Recommendations
0 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 | 200.47 | SELL 30% | 200.47 | 182.81 | 202.17 (+0.8%) | 205.52 (+2.5%) | 211.37 (+5.4%) | Banking a third of a position whose risk line has walked 34% away from it without a single share ever being sold. KLA's business case is intact and is not what is being cut - Q2 CY2026 revenue rose 15.2% to 66bn with a 0bn next-quarter guide, the shares are up 59% in six months against a 44% industry, Jefferies has it as a 2026 top pick, and a quarterly dividend was declared 06-Aug - but a 1000-share holding carrying a stop that has fallen from 272.00 to 180.13 with zero size reduction is a risk failure independent of price, and stay long the thesis into the AMAT FQ3 read-through on 13-Aug. OPERATIONAL: ADR-0020 Trail Erosion FIRST FIRE - Peak Stop 272.00 (2026-07-01 post-split) vs live 180.13 = -33.78%, never size-cut since the 07-04 opening balance; US-36 makes this first fire IMMUNE to the held-name sell veto stack (US-17/US-18/US-34/US-22/US-30) and fixes the tier at the 30% floor. The signal leg is deliberately absent (NO TRADE cv0, 'Consensus conflict + low conviction') and is NOT the basis. ADR-0012 profit-take TESTED AND DECLINED: armed via the T12_Progress 94.84 limb (Gain_ATR 2.10 < 3.0, US-26 passes at +14.07%), but the conviction drop is a FLIP TO NO TRADE, i.e. a label change not a comparable conviction decay - declined on GOOGL 08-05 and KLAC 08-05 for the same reason and declined consistently here, or tick (b) becomes a back door around US-17; (a) absent, (d) one print not two, US-33 bars the (c)-only path on a T12-limb arm. Valuation amplify (MoS_FV -5.68 <= -3, would bump SUPPRESSED BY G2: Target_12Week 211.37 vs DCF_per_share 83.43 = 2.53x overshoot. Re-price of the unexecuted 08-11 card 192.74 -> 200.47 at the same tier is a RE-PRICE, not a US-19 escalation. PROVENANCE: the anchor row prints NO Entry_Price (NO TRADE), so the reference is the anchor settled close 200.47; the stale 07-01 @280 GTC is dead and is NOT republished. Entry range 197.53-203.41 (+/-0.25 ATR). Proceeds realized ~+ (+14.07%) vs the 175.75 cost, keep 700. Trail on the remainder 180.13 -> 182.81 (1.5 ATR; erosion improves to -32.79%). Forward view IMPROVED vs 08-11 (1wk +0.85%, 4wk +2.52%, 12wk +5.44%) - this cut costs more in expected return tonight than it did a session ago and is taken anyway. 12wk 211.37 exceeds the 83.43 DCF anchor. TENTH consecutive session unexecuted (US-37 stalemate). SHARED with ustech/ussemicon - EXECUTE ONCE (L7). · news |
| GOOGL | Alphabet, Inc.- Class A | 343.80 | SELL 30% | 343.80 | 325.72 | 345.77 (+0.6%) | 357.91 (+4.1%) | 387.04 (+12.6%) | Taking 30% off the book's largest winner after it broke a stop line that was drawn in numbers a week ago, on news that is specific to the company rather than to the tape. Alphabet's DeepMind turnover deepened on 11-Aug - Demis Hassabis stepped down as DeepMind CEO to become chairman, veteran AI scientist Jeff Dean is leaving to start his own venture, and Koray Kavukcuoglu now reports straight to Pichai - which is dated, senior talent flight at the top of the AI organisation, landing on a capex guide already raised to -205bn. The 12-week view is still +12.6% and the close holds above its 200-day, so this is a de-risk on a position up 179% from cost, not an exit: keep running with a reset stop just above the 200-day. OPERATIONAL: THE 08-08 PRE-COMMITTED DECISIVE-BREAK RE-ARM FIRES AS WRITTEN - 'a DECISIVE settled close below 356.00 - 1.5% or more, i.e. <= - fires a further 30% trim'. Actual settled close 343.80 = -3.43% below the stop and -1.96% below the threshold. Tier 30% is PRE-COMMITTED: not escalated (US-19), not de-escalated (US-32). A decisive stop-loss break is a valid CLAUDE.md section 9 exit and is EXEMPT from G1/G2/G3; US-34's bar is met on its 'close below invalidation' limb, not a 2nd print. This is NOT the L2 combo - Target_12Week 387.04 is POSITIVE at +12.58% - which is exactly why the tier stays at 30% (US-22: