Recommendations
0 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 | 192.74 | SELL 30% | 192.74 | 180.13 | 194.22 (+0.8%) | 192.50 (-0.1%) | 194.10 (+0.7%) | Bank a third of a position that has never been reduced in ten weeks. KLA's business case is intact - Q2 CY2026 revenue rose 15.2% to 66bn and next-quarter guidance of 0bn came in above consensus - but the shares are down about 11% on the month on China export-rule risk and a broad semiconductor sell-off, and the model's own forward view is flat across all three horizons (1wk +0.8%, 4wk -0.1%, 12wk +0.7%). Taking 30% off at roughly banks a 9.7% gain against the cost and costs nothing in expected return, while stay long the AI capex thesis into the Applied Materials read-through on 08-13. OPERATIONAL: ADR-0020 Trail Erosion FIRST FIRE - Peak Stop 272.00 (2026-07-01, post-split) vs live 180.13 = -33.78% with zero size reduction ever recorded. Per US-36 this first fire is IMMUNE to the held-name sell veto stack (US-17, US-18/34, US-22, US-30); the signal leg is deliberately absent (AVOID cv61, 0, blank Sell_Rank) and is NOT the basis. Tier held at the 30% ADR-0020 floor: the MoS_FV -5.68 one-tier amplify to 50% is SUPPRESSED by G2 (Target_12Week 194.10 vs DCF_per_share 83.43 = 2.33x overshoot, DCF anchor unreliable). Re-pricing the unexecuted 08-08 card 198.11 -> 192.74 at the same tier is a re-price, not a US-19 escalation. Model Stop_Price 216.96 prints ABOVE the close - US-29 artifact, REJECTED; stop on the 700 remainder stays 180.13. Ninth consecutive session unexecuted (US-37 stalemate). · news |
| SPCX | Space Exploration Tech Corp | — | SELL 50% | 133.11 | 114.92 | n/a | n/a | n/a | Halve the position into a supply schedule that is not finished. The 06-Aug lock-up released 911.5 million shares - roughly $116bn, more than doubling the float - and the market absorbed it well, with the stock up 6.1% that session and a further 15.8% on 07-Aug after an Argus upgrade to Buy with a target. But a second tranche of roughly 319 million shares becomes eligible on 12-Aug, with further unlocks reported around 21-Aug and 10-Sep, against a share price already about 40% below its post-IPO highs. Selling half here keeps in a business whose fundamentals nobody disputes while stepping out of the way of three more supply events in five weeks. OPERATIONAL: carried disposition from the 08-08 card, session 4 unexecuted (US-37 stalemate - record TAKEN or SKIPPED). Permanently muted on signals (private-market row, no public quote, no_close) so this is a qualitative news assessment only, not a signal read - the card is re-affirmed on NEW information (the 08-12 tranche), which is not the re-litigation US-37 warns against. Action price left at the 08-07 reference 133.11 - no Aug-10 print exists. One search snippet quoted a level irreconcilable with the book's reference; treated as unusable. Manual stop 114.92 stands on the 80-share remainder. 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. | 184.11 | HOLD | 184.11 | 183.00 | 188.77 (+2.5%) | 192.01 (+4.3%) | 219.72 (+19.3%) | Hold all 300 through Wednesday's results, then de-risk regardless of the outcome. Nebius reports Q2 before the open on 12-Aug - consensus looks for roughly of revenue and a loss per share, off a quarter in which revenue grew 684% year on year, with a $27bn five-year Meta compute deal and 4GW of guided 2026 capacity now needing to convert into delivered revenue. That is a genuinely two-sided event worth perhaps 20% either way, and selling a losing position into it the day before is the wrong trade; the stop at 183.00, barely 0.6% below the close, already defines the downside. What is not defensible is carrying the full position indefinitely, so the reduction is committed to a date rather than left to judgement. OPERATIONAL: ADR-0020 Trail Erosion DETECTED FOR THE FIRST TIME and DEFERRED one session on the CLAUDE.md section 9 catalyst-aware hold. Peak Stop 270.00 (2026-06-23) vs live 183.00 = -32.2%, and us_transactions.jsonl shows NBIS untouched at 300 since the 07-04 opening balance (the 07-17 50% was withdrawn 07-18) - NEVER size-cut, contradicting