US — End-of-Day Verdict
ACQUIRE 1 of <=3 - DELL BUY @ ~464.24 (buy-stop, max fill 474.18), stop 404.50, adj 76, Quality_Pass TRUE; the first admissible new BUY in 15 sessions (US-20 drought broken). ANTICIPATE 0 of <=1 - no BUY - ANTICIPATE row anywhere in the anchor; VIST decayed to WATCH cv49, the US-35 trigger stands unfired. DISPOSE 3 of <=3 - KLAC SELL @ ~195.45 on ADR-0020 Trail Erosion alone, de-escalated 50% -> 30% per US-32; SPCX SELL at market on a decisive settled close through the pre-committed 135.00 line ahead of the 08-06 lock-up; NVDA SELL @ ~211.94 on two consecutive settled STRONG SELL - REDUCE prints, Sell_Rank 1. HOLDS: GOOGL, LLY, MU, NBIS, NET, OKTA + the retained SPCX 80. : buys vs sells - by ~3x; -> (22.8% of a mark).
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 |
|---|---|---|---|---|---|---|---|---|---|
| DELL | Dell Technologies Inc | 467.27 | BUY | 464.24 | 404.50 | 469.34 (+1.1%) | 502.51 (+8.2%) | 613.16 (+32.1%) | Dell has turned into one of the cleanest AI-infrastructure businesses in the market: AI-optimised server revenue grew 757% year-on-year to 1bn and management raised FY27 AI-server revenue guidance to $60bn, and the shares just broke to a new 52-week high on that demand rather than on an earnings gap. It is the first STRONG BUY in fifteen sessions to clear the quality screen (F-Score 8, ROIC 31%, 4bn free cash flow, FAIR_VALUE) while still trading at a sane RSI of 61 and sitting right at the model's entry, so the book gets a high-conviction breakout without chasing it. Breaks the 15-session US-20 buy drought; the gate passes here because Dell is cash-generative hardware the F-Score/FCF screen can actually score, unlike the growth and financial names it has been rejecting. Sized 0.75x on MoS_FV -2.59: Portion 85.49% x x 20% buy_pct = x 0.75 = / 464.24 =. Stop set at 404.50 (~1.5 ATR) NOT the model's 340.54 (-26.6%, risk); 404.50 risks = 1.64% of the base. HARD NO-CHASE GUARD (US-9/L6): MAX FILL 474.18 = top of the entry range; if the 08-05 open gaps above that, do not fill - re-queue at the live Entry_Price. Est_12W 613.16 exceeds the 485.61 DCF anchor - projection assumes growth beyond priced-in. L6 clear: last earnings May 2026, next print late-August, so this is not a post-earnings gap. Noted against: +234% YTD and UBS argues limited further upside - which is why it is 0.75x, not full. · news |
| KLAC | KLA Corporation | 195.45 | SELL 30% | 195.45 | 175.00 | 195.90 (+0.2%) | 201.47 (+3.1%) | 207.00 (+5.9%) | KLA is a quarter of the whole portfolio and its protective stop has been walked down from to - a 38% retreat of the risk line - without a single share ever being sold, so this trim is about restoring risk discipline rather than a view on the company. The shares are still 36% below their 52-week high and below their 20-, 50- and 100-day averages despite Tuesday's 7% bounce, KLA's own record quarter was sold down 10% on the print, and the stock trades at 49.6x forward earnings against a 39.6x industry average. Cutting banks about and takes the position from 27.1% of the book to 19.0%. ADR-0020 TRAIL EROSION is the ONLY basis tonight and this is still the FIRST fire on the 272.00 Peak Stop (the 08-04 fire was never executed), so it is EXPLICITLY IMMUNE to the held-name sell vetoes US-17, US-18/34, US-22 and US-30. DE-ESCALATED 50% -> 30% per US-32: the signal leg dissolved (3x STRONG SELL - REDUCE -> NO TRADE cv0 on a +6.95% session) and US-17 bars a bare NO TRADE from originating, leaving only the ADR-0020 mandatory 30% floor. ADR-0012 is armed and US-26-verified (Gain_ATR 1.44, +11.21%, T12 94.42) but NOT fired - tick (a) absent, tick (b) declined as a label change not a conviction decay, tick (c) alive (Est_12W 207.00 vs 199.36 threshold), tick (d) only 1 NO TRADE print - so it contributes nothing. Valuation amplify SUPPRESSED by G2: Est_12W 207.00 is 2.07x the 99.91 DCF anchor. Est_12W exceeds DCF anchor. Trail on the retained 700 rises 168.00 -> 175.00 (~1.5 ATR); this does NOT discharge the fire (erosion still -35.7%) - only the size cut does - and the watermark re-bases to 175.00 on execution. New adverse read-through: AMD beat Q2 and guided strongly on 08-04 AMC and fell 8% after-hours, landing on this trim's execution session. US-37 STALEMATE session 5 - record TAKEN or SKIPPED explicitly. · news |
