW the DCF anchor - no growth-beyond-priced-in warning needed. Earnings 28-Oct, outside the 5-session window. Reprice if the open gaps >1.5%. · news |
| OKTA | Okta Inc. | 167.60 | NO ACTION | 167.59 | 163.59 | 169.61 (+1.2%) | 178.75 (+6.7%) | 189.48 (+13.1%) | Okta has more than doubled this year and we are up 19.9% on it, but the model's conviction in the position collapsed from 61 to zero in a single session and the stock trades on roughly 100x earnings against a mean analyst target of versus today's. A profit-take is warranted - it simply ranks fourth tonight behind three larger, mandatory cards. OPERATIONAL: ADR-0012 armed + FIRED and therefore a MANDATORY dispose candidate, not optional advice. Arm US-26-VERIFIED (Gain_ATR 3.22, Unreal_PnL +19.88%, both positive; genuine gain limb). Fire tick (b): conviction, a 61-point drop against a 20-point bar. Tick (d) does NOT fire (one NO TRADE print, not two); tick (c) does not fire (12wk 189.48 = +13.1% alive). Excluded ONLY by the <=3 DISPOSE slot cap - it ranks 4th on banked dollars at behind KLAC LLY and GOOGL Ledgered direction=none / optional=true so the counterfactual grader prices what deferring cost (US-13 forbids prose-only advice; this is a ledgered card with price and volume, not prose). PRE-COMMITMENT IN WRITING (US-35): this fires FIRST at the next EOD ahead of any new candidate while the arm holds, and the MoS -6.27 valuation amplify then bumps it 30% -> 50%. Tonight: trail HOLDS at 163.59 (chandelier 150.34 is lower; ratchet). Quality_Pass FALSE blocks new entries only, never a held name. · news |
| DELL | Dell Technologies Inc | 533.88 | BUY-STOP | 538.31 | 438.83 | 538.91 (+0.1%) | 542.68 (+0.8%) | 664.22 (+23.4%) | Dell finally clears the two-of-three anticipation test, but the model's conviction in it has fallen hard - to 63 across three sessions - and the stock sits 12.3% above its volume-weighted average price, which is stretched. We pass rather than buy strength into a fading signal. OPERATIONAL: slot-eligible (ANTICIPATE on 09-04 and 09-08; raw cv 63 >= 55) but adjusted conviction is 53 (63 - 10 val_adj on MoS -2.77), two points under the 55 anticipate floor. ASSUMPTION STATED: the floor is read on the ADJUSTED scale, consistent with every other gate in Steps 6-7; on a raw reading it would have qualified. Supported independently by the counterfactual aggregate - anticipate is now n=10, median wh_1w -3.46% [SAVED], i.e. in this regime NOT taking the anticipate has on balance been the better side, which materially reverses the optimism of US-21 (written at n=1). Advisory only per US-28, never blended into conviction. PRE-COMMITMENT (US-35): re-arms if conviction recovers such that adjusted clears 55 while the 2-of-3 print holds. |
| MU | Micron Technology, Inc. | 1000.26 | HOLD | 1000.26 | 929.07 | 1006.83 (+0.7%) | 1139.36 (+13.9%) | 1396.29 (+39.6%) | Micron holds above and we stay long all. We are up 16.4% and the twelve-week view is +39.6% - there is no reason to cut a winner that is still working. OPERATIONAL: PROFIT_ARMED and US-26-VERIFIED (Gain_ATR 3.14, Unreal_PnL +16.38%) but ARMED-AND-NOT-FIRED - no second warning code, conviction is only 8 points against the 20-point tick (b) bar, 12wk alive so tick (c) is silent, and the label is HOLD not NO TRADE so tick (d) is silent. Armed without a tick is a MANDATORY trail-raise, not a trim - but the model's own PROFIT_ARMED chandelier prints 910.46 (2.0xATR on a 44.90 ATR), BELOW the standing 929.07, so the ratchet holds and no numeric raise is available tonight. Stated rather than silently skipped. |
| NVDA | NVIDIA Corporation | 225.73 | HOLD | 225.73 | 220.48 | 227.49 (+0.8%) | 227.84 (+0.9%) | 233.38 (+3.4%) | We hold all 100 Nvidia shares with the stop at, four dollars below Tuesday's close. A profit-take would otherwise be due here, but we cannot verify our own on this position, and the book's rule is never to bank a profit it cannot prove exists. OPERATIONAL: ADR-0012 WOULD fire - PROFIT_ARMED plus tick (b), conviction, a 93-point collapse. BLOCKED for a FIFTH consecutive session by US-26: input/usportfolio.json carries no Cost for NVDA, so Gain_ATR and Unreal_PnL_Pct are blank and the arm cannot be verified. Note the arm here is the NON-artifact shape (T12_Progress 96.72 with Target_12Week 233.38 ABOVE the 225.73 close), which is exactly the case US-26 was not written to catch - but the rule demands the gain legs, not judgement. US-17 independently bars disposing on the bare NO TRADE label. Chandelier 209.97 is below the standing stop; ratchet holds at 220.48. ESCALATED SYSTEMIC S7 - filling one Cost line in the portfolio JSON is worth more to this book than any signal change this week. |