sector-driven-plus-idiosyncratic break = trim and keep core; the thesis-crack full exit is not warranted). ADR-0012 CHECKED AND NOT THE BASIS: armed genuinely (Avg_Cost 123.00, Gain_ATR 18.32, Unreal_PnL +179.51%, T12_Progress 88.83 - US-26 passes cleanly) with ZERO ticks - (a) no code beyond PROFIT_ARMED, (b) = -4, (c) 387.04 = 1.126x Close, alive, (d) HOLD not NO TRADE. ROBUSTNESS: a Yahoo summary quotes an 11-Aug close of vs the scan's 343.80 (systemic #4) - BOTH are <= 350.66 so the trigger fires on either reference; the scan's settled close is authoritative per L4 and sets the action price. Provenance clean: anchor Entry_Price 343.80, Entry_Distance_ATR 0.0. Entry range 340.79-346.81. Proceeds realized ~+ (+179.5%) vs the 123.00 cost, keep 112. REMAINDER STOP RESET 356.00 -> 325.72 (1.5 ATR) - the broken line must sit below the price and 325.72 is just above EMA200 323.26; the model's 319.70 is looser, REJECTED per US-29. Close is below EMA20 351.99 / EMA50 353.64, RSI 45.9. Tape was soft too (S&P -0.3%, Nasdaq -0.6%, Big Tech leading lower into the 08-12 CPI on the US-Iran Hormuz standoff). Per US-31 the 38-day-stale overlay override ('<345', now invalidated by the close) has NO originating force - the pre-committed stop line does. 12wk 387.04 exceeds the 145.29 DCF anchor (2.66x). · news |
| SPCX | Space Exploration Tech Corp | — | SELL 50% | 133.29 | 114.92 | n/a | n/a | n/a | Halving the SpaceX position the session the second block of insider shares comes free. Tranche 1 released up to 911.5m shares (~$101bn) on 06-Aug and the market absorbed it outright - the stock rose 6.1% that day and 21% over the following week - but the release schedule is not finished: ~319m more shares unlock on 12-Aug, with further tranches reported around 21-Aug and 10-Sep under a staggered nine-stage plan, against a stock down 11% year-to-date and trading below its IPO price. Morgan Stanley reads the fundamentals as largely unchanged and the level as an attractive entry, which is why this is half the position rather than all of it - stay long the story through the remaining supply. OPERATIONAL: CARRIED DISPOSITION, SESSION 5 UNEXECUTED (US-37 stalemate - record TAKEN or SKIPPED). The name is PERMANENTLY MUTED on signals (private-market row, no public quote, no_close), so this is a QUALITATIVE NEWS ASSESSMENT ONLY - no Action, conviction, entry, stop or targets are synthesized, and estimates are n/a because there is no scan row and therefore no Est_* columns. News classification CONCERN, REINFORCED by the tranche-2 date rather than repudiated - re-affirming on NEW information is not the re-litigation US-37 warns against. ACTION PRICE RE-REFERENCED 133.11 -> 133.29 (last confirmed settled close, Aug-10, -3.93% on the day); after-hours 132.75, and a separate search summary quotes ~131.78 - the quote-page close is used and the discrepancy is RECORDED, not averaged. Proceeds keep 80. No Cost in the portfolio file so realized is UNKNOWN (systemic #3). Manual stop 114.92 on the remainder, unchanged. AUTO-UNMUTE to full signal+fusion treatment the moment SPCX appears in the scan CSV with a non-null Close. · news |
| NBIS | Nebius Group, N.V. | 193.23 | HOLD | — | 183.00 | 197.90 | 206.27 | 234.35 | Holding all through tomorrow's Q2 print rather than cutting into it. Nebius reports before the open on 12-Aug with the call at 08:00 ET; the street is at roughly of revenue and a per-share loss, and the whole case rests on how fast the $27bn five-year Meta compute deal converts into booked capacity. The stock rallied 5% into the print and the stop now sits a comfortable 5.6% below the price rather than the 0.6% it was binding at a session ago, so there is room to let the result speak. The de-risk is not cancelled, only sequenced - it executes the evening after the numbers land, whichever way they go. OPERATIONAL: THE ADR-0020 UNCONDITIONAL FIRST FIRE IS NOT YET DUE. The 08-11 pre-commitment reads 'executes UNCONDITIONALLY at the 08-12 EOD, POST-PRINT'; tonight's anchor is the Aug-11 ET settled