| SPCX | Space Exploration Tech Corp | — | SELL 50% | 124.46 | 135.00 | n/a | n/a | n/a | SpaceX's first-ever quarterly report beat handsomely - 8bn revenue against 81bn expected, 5bn adjusted EBITDA against 0bn, Starlink at 12 million subscribers - yet the shares closed below our risk line and the market's focus is squarely on cash burn: 4bn of quarterly capex and a 26bn operating loss in the AI segment. On Thursday roughly 911.5 million shares, about triple the current tradable float, become free to trade as the first lock-up tranche expires. Halving the position ahead of that supply event respects the broken risk line while keeping half the shares for a business whose top line is not in question. THIS CARD IS WRITTEN ON NEWS-SOURCED PRICES, NOT MODEL OUTPUT - SPCX has no row in any scan snapshot, so there are no Est_* values, no conviction, no pillars and no targets; none were synthesized. Basis: a decisive SETTLED regular-session close (~124.46, +8.67% on the day, into the AMC print) through the pre-committed 135.00 manual stop, post-maiden-print, which under US-27 OPENS the exit case; L4 is satisfied because this is a settled close, not an intraday or after-hours print. Sized 50% rather than the mute rule's nominal full exit on three grounds: the price is unverifiable by the platform, US-27 as narrowed says a broken line opens a case rather than mandating a full exit, and the print was a beat with the selling driven by AI capex and the unlock. PRE-COMMITTED ESCALATION: if SPCX ever appears in the scan CSV with a non-null Close below 135.00 it auto-unmutes and the remaining 80 become a FULL-EXIT candidate; independently a settled close below 110.00 after the 08-06 unlock fires the remainder. Manual stop on the retained 80 stays 135.00, deliberately above the market. Counterweight on the record: Morgan Stanley says fundamentals are largely unchanged, and on the call Musk committed SpaceX to building its AI exclusively on Nvidia (Starmind AI-1, Vera Rubin NVL72) - a stronger read-through for NVDA than for SPCX. · news |
| NVDA | NVIDIA Corporation | 211.94 | SELL 30% | 211.94 | 200.00 | 212.52 (+0.3%) | 210.95 (-0.5%) | 217.45 (+2.6%) | Nvidia has now printed a sell signal on two consecutive closes and ranks first in the book's sell queue, on the signal family with the best track record here, while the model's own forward view is flat at every horizon - roughly break-even at one week, slightly negative at four, and up only 2.6% at twelve. Trimming 30 of banks about near the highs and leaves - the clear majority - exposed to the upside, which matters because Elon Musk committed SpaceX on Tuesday's call to building its AI exclusively on Nvidia systems. Stable REDUCE, clearing US-18/34 on its own terms: A03C cv76 and A04C cv69 are two consecutive settled STRONG SELL - REDUCE prints, so unlike the 08-04 card no override of the lone-flip bar is needed. Sell_Rank 1; FailedBreakout_Down_20D|sell +3 (n=15, hit_1w.636). ADR-0012 remains BLOCKED by US-26 - no Cost in usportfolio.json so Gain_ATR/Unreal_PnL_Pct are null - and US-33 retires the T12-arm + tick-(c) path; the card rests on the signal alone. Stated against it: pillars contradict the label (Mostly_Aligned, Trend 72, Momentum 67), close above all four EMAs, conviction decaying while price rose; the discriminator versus GOOGL, applied consistently across both nights, is the forward