| NET | Cloudflare, Inc. | 284.30 | HOLD | 284.30 | 283.72 | 286.05 (+0.6%) | 287.23 (+1.0%) | 297.68 (+4.7%) | Cloudflare has climbed back above its stop - against - so the breach that had been open since 04-Sep is healed and we hold all. OPERATIONAL: L2 fails on its second leg anyway (12wk 297.68 = +4.7%, positive), so there was never a trim basis. WARNING S9 REMAINS OPEN: input/usportfolio.json still carries TWO NET rows ( / Cost 274.25 and a duplicate); the scan resolves the zero-duplicate and prints Held=N, and portfolio_monitor.py consequently reports this as unledgered_drift rather than a position. The breach healed on its own this time - next time it may not. Delete the duplicate row. |
| SPCX | Space Exploration Tech Corp | 153.47 | HOLD | 153.47 | 140.06 | 153.42 (-0.0%) | 152.87 (-0.4%) | 164.45 (+7.2%) | SpaceX has finally entered the scan universe with a live quote and is no longer flying blind in this book. It closed at against our cost - up 135% - and we raise the protective stop from to rather than trim. The reason to tighten now: a Day-90 underwriter lock-up tranche releases as many as 319 million shares on 09-Sep, which is real supply. OPERATIONAL: AUTO-UNMUTED per the skill's unpriceable-holding rule - first appearance in the scan CSV with a non-null Close, so full signal and fusion treatment resumes and systemic finding S8 is RESOLVED after months open. The stale 2026-07-05 overlay override on SPCX ('stop set + signals unmuted -> normal treatment') is thereby SATISFIED. PROFIT_ARMED and US-26-VERIFIED (Gain_ATR 13.16, Unreal_PnL +135.09%) but ARMED-AND-NOT-FIRED: no second warning code, 12wk 164.45 = +7.2% alive, label AVOID not NO TRADE, and no prior snapshot exists for a conviction delta. Armed without a tick is a mandatory trail-raise - chandelier 2.0xATR = 140.06, ABOVE the standing 129.00, so the ratchet genuinely lifts. The lock-up is an Event Horizon advisory and cannot originate a sell; the raised stop is the correct response to it. · news |
| CRWV | CoreWeave, Inc. | 99.83 | NEAR-MISS | 99.83 | 92.88 | 99.98 (+0.2%) | 107.96 (+8.1%) | 116.73 (+16.9%) | CoreWeave was the highest-ranked buy signal on the board tonight and we are still not buying it. It fails our fundamental quality screen, and the move that produced the signal was an 11.7% one-day jump out of three straight avoid readings - buying that is chasing, not entering. OPERATIONAL: adjusted conviction 79 clears the 65 floor and raw 82 clears the US-10 chase bar of 75, so the BLOCK is the hard quality gate - Quality_Pass FALSE, F-Score 4. Corroborating demotions: volatility pillar BEARISH against a Mixed consensus, VWAP distance +3.66%, and its own Est_1W of 99.98 is +0.15% against the 99.83 action price - a flat forward on a vertical move. val_adj set to 0 because Val_Signal is NA (no resolvable fair value); ASSUMPTION STATED - the row nonetheless emits MoS_FV -57.16, and if that number were taken literally the haircut would be -10 and the name would fail the floor as well. Logged as new systemic finding S11. |
| MSFT | Microsoft Corp | 493.95 | NEAR-MISS | 497.20 | 484.25 | 499.36 (+0.4%) | 535.74 (+7.8%) | 536.36 (+7.9%) | Microsoft is the closest thing to a miss on the buy side - a clean pullback entry on the one signal family this book actually trusts, and it passes the quality screen - but at it is still priced for growth we would be paying for in advance. Two points short. OPERATIONAL: adjusted conviction 63 versus the 65 floor (71 raw, -9 on MoS -1.75, +1 on the PullbackBuy_VWAP|buy scorecard cell, n=28 hit_1w.542). Quality_Pass TRUE (F-Score 6), Entry_Distance_ATR +0.29 so the limit at 497.20 sits above the close - not extended. Ledgered so the counterfactual grader prices the pass. Re-arms on any conviction recovery to 73+ or a pullback that improves the margin of safety. |
Outcome & track record
Accuracy at the time of this record.
197 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 →
-0.171 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.