Recommendations
3 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 |
|---|---|---|---|---|---|---|---|---|---|
| DELL | Dell Technologies Inc | 492.20 | BUY | 492.20 | 432.27 | 498.38 (+1.3%) | 553.73 (+12.5%) | 661.19 (+34.3%) | Dell just delivered the loudest AI print of the quarter: record revenue of $47bn (+58% year on year), adjusted EPS of (+203%), a record 9bn of AI server orders and a record $95bn backlog, with full-year revenue guidance lifted from -169bn to $192bn and adjusted EPS from to. The signal engine ranks it the single strongest buy in the book, its balance-sheet quality screen passes cleanly, and on a discounted-cash-flow basis the shares still sit close to fair value rather than stretched. We are taking a deliberately small starter because the stock gapped nearly 16% in one session on this news, and gap-chasing is a documented weakness of this book. OPERATIONAL: L6/US-9 conflict stated and overridden on the record - model prints Market at close with Entry_Distance_ATR 0.00, so there is no lower level to queue at and inventing one would breach the action-price provenance rule; the x0.75 size modulator already caps this at 2.4% of the book. Valuation layer is STALE on this name (FetchedAt 09-01 18:05 ET, pre-print TTM; FV_method chose the EPV branch at 112.78 while the DCF branch reads 485.61 ~= the close) so the -10 haircut is likely harsh - applied unmodified anyway, no ad-hoc relief. US-16 chase-family bar cleared (raw cv 83 >= 75). Pre-committed abort: settled close below 432.27 exits the starter; if it round-trips >8% within two sessions, do NOT average down. · news |
| VIST | Vista Energy, S.A.B. de | 77.03 | BUY | 76.21 | 65.28 | 77.72 (+2.0%) | 79.58 (+4.4%) | 83.88 (+10.1%) | Vista Energy is the only genuinely cheap name in this entire book and it has just broken out to a new 52-week high. It trades on roughly 7.6 times forward earnings against an energy-sector average near 12.8, eight of nine covering analysts rate it a strong buy with an average target about 17% above our entry, and Peter Thiel has disclosed a stake of around 1%. The operating story supports the price: 2026 production guidance was raised to 158 thousand barrels of oil equivalent per day, the Equinor asset consolidation adds roughly 22 kboe/d and about of EBITDA this year, and management targets 250 kboe/d by 2030. Uniquely among tonight's buys, our twelve-week projection sits BELOW our own discounted-cash-flow anchor - this is the one name we are not paying up for future growth to own. OPERATIONAL: US-35 pre-committed trigger from 09-01 fired STRONGER than written (a full BUY, not a second ANTICIPATE), so the name graduates out of the anticipate slot and frees it for AAPL. Model prints Buy Stop @75.40 but the close is 77.03 (Entry_Distance_ATR -0.50) - S3 anomaly recurred, so fill is CAPPED at the top of the entry range, 76.21. Earns the x1.25 UNDERVALUED + Quality_Pass up-modifier, first time in weeks. Auto-withdraw: family leaves BreakoutUp_52W, conviction <65, Quality_Pass flips False, or a settled close below 65.28. If the 09-03 open gaps above 77.71 (+1.5%), do NOT fill - reprice at the next EOD (US-9). · news |
| AAPL | Apple | 324.96 | BUY-STOP — QUEUED @328.04 | 328.04 | 305.86 | 323.57 (-1.4%) | 317.41 (-3.2%) | 334.76 (+2.0%) | Apple has a dense catalyst window: the 'Surprise and shine' launch event lands on 9 September with the iPhone 18, the company's first foldable phone (prediction markets put it near 81%) and a rebuilt Siri, and John Ternus has just taken over as chief executive from Tim Cook. Morgan Stanley expects a roughly iPhone price increase worth 2-4% to near-term earnings. It carries the strongest balance-sheet quality score in the book. We are buying only a half-size starter, and only if the shares trade UP through our trigger, because the honest read is that the reward looks capped - our own one-week and four-week projections both sit below the entry price, and the Street's average target is barely above it. This is a small, strength-confirmed position into an event, not a conviction bet. OPERATIONAL: REPRICED from the unfilled 323.17 to the live 328.04 per the action-price provenance rule - the stale limit is NOT carried forward. S3 anomaly self-corrected this session (stop is now 0.95% above the close, so no fill cap needed). Forward DETERIORATED vs 09-01: 4wk went -1.51% to -3.24% versus entry, conviction. NOT withdrawn because none of the 09-01 pre-committed triggers fired (family intact, cv 77 >= 65, Quality_Pass True, expiry not reached); killing it on estimates alone would be the ad-hoc override US-19 bars. counterfactuals.anticipate n=8, median -3.46%/1wk and -9.86%/4wk [saved] - cited per Step 4.4, NEVER blended into conviction. NEW pre-committed withdraw trigger: if the 09-03 settled close leaves Est_4W_Point below the action price AND conviction <70, withdraw outright rather than reprice a second time. · news |