US-36's evidence line (systemic finding 1). Mitigating: the stop is in a RISING phase (135.00 on 07-30 -> -> 183.00) i.e. the risk line is being rebuilt toward the position, the opposite of the failure ADR-0020 targets. TWO PRE-COMMITMENTS, BOTH BINDING: (1) any settled close below 183.00, or a 2nd consecutive settled STRONG SELL - REDUCE, fires a mandatory 50% trim; (2) the ADR-0020 30% first fire executes UNCONDITIONALLY at the 08-12 EOD, post-print, whichever way it resolves, unless the position has been size-cut by then. Tonight AVOID cv64 (4th consecutive non-actionable print), -12.99% vs the 211.60 cost. Not armed - Gain_ATR -1.24, -12.99%, US-26 declines outright. Stop_Price prints 184.11 = Close, the US-29 artifact, REJECTED. 1.5 ATR = 150.98, no raise. · news |
| LLY | Eli Lilly & Co | 1231.94 | HOLD | 1231.94 | 1159.23 | 1238.60 (+0.5%) | 1258.56 (+2.2%) | 1320.74 (+7.2%) | Hold all 70 and raise the trailing stop - this is now the strongest holding in the book. Lilly gained 3.9% to 1231.94 on the session and trades above every major moving average, with conviction jumping from. The trim proposed on 07-Aug was withdrawn on 08-Aug and that call has already paid: selling at 1191.94 would have cost about The twelve-week view still points 7% higher, so the position is worth keeping intact with a tighter stop underneath it rather than cut. OPERATIONAL: warnings print RSI_DIV,PROFIT_ARMED but US-26 DECLINES THE ARM - no Cost in the portfolio file, so Gain_ATR and Unreal_PnL_Pct are both blank and the gain is unverifiable (systemic finding 5 - fill the Cost field). Even with a valid arm the trim-discipline gates block it: G3 (Target_12Week 1320.74 is +7.2% above the close with a bullish cv93 signal - cutting the winner) and G2 (DCF_per_share 387.04 vs a 1231.94 close makes the valuation anchor unusable, so GROWTH_PRICED_IN / MoS_FV -3.36 cannot amplify). Trail raised 1155.71 -> 1159.23 (1.5 ATR). |
| NVDA | NVIDIA Corporation | 217.55 | HOLD | 217.55 | 212.52 | 218.93 (+0.6%) | 221.32 (+1.7%) | 231.46 (+6.4%) | Hold all 100 with the stop unchanged. Nvidia gave back 2.9% to 217.55 and the signal engine went quiet, but the stock still sits above its 20-, 50- and 200-day averages and 2.4% above the 212.52 stop, with the twelve-week view pointing to 231.46. A quiet signal on a position that has not broken anything is not a reason to sell; earnings on 26-Aug are the next real test. OPERATIONAL: conviction collapsed, a textbook ADR-0012 tick (b) at >=20 points, and PROFIT_ARMED printed for a 6th session - but US-26 DECLINES THE ARM (no Cost in the portfolio file, so Gain_ATR / Unreal_PnL_Pct are blank and the gain is unverifiable), so there is no valid armed position for the tick to fire against. With the arm declined all that remains is a bare NO TRADE, which US-17 bars from originating a dispose. This is a data gap costing the book a valid profit-take - see systemic finding 5. 1.5 ATR = 206.19, below the standing stop, no raise. SHARED WITH ustech/ussemicon - execute once (L7). |
| NET | Cloudflare, Inc. | 310.59 | HOLD | 310.59 | 279.08 | 310.03 (-0.2%) | 313.54 (+0.9%) | 336.27 (+8.3%) | Hold all 150 and raise the stop, but do not add here. Cloudflare's Q2 was a clear beat and raise - revenue up 36% to, adjusted EPS of up 38%, more than 80,000 net new paying customers, and full-year guidance lifted to -2.87bn - and the shares have responded, up 9.2% over two sessions to 310.59. The business case is confirmed; the entry price is not. Buying more at the top of a two-day earnings pop, with the stock overbought and the model's own one-week estimate of 310.03 sitting slightly BELOW today's price and its four-week estimate just 1% above, is paying up for a flat expected week. Better to hold the existing stake, lift the stop to 279.08, and add on a pullback. OPERATIONAL: the ADD arithmetic CLEARS - adj = 93 - 10 (MoS_FV -943.71, an EPV artifact on a name whose FV_per_share is) - 7 (FailedBreakdown_Up_20D|buy cell) = 76, eleven over the floor, and the quality gate cannot block it because NET is held. DECLINED on L6 (day 2 of a post-earnings