close, i.e. PRE-PRINT, and firing now would be the pre-trim into a catalyst that CLAUDE.md section 9's catalyst-aware hold forbids and that caused the deferral in the first place. PRE-COMMITMENT RESTATED AND DATE-DISAMBIGUATED SO IT CANNOT BE MISREAD AGAIN: (1) the ADR-0020 30% FIRST FIRE EXECUTES UNCONDITIONALLY at the EOD run anchored on output/us_signals_20260813_06*.csv (= the Wed 12-Aug ET settled close, POST-PRINT), whichever way the print resolves, unless the position has been size-cut by then; (2) any settled close below 183.00, or a 2nd consecutive settled STRONG SELL - REDUCE, fires a MANDATORY 50% trim. Trail erosion stands at Peak Stop 270.00 (2026-06-23) vs live 183.00 = -32.2%, NEVER size-cut ( since the 07-04 opening balance; the 07-17 50% was withdrawn 07-18). Tonight NO TRADE cv0 ('Consensus conflict + low conviction'), close 193.23 (+4.95%), -8.68% vs the 211.60 cost (improved from -12.99%). Not armed (Gain_ATR -0.92, -8.68% - US-26 declines outright). 1.5 ATR = 163.27, no raise. Est bands ~80% coverage: 1wk 197.90 [164.06-237.55], 4wk 206.27 [165.49-292.34], 12wk 234.35 [126.45-532.33] - the widest bands in the book, consistent with the horizon's +/-20% two-sided event band at high confidence. · news |
| NVDA | NVIDIA Corporation | 217.50 | HOLD | — | 212.52 | 218.76 | 221.97 | 232.73 | Holding all. Nvidia closed flat on a soft day for large-cap tech, still 2.3% above its stop and above its 20-, 50- and 200-day averages, with earnings on 26-Aug the next real event. The model has stopped issuing a view on it - two straight sessions of no-trade on conflicting internals - which argues for patience rather than action. OPERATIONAL: A VALID PROFIT-TAKE IS BEING SURRENDERED TO A DATA GAP FOR THE SECOND CONSECUTIVE SESSION. Tick (d) fires squarely - TWO consecutive settled NO TRADE prints (A10C, A11C) - and PROFIT_ARMED printed for a 7th session with T12_Progress 93.46 >= 90 arming the position on the T12 limb. US-26 DECLINES THE ARM OUTRIGHT: Avg_Cost is blank in input/usportfolio.json so Gain_ATR and Unreal_PnL_Pct are blank and the gain is unverifiable. With no valid armed position there is nothing for tick (d) to fire against, and a bare NO TRADE cannot originate a dispose (US-17). NoTrade_Reason verbatim: 'Consensus conflict + low conviction'. Close 217.50 (-0.02%), above EMA20 209.64 / EMA50 206.71 / EMA200 195.40, RSI 58.0. 1.5 ATR = 206.64, below the standing stop - no raise. FILL THE COST FIELD (systemic #3) - GOOGL's was filled and worked perfectly tonight. SHARED WITH ustech/ussemicon - EXECUTE ONCE (L7). |
| LLY | Eli Lilly & Co | 1215.02 | HOLD | — | 1159.23 | 1225.56 | 1245.46 | 1291.81 | Holding all despite a sharp cooling in the model's enthusiasm. Lilly gave back 1.4% but still trades above its 20-, 50- and 200-day averages with a 12-week view 6.3% above the close, and the 08-08 decision to withdraw a trim here has already been vindicated once. Cutting a name whose forward view is still positive and whose trend is intact is what the book's own rules were written to prevent. OPERATIONAL: conviction is a GENUINE within-label tick (b) (-31 pts, a decay inside HOLD, not a flip to NO TRADE - so the US-17 back-door objection that voided KLAC's tick does NOT apply here). It still does not fire: US-26 DECLINES THE ARM (Avg_Cost blank, Gain_ATR and Unreal_PnL_Pct blank, gain unverifiable) so there is no armed position for tick (b) to act on. Even with a valid arm, G3 BLOCKS - Target_12Week 1291.81 is +6.3% above the close on a bullish HOLD print, above EMA20 1179.85 / EMA50 1148.10 / EMA200 1029.71, RSI 57.3 - that is cutting the winner; and G2 independently disqualifies the valuation amplify (DCF_per_share 387.04 vs a 1215.02 close = 3.14x overshoot, so MoS_FV -3.36 cannot bump a tier). Signal rotated AnticipationUp_VCP -> FailedBreakdown_Up_20D; RSI_DIV dropped off the warning row leaving only PROFIT_ARMED. 1.5 ATR = 1148.31, below the standing line - no raise. FILL THE COST FIELD (systemic #3). |