| KLAC | KLA Corporation | 172.26 | SELL 30% | 172.26 | 160.55 | 173.51 (+0.7%) | 179.81 (+4.4%) | 179.81 (+4.4%) | This is a risk-discipline trim, not a call on the company. KLA's business is doing well - record June-quarter revenue of 66bn, a fifth consecutive earnings beat, September-quarter guidance of about $4bn, and 29 analysts split 5 strong buy / 13 buy / 11 hold with zero sells. The problem is our own risk management: the protective stop on this position has been walked down from to over two months - a 41% retreat - while the position size has never once been reduced. A risk line that keeps moving away from an untouched position is a discipline failure regardless of how good the story is, so we take the mandatory 30% off and keep working. OPERATIONAL: ADR-0020 / US-36 Trail Erosion, the third origination class. Peak stop 272.00 (07-01) -> live 160.55 = 40.97% erosion, zero size reduction since that peak, still the FIRST fire on this Peak Stop - so the held-name sell veto stack (US-17, US-18/34, US-22, US-30) EXPRESSLY DOES NOT APPLY. Restate this every session or US-17 kills the rule. No signal basis claimed: AVOID on a long-only book is not a sell instruction (US-29) and the printed Stop_Price 185.13 sits ABOVE the 172.26 close (bearish-family artifact) - both refused. US-26 declines the PROFIT_ARMED flag as an armed-loser artifact (Gain_ATR -0.54, Unreal -1.99%). Valuation amplify SUPPRESSED by G2 (Est_12W 179.81 / DCF 61.80 = 2.91x, anchor unreliable) so no bump. Residual on 160.55; model 185.13 refused as artifact, 2.0xATR chandelier 159.39 refused as looser (US-29). Peak stop re-bases on fill. US-37 STALEMATE: 6th consecutive unexecuted KLAC card (08-25/26/27/28, 09-01, tonight), zero shares sold in two weeks - requires an explicit Disposition, not a 7th silent re-fire. · news |
| NBIS | Nebius Group, N.V. | 204.09 | SELL 30% | 204.09 | 191.02 | 211.24 (+3.5%) | 227.17 (+11.3%) | 286.98 (+40.6%) | Two things argue for taking 30% off Nebius while keeping the core. First, our protective stop has fallen 21% below its high-water mark without the position ever being trimmed, which is the same risk failure we are correcting in KLA. Second, there is a dated supply event six sessions away: Nvidia's filing discloses a 9.3% stake including about 21 million shares underlying a warrant that cannot be exercised before 11 September - and the restriction on selling those shares lapses the same day. The stock has bounced twice but has not reclaimed the level we committed to in writing. We are trimming rather than exiting because the business case is intact and our own twelve-week projection is still roughly 40% above the current price - so stay long the story. OPERATIONAL: PRIMARY basis is ADR-0020 / US-36 Trail Erosion, FIRST fire on NBIS - peak stop 235.69 (08-17) -> live 185.70 = 21.21% erosion, no size cut since that peak, so veto-immune. Note the mechanism: last night's own reset of the line 213.00 -> 185.70 deepened the erosion, which is exactly the pathology ADR-0020 catches. SECONDARY: the pre-committed 213.00 line is still broken (close 204.09, -4.18%), so US-22's settled-close self-heal has NOT triggered. US-17 respected - the bare NO TRADE cv0 label is not the basis; L2 does not fire (Est_12W +40.61%, positive). Valuation amplify SUPPRESSED by G2 (DCF_per_share blank, Rev_Growth 8.55x hyper-growth) so the conservative 30% tier stands, no bump. Residual on a stop RAISED 185.70 -> 191.02 (close - 1.0 x ATR 13.068). US-31: the 'keep core' overlay override is 60 days stale, no originating force, and it argues FOR holding so its staleness cuts against the core. AUTO-WITHDRAW: if NBIS reclaims 213.00 on the next settled close, this card is withdrawn. 2nd consecutive unexecuted NBIS card. · news |
| NVDA | NVIDIA Corporation | 224.41 | HOLD | 224.41 | 217.07 | 226.08 (+0.7%) | 226.37 (+0.9%) | 232.31 (+3.5%) | Holding all. Nvidia rose 3.2% and the engine's confidence in it actually increased, so there is nothing here that argues for taking profit; the trailing stop at now sits a comfortable 3.3% below the price rather than the hair's breadth it was yesterday. OPERATIONAL: PROFIT_ARMED (T12_Progress 96.60) but NO tick fires - (a) no second code; (b) conviction ROSE; (c) Target_12Week 232.31 > 228.90, forward alive; (d) no NO TRADE (HOLD x4). US-26 UNVERIFIABLE - no Cost in input/usportfolio.json so Gain_ATR / Unreal_PnL_Pct are blank and the arm cannot be confirmed real; declined to originate rather than guessing (systemic finding S7). Trail 217.07 held - model Stop_Price 207.48 and the 2.0xATR chandelier both refused as looser (US-29). The 09-01 warning that this line sat 0.17% under the close has resolved. |