move, +9.2% cumulative, RSI 68.5, at-market entry with no pullback offered) plus a sub-zero Est_1W. RE-ARM TRIGGER, PRE-COMMITTED IN WRITING: a pullback to the 279.96 EMA20 +/-0.25 ATR with conviction still >=65 re-opens the ADD at the live Entry_Price. Trail 272.00 -> 279.08 (1.5 ATR); the model's 290.49 at 0.96 ATR is tighter - noted, NOT adopted (whipsaw on a post-earnings name). WARNING 11TH SESSION: two NET rows in input/usportfolio.json (150 / 0), loader takes the last, so this position reads Held=N and tonight's STRONG BUY Buy_Rank 1 was evaluated as a quality-gate-eligible NEW ENTRY rather than an ADD. DE-DUPLICATE. · news |
| MU | Micron Technology, Inc. | 861.00 | HOLD | 861.00 | 770.00 | 870.19 (+1.1%) | 956.04 (+11.0%) | 1102.69 (+28.1%) | Hold all 100 with the stop unchanged. Micron eased 1.9% to 861.00, essentially flat against the 859.49 cost, and the engine has now printed no actionable signal for nine sessions running - the stated reason being conflicting indicators rather than deterioration. The stock sits well above its 200-day average and the twelve-week view still points 28% higher, so there is nothing here that justifies acting; the 770.00 stop, 10.6% below, defines the risk. OPERATIONAL: ADR-0020 Trail Erosion CHECKED AND DOES NOT FIRE. The naive all-time peak stop of 1102.16 (07-01) is a US-29 flush artifact - a bearish-family stop printed above the price during the 07-02 rout and reset to 937.03 within two sessions - and the position was ADDED to on 07-21 @ 868.00), not cut. The defensible watermark is the highest stop set on the current 100-share position, 880.00 (07-22), against a live 770.00 = -12.5%, inside the 15% threshold. No fire. Note US-36's evidence line wrongly describes MU as already size-cut (systemic finding 1). NoTrade_Reason verbatim: Consensus conflict + low conviction. Not armed (Gain_ATR 0.02, T12_Progress 78.08). 1.5 ATR = 755.70, below the standing stop, no raise. |
| OKTA | Okta Inc. | 150.77 | HOLD | 150.77 | 141.14 | 151.23 (+0.3%) | 156.43 (+3.8%) | 167.45 (+11.1%) | Hold all 200 and raise the stop above cost. Okta added 1.7% to 150.77 and is now 7.9% ahead of the 139.80 entry, with conviction firming from and the twelve-week view at 167.45. Lifting the trailing stop to 141.14 locks the position into a profit whatever happens next, which is the right response to a winner that is working rather than trimming it. Earnings land 26-Aug. OPERATIONAL: PROFIT_ARMED prints and US-26 PASSES (real gain, +7.85%), but the arm comes via the T12_Progress 90.04 limb only - Gain_ATR 1.71 is well under 3.0 - so US-33 requires a non-(c) tick. NONE FIRES: (a) no second warning code, (b) conviction ROSE 7 points, (c) Target_12Week 167.45 = 1.111x Close so not dead in any case, (d) not a NO TRADE. Armed-but-not-fired = mandatory trail-raise, not a trim. Stop 140.00 -> 141.14 (1.5 ATR). Model Stop_Price 137.93 is LOOSER - rejected per US-29 discipline. |
| DELL | Dell Technologies Inc | 457.88 | WATCH | 457.88 | 407.00 | 463.26 (+1.2%) | 505.90 (+10.5%) | 593.69 (+29.7%) | The book cannot tell whether it owns this. A buy-stop order placed on 05-Aug may have filled at around 466.60; no Dell position with a non-zero quantity exists in the holdings file four sessions later. If it did fill, the position is now worth 457.88, down 1.9% or about - improved again from last session. Confirm it either way before anything else can be decided on the name. OPERATIONAL: ADR-0019 / US-37 disposition owed, session 4 - record TAKEN (trade buy) or SKIPPED (trade skip us DELL). Separately, as a fresh acquire DELL is BELOW FLOOR tonight: adj = 69 - 10 (MoS_FV -2.88, clamped) - 7 (FailedBreakdown_Up_20D|buy cell) = 52. Worth flagging that DELL is the ONLY Quality_Pass True / Bargain True name in the entire buy pool (F_Score 8, Quality_Rank 9, FAIR_VALUE) with DCF_per_share 485.61 against a 457.88 close, i.e. +6.1% of DCF upside, and the fusion still applies the maximum -10 valuation penalty - see systemic finding 3. Stop 407.00 if held. |
Outcome & track record
Accuracy at the time of this record.
124 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.