US — End-of-Day Verdict
1 ACQUIRE (QUEUED AAPL @316.32, adj 93) | DISPOSE fired: KLAC -300 (30%), META exit 60, GOOGL exit 100 | DEFERRED by cap: NVDA -30, MU -30 | HOLD NBIS through Q2 (marginal break, +58% fwd), LLY trail 1134 | BOOK UNRECONCILED - 07-25 cards unexecuted 5th session
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 |
|---|---|---|---|---|---|---|---|---|---|
| AAPL | Apple | 336.91 | BUY — QUEUED @316.32 | 316.32 | 306.90 | 339.30 (+7.3%) | 339.98 (+7.5%) | 347.89 (+10.0%) | Apple is the single strongest buy setup the book has produced this month: a STRONG BUY at maximum conviction (100) on PullbackBuy_VWAP -- the only signal family with a proven positive record here -- and the only candidate that also clears the fundamental quality screen (F-Score 9, best in the pool). Apple retook the world's-most-valuable-company spot from Nvidia on Monday's chip selloff, which is exactly the defensive-megacap rotation this setup is built to catch. The catch is price: the shares closed at 336.91, about 2.5 ATR above the model's 316.32 entry, so buying here would be chasing. We queue a good-till-cancelled limit at the model entry instead and let the pullback come to us. Valuation is the honest offset -- at a 3.2x negative margin of safety it is priced for growth, which halves the position size. OPERATIONAL: repriced from the prior 308.38 queue to the live 316.32 entry (provenance rule / US-9); size 0.5x on MoS <= -.8549 x x 20% x 0.5 = ->; adj conviction 100 -10 (val) +3 (PullbackBuy cell) = 93; expires 2026-07-30 (anchor + 3 sessions). CAVEAT: the model's own 1wk-low estimate (319.71) sits ABOVE the 316.32 limit, so a ~6.1% retrace is needed to fill -- the no_chase_queue counterfactual (n=13, +2.1% cost) says these limits mostly do not fill. Re-verify the PullbackBuy family + quality gate at any fill (US-8). 12wk 347.89 exceeds the DCF anchor 208.95. |
| KLAC | KLA Corporation | 203.36 | SELL 30% | 203.36 | 198.40 | 206.23 (+1.4%) | 206.52 (+1.6%) | 214.52 (+5.5%) | Take 30% off KLA into its earnings print. This is the book's largest position by a wide margin (, about 29% of the book) and the setup deteriorated decisively on Monday: after three sessions of the engine printing nothing, the signal flipped outright bearish (RallySell_BearTrend) and the shares broke the 210 trailing stop by a decisive 3.2%, closing at 203.36 and losing the 50-day average. KLA reports fiscal Q4 after Tuesday's close, guiding to 575bn revenue against a 61bn consensus, with elevated DRAM input costs already flagged as a ~100bp gross-margin headwind. Carrying a concentrated, technically broken position into a binary print is the risk we are paid to avoid -- so we bank roughly a third at a +15.7% gain and keep 70% for the AI-capex trend, which the model still projects +5.5% higher over twelve weeks. OPERATIONAL: @ ~203.36, keep 700 on a reset stop of 198.40 (EMA100). Basis is signal degradation + decisive trail break (CLAUDE.md 9), NOT valuation -- the MoS -5.44 amplify that would have bumped this to 50% is SUPPRESSED by gate G2 (DCF anchor 93.26 unreliable vs a 33% revenue grower). Honours the prior card's pre-commitment ('a decisive <210 settle -> trim 30% next EOD'). The catalyst-aware hold rule permits this because a decisive stop break is its stated exception. Tension noted: the optional_trim counterfactual (n=12, median -10.85% at 4wk) says trimming winners has cost this book; the hard trail break plus the bearish flip overrides. An earnings gap through 198.40 triggers the next tranche at the NEXT settled close, not intraday (L4/US-7). 12wk 214.52 exceeds the DCF anchor 93.26. · news |