| LLY | Eli Lilly & Co | 1160.08 | HOLD | 1160.08 | 1168.46⚠ | 1164.99 (+0.4%) | 1207.71 (+4.1%) | 1182.57 (+1.9%) | Holding all despite a second consecutive close marginally below our stop. The shortfall is under 1%, the twelve-week outlook is still positive, and we committed in writing yesterday that only a close below would open the exit case - that line has not been touched. We honour the line we wrote rather than reacting to a 0.7% brush. OPERATIONAL: PROFIT_ARMED via the T12 limb ONLY (T12_Progress 98.10; no Cost recorded -> Gain_ATR / Unreal blank). Ticks (c) AND (d) technically fire - Target_12Week 1182.57 < 1.02 x 1160.08 = 1183.28, by / 0.06%, and two consecutive NO TRADE prints (09-01, 09-02) - BUT US-33 disqualifies (c) as the same metric the arm reads, and US-26 cannot be verified without Cost, so the profit-take is DECLINED to originate (systemic finding S7). Stop breach now 2 sessions old (1160.00 then 1160.08 vs 1168.46, -0.72%) but NOT decisive: L2 does not fire (Est_12W +1.94%, positive). Decisive line 1142.00 RE-AFFIRMED. Stop 1168.46 unchanged. |
| MU | Micron Technology, Inc. | 956.08 | HOLD | 956.08 | 929.07 | 970.64 (+1.5%) | 1127.25 (+17.9%) | 1328.71 (+39.0%) | Holding all. Micron gained 2.4% and is now a comfortable 2.9% clear of its trailing stop, the position is up 11% on cost, and nothing in the signal or the forward view argues for reducing. OPERATIONAL: NOT armed - Gain_ATR 2.31 < 3.0 and T12_Progress 71.96 < 90, no Exit_Warning printed, so ADR-0012 does not reach. Trail 929.07 held; model Stop_Price 927.16 refused as looser. The 09-01 warning that MU sat 0.47% above its trail has resolved. |
| OKTA | Okta Inc. | 163.15 | HOLD | 163.15 | 160.20 | 164.55 (+0.9%) | 174.50 (+7.0%) | 187.46 (+14.9%) | Holding all. Okta was the only holding to fall today, down 2.0%, but it remains 1.8% above its trailing stop and up roughly 17% on cost, and a quiet signal on its own has never been a good reason to sell in this book. OPERATIONAL: not armed (no Exit_Warning; Gain_ATR 1.94, T12_Progress 87.03). US-17 - bare NO TRADE cv0 is not a dispose basis. Trail 160.20 held (close - 2.0xATR = 139.12, refused as looser). |
| NET | Cloudflare, Inc. | 272.74 | HOLD | 272.74 | 283.72 | 274.00 (+0.5%) | 274.30 (+0.6%) | 284.93 (+4.5%) | Cloudflare fell 4.5% and has closed 3.9% below our protective stop - a genuine risk event that our monitoring missed because the holdings file lists this position twice and the scan therefore treats it as unowned. We are not selling tonight: the twelve-week outlook is still about 4.5% above the current price and there is no confirming sell signal. But we are writing the line down now - a settled close below opens the exit case - and the duplicate holdings entry needs fixing before this happens again. OPERATIONAL: MASKED STOP BREACH. input/usportfolio.json carries two NET rows ('Clouadflare Inc' and 'Cloudflare, Inc.'), so the scan reads Held=N and portfolio_monitor.py reported 'unledgered drift' instead of a breach (systemic finding S9). Not a dispose card: L2 does not fire (Est_12W 284.93 = +4.47%, positive), no signal basis (NO TRADE cv0), and ADR-0020 does not reach (peak 301.75 -> 283.72 = 5.98% erosion, under the 15% threshold). PRE-COMMITTED DECISIVE LINE: settled close below 270.58 (283.72 - 0.75 x ATR 17.520). Run swingtrade2.py trade reconcile and dedupe the file - 2nd consecutive session flagged. |
| SPCX | Space Exploration Tech Corp | — | HOLD | — | 129.00 | n/a | n/a | n/a | Holding all on the manual stop. SpaceX is now a public company - it listed at in the largest IPO on record, raising 7bn, jumped about 20% on debut and now carries a market value near trillion on under $19bn of 2025 revenue. There is no crack in the business case; the live risk is share supply, because insider lock-ups release in stages through mid-December. We keep the position and watch the release calendar. OPERATIONAL: MUTED on the verdict - absent from the scan CSV, so no Action, conviction, entry or targets are synthesized (never guessed). News-assessed as required by the skill's unpriceable-holding rule: classified 'watch'. Lock-up schedule: up to 20% of insider shares freed late-Jul->Aug, then 7% tranches at 70/90/105/120/135 days post-IPO, a further 28% after the Q3 call, all restrictions lifting ~mid-December. Manual stop 129.00 stands. SYSTEMIC FINDING S8: the permanent-mute rationale ('private, no public quote') has EXPIRED - SPCX trades on Nasdaq and must be added to the scan universe; a 300-share position is currently running with no signal and no automated stop check. · news |
Outcome & track record
Accuracy at the time of this record.
175 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.