| MU | Micron Technology, Inc. | 868.52 | HOLD | — | 774.61 | 878.50 | 978.23 | 1118.48 | Holding all and lifting the stop for the first time in weeks. Micron edged up 0.9% and now sits fractionally above cost, with a 12-week view 28.8% higher and improving - the memory cycle case is intact even though the model has printed no view on the name for ten straight sessions. Applied Materials' results on 13-Aug are the next read-through for the whole equipment and memory complex. OPERATIONAL: TRAIL RAISE 770.00 -> 774.61 (1.5 ATR) - the first raise available in weeks; it also improves ADR-0020 erosion from -12.5% to -11.98%. ADR-0020 RE-CHECKED AND DOES NOT FIRE: the defensible watermark is Peak Stop 880.00 (07-22, the highest stop on the CURRENT 100-share position) vs a live 770.00 = -12.5%, INSIDE the 15% threshold; the naive 1102.16 all-time print (07-01) remains a US-29 flush artifact and the position was ADDED to on 07-21 @ 868.00), not cut. TENTH consecutive NO TRADE cv0, NoTrade_Reason verbatim: 'Consensus conflict + low conviction' - US-17 bars a bare NO TRADE from originating. Close 868.52 (+0.87%), +1.05% over the 859.49 cost, below EMA20 889.72 / EMA50 887.62, far above EMA200 614.08. Not armed (Gain_ATR 0.14, T12_Progress 77.65). SHARED with ussemicon - EXECUTE ONCE (L7). |
| OKTA | Okta Inc. | 150.33 | HOLD | — | 141.21 | 150.65 | 157.87 | 164.33 | Holding all and nudging the stop up again. Okta was flat on the day, sits 7.5% above cost with the model's conviction rising rather than falling, and the stop is now comfortably above the purchase price - the position is protected into a profit whatever happens next. Earnings land 26-Aug. OPERATIONAL: ARMED BUT NOT FIRED, fourth session running. PROFIT_ARMED prints and US-26 PASSES (real, verified gain, +7.53% on a 139.80 cost), but the arm is via the T12_Progress 91.48 limb ONLY - Gain_ATR 1.73 is well under 3.0 - so US-33 requires a non-(c) tick. NONE FIRES: (a) no second warning code, (b) conviction ROSE, (c) Target_12Week 164.33 = 1.093x Close so not dead in any case, (d) not a NO TRADE. Armed-but-not-fired = MANDATORY TRAIL-RAISE, NOT A TRIM: stop 141.14 -> 141.21 (1.5 ATR; a token, applied because the rule is mechanical, not because it is material). Model Stop_Price 138.17 is LOOSER - rejected per US-29. Close 150.33 (-0.29%), above EMA20 143.23 / EMA50 131.89 / EMA200 105.08, RSI 62.4. |
| NET | Cloudflare, Inc. | 306.97 | HOLD | — | 279.08 | 306.72 | 317.11 | 331.60 | Holding the and declining to add more. Cloudflare's Q2 beat-and-raise is real and the position is up handsomely on it, but the shares have already run about 8% since the print and the model's own one-week estimate is essentially flat against today's price - paying up at the top of a post-earnings move for no projected week is not an add, it is a chase. The add re-opens if the stock comes back to its 20-day average. OPERATIONAL: THE ADD NOW FAILS ON ARITHMETIC, NOT JUDGEMENT. STRONG BUY cv93 -> BUY cv68, Buy_Rank 1, -1.17% to 306.97, halved 85.49 -> 49.24. adj = 68 - 10 (MoS_FV -943.71, an EPV artifact on a name whose FV_per_share is) - 7 (FailedBreakdown_Up_20D|buy cell) = 51, FOURTEEN below the 65 floor, where it cleared at 76 a session ago. Quality gate is inapplicable (NET is held). RE-ARM TRIGGER CARRIED AND RE-BASED: a pullback to the live EMA20 282.53 +/-0.25 ATR (277.42-287.65) with conviction still >= 65 re-opens the ADD at the then-live Entry_Price; the old 279.96 reference is superseded and is NOT triggered (close is +8.7% above EMA20). Trail: 1.5 ATR = 276.28, BELOW the standing 279.08, so STOP 279.08 STANDS; the model's 295.50 at 0.56 ATR is tighter - noted, NOT adopted (whipsaw on a post-earnings name). WARNING, 12TH SESSION: two NET rows in input/usportfolio.json (150 / 0), the loader takes the last, so this position reads Held=N to the scan and unledgered_drift to the monitor. DE-DUPLICATE. |
Outcome & track record
Accuracy at the time of this record.
125 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 →
-0.148 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.