| META | Meta Platforms | 593.87 | SELL ALL | 593.87 | 620.00⚠ | 592.50 (-0.2%) | 571.38 (-3.8%) | 592.91 (-0.2%) | Close the Meta position entirely. It is 6.5% underwater on a 635 cost, has been below its 620 stop for four straight sessions (593.87, -4.2%), and now trades under its 50-, 100- and 200-day averages -- broken on both price and structure. The model sees no recovery to wait for: flat at one week (-0.2%), down 3.8% at four weeks, flat at twelve. Meta reports Q2 on Wednesday after the close and the swing factor is capex, which management already raised to -145bn from -135bn. In a tape that has just punished Alphabet 6% for exactly the same disclosure, carrying a losing position with no modelled upside into that print is an uncompensated bet. Exit and redeploy. OPERATIONAL: @ ~593.87, largest loss-prevention dollar in the book. Basis is L2 (decisive stop break AND negative 12wk), not the bare NO TRADE label -- US-17 bars label-only sells, and the NoTrade|sell cell carries a -3 bias which we absorb. PROFIT_ARMED on this row is the known loss-artifact (Gain_ATR -1.83) and is ignored. 4th re-fire of an exit first carded 07-22; the book has not executed it (L3). · news |
| GOOGL | Alphabet, Inc.- Class A | 326.56 | SELL ALL | 326.56 | 345.00⚠ | 335.18 (+2.6%) | 337.18 (+3.3%) | 375.23 (+14.9%) | Exit Alphabet in full on the pre-committed stop. We drew a hard line at 345 on 22-Jul after the Q2 capex shock -- 2026 capital spending lifted to -205bn, free cash flow turning negative, buyback suspended -- and the shares have now spent six sessions below it, closing at 326.56, still 5.3% under the line. Monday's +2.1% bounce is the second up-day of a recovery attempt but does not reclaim the level, and a pre-committed stop is only worth having if it is honoured. We are candid about the other side: the model projects +14.9% over twelve weeks and Bank of America has reiterated Buy with a target, arguing the market is mispricing what the spending is already producing. That is a real bull case -- but it is a re-entry thesis, not a reason to keep sitting under a broken line. OPERATIONAL: @ ~326.56. Basis = triggered overlay invalidation ('close below 345') with news-override force + name-specific breakdown -> US-12 gives idiosyncratic cracks a FULL exit, not a trim. L2 does NOT fire (12wk is +14.9%, positive) -- US-23 governs: on a fresh breakdown the forward lags, the signal/stop leads. 5th re-fire of an unexecuted exit (L3). US-22 self-heal: if GOOGL settles back ABOVE 345 before execution, this exit auto-withdraws. 12wk 375.23 exceeds the DCF anchor 154.23. · news |
| MU | Micron Technology, Inc. | 900.20 | NO ACTION | 900.20 | 880.00 | 900.30 (+0.0%) | 956.27 (+6.2%) | 971.08 (+7.9%) | Micron mechanically qualifies for a 30% profit-take and is held back only by the three-slot cap -- this is a deferral, not an optional idea. The position is armed (93% of its 12-week target, +4.7% on cost) and has printed NO TRADE two sessions running, the ADR-0012 fire tick (d). Against that: it held its 880 trailing stop (closed 900.20, -2.3% on the broad chip selloff) and the forward view is very much alive -- +6.2% at four weeks, +7.9% at twelve -- so the prior card's pre-commitment, which required a second NO TRADE AND a rolling-over forward, is only half met. Hold the on the 880 stop through SK Hynix's HBM read-across and KLA's print. OPERATIONAL: deferred trim would be @ ~900.20. Ranked 4th of 5 dispose candidates on dollar terms behind KLAC/META/GOOGL. Ledgered direction=none/optional=true so the counterfactual grader prices what not acting cost. Trail stays 880 (never ratchets down). Re-fires as a full DISPOSE card at the next EOD if a third NO TRADE prints or 880 breaks. |
| NVDA | NVIDIA Corporation | 196.51 | NO ACTION | 196.51 | 201.00⚠ | 195.67 (-0.4%) | 196.71 (+0.1%) | 197.88 (+0.7%) | Nvidia's 30% profit-take has hardened and is deferred only by the three-slot cap -- it is card #1 at the next EOD. The shares fell nearly 5% to 196.51 on renewed circular-financing worries after reports Nvidia would backstop $250bn of OpenAI, breaking the 201 trailing stop by 2.2% and losing both the 50- and 100-day averages. The position sits at 99% of its 12-week target with a completely dead forward view -- +0.2% at one week, +0.1% at four, +0.7% at twelve -- so the last third is carrying risk for no modelled return. OPERATIONAL: deferred trim would be @ ~196.51, the smallest dollar of the five candidates, hence last in the ranking. Armed (T12 99.3%) + soft ticks (c) dead forward and (d) 3 consecutive NO TRADE, PLUS a hard tick (stop break) which under US-19 is immediately executable. Held at 30%, NOT escalated to 50%, per US-19 (no day-by-day escalation on the same name). Stop stays 201.00 -- BROKEN, deliberately not lowered. Shared with ussemicon -> execute ONCE (L7). Same fire first carded 07-16/07-23/07-24/07-25, never executed. · news |
| NBIS | Nebius Group, N.V. | 187.88 | HOLD | 187.88 | 190.00⚠ | 189.97 (+1.1%) | 204.29 (+8.7%) | 297.04 (+58.1%) | Hold all 300 Nebius shares on the 190 stop -- do not trim into the print. The shares stabilised at 187.88 after last week's 15% crash, leaving them just 1.1% below the stop, a marginal rather than decisive break. More importantly the bear catalyst has weakened: the dilution fear from the 75bn convertible has been partly answered by a GPU-backed secured facility that funds the buildout without issuing, and the Meta contract at the centre of the thesis is intact -- $12bn of compute from 2027, inside more than $40bn of contracted revenue from investment-grade customers including Microsoft. The model still projects +58% over twelve weeks, the strongest forward in the book. OPERATIONAL: L2 does NOT fire (needs stop break AND negative 12wk; 12wk is +58.1%). The standing overlay invalidation ('Meta deal materially at risk / builds own cloud') is NOT triggered. Q2 results expected 07-29 (date unconfirmed) with a +/-15% band -- the catalyst-aware hold rule bars pre-trimming into it. Keep 190.00, do NOT widen. Trigger: a SECOND decisive settle below 190 or a confirmed Meta-deal impairment -> full exit next EOD. Note the forward is compressing (12wk, 4wk -- watch it. · news |
| LLY | Eli Lilly & Co | 1197.53 | HOLD | 1197.53 | 1133.58 | 1205.12 (+0.6%) | 1187.44 (-0.8%) | 1236.80 (+3.3%) | Hold 70 Eli Lilly and raise the trailing stop from 1126..58. Lilly was the book's one green name again, closing +0.1% at 1197.53 while the AI complex sold off -- the defensive-rotation hedge doing its job. The position is armed at 97% of its 12-week target but nothing has deteriorated: conviction eased (well short of the 20-point collapse that would trigger a trim) and the forward remains alive at +3.3%. Armed-but-not-fired means tighten the stop, not sell. OPERATIONAL: new stop 1133.58 = the model Stop_Price. Q2 earnings 08-05 (BMO); pharma-onshoring tariff decision 07-31 (+/-3%). |
| OKTA | Okta Inc. | 137.43 | HOLD | 137.43 | 124.00 | 138.16 (+0.5%) | 136.47 (-0.7%) | 141.54 (+3.0%) | Hold 200 Okta on the 124 stop. The 22-Jul entry at 139.80 is 1.7% underwater at 137.43, but the name has gone quiet rather than bad -- it is back on PullbackBuy_VWAP, the one signal family with a positive record here, conviction firmed, and it passes the quality screen (F-Score 7). The forward is modest (+3.0% at twelve weeks) but positive, and it is one of only two non-AI positions left. OPERATIONAL: PROFIT_ARMED is again the known loss-artifact (Gain_ATR -0.35 on an underwater position) -- declined. The model Stop_Price has risen to 130.45; NOT adopted, because a stop 5% under an underwater position with a +3% forward invites a whipsaw exit. Keep 124.00. |
| SPCX | Space Exploration Tech Corp (muted) | — | HOLD | — | 135.00 | n/a | n/a | n/a | Hold 160 SpaceX on the manual 135 stop; the name is private/unquoted, absent from the scan, so it is muted on signals and assessed on news only. The read stays CONCERN and the calendar is now firm: maiden Q2 results on 04-Aug after the close, then roughly 911.5m insider shares -- about $116bn of stock -- unlock two trading days later on 06-Aug. That is a genuine supply overhang into a first-ever public print. One mitigant: because the shares have traded below the contingent second block of 455.8m shares stays locked, and Morgan Stanley argues fundamentals are largely unchanged and the level an attractive entry; Musk's own stake is locked until mid-2027. OPERATIONAL: auto-unmute and resume full signal treatment the moment a scan Close prints; a Close < 135 is the full-exit trigger. · news |
Outcome & track record
Accuracy at the time of this record.
96 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 →
-0.132 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.
23 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.