view. News NEUTRAL (mixed) not an override: SpaceX/Nvidia exclusivity bullish, AMD -8% after-hours on a beat bearish; the Nvidia news broke AFTER the anchor close so it changes the 08-05 open, not the anchor print. GAP RULE (US-9/L8): if the 08-05 open gaps above ~215.1 (>1.5%), REPRICE the trim to the live open - trimming into strength is better, not worse - do NOT withdraw. Est_12W exceeds the 163.98 DCF anchor. Trail on the retained 70 rises 194.00 -> 200.00 (~1.49 ATR, just under EMA20/50), closing the Trail Erosion gap to -8.35%. SHARED WITH ussemicon - EXECUTE ONCE (L7). US-37 STALEMATE session 16 - record TAKEN or SKIPPED explicitly. · news |
| GOOGL | Alphabet, Inc.- Class A | 377.65 | HOLD | 377.65 | 356.00 | 382.64 (+1.3%) | 384.53 (+1.8%) | 409.36 (+8.4%) | Alphabet flipped back to a bullish signal at conviction 77 and closed nearly 10% clear of its stop, so the trim case that was building last week has evaporated and the position is left alone with a tighter trailing stop. The holding is up more than 200% on cost and the model still projects roughly 8% upside over twelve weeks. The 08-04 pre-commitment (a THIRD consecutive STRONG SELL - REDUCE, or a close below 343.50) DID NOT FIRE and is discharged - label went cv67 SELL -> HOLD cv77 AnticipationUp_VCP; yesterday's demotion vindicated in one session. ADR-0012 armed and US-26-verified (Gain_ATR 17.48, +207.03%, T12 92.25) with ZERO ticks: (a) no second code, (b) conviction ROSE, (c) Est_12W 409.36 vs the 385.20 threshold alive, (d) no NO-TRADE run - so armed-but-not-fired resolves to a MANDATORY trail-raise. Stop 343.50 -> 356.00 (~1.49 ATR below close, just above EMA50 353.38/EMA20 350.43), which also closes the Trail Erosion gap from -4.05% to -0.56%. Est_12W exceeds the 145.29 DCF anchor. Overlay '<345' is healed and 31 days stale (US-31, no originating force). RE-ARM restated: two consecutive settled STRONG SELL - REDUCE prints, or any settled close below 356.00, fires the 30% trim without further debate. |
| LLY | Eli Lilly & Co | 1115.68 | HOLD | 1115.68 | 1155.71⚠ | 1123.66 (+0.7%) | 1111.71 (-0.4%) | 1170.95 (+5.0%) | Lilly reports second-quarter results before the open on Wednesday, so the position is held through its own earnings rather than sold into them - the long-standing rule here is to trim only on a confirmed breakdown after a catalyst, never ahead of one. The shares sit about 3.5% below the trailing stop for a fourth session and the oral GLP-1 launch is the swing factor: Foundayo prescriptions have stalled for five straight weeks and Goldman has cut its 2026 US estimate for the drug from 1bn to, though it still models 8bn of quarterly revenue. The twelve-week view remains positive, and the decision point is one session away. THE LEAST COMFORTABLE CALL IN THE BOOK, recorded as such: 4th consecutive breach at -3.46%, and the close 1115.68 has now slipped BELOW the pre-committed 1120.00 action line. Held on four counts: (i) the pre-commitment is explicitly conditioned on a close below 1120.00 AFTER the print, and Q2 is 08-05 BMO, still ahead of the anchor; (ii) L2 needs both legs and Est_12W 1170.95 (+4.95%) is alive vs the 1138.00 threshold; (iii) label is NO TRADE cv0 and US-17 bars a bare NO TRADE from originating; (iv) PROFIT_ARMED is unverifiable - no Cost in the portfolio file, Gain_ATR/ null - so US-26 declines the arm by rule. Lost EMA20 1165.82; EMA50 1136.62 is the line that matters into the print. Novo Nordisk reports into the same session so incretin share gets marked directly. PRE-COMMITMENT HARDENED AND NOW SIZED: any settled close below 1120.00 AFTER the 08-05 print fires a 30% trim without further debate; if the 12wk estimate also rolls negative on that same close, the tier escalates to 50%. A written line breached for four sessions must not be allowed to rot (US-37). · news |
| NBIS | Nebius Group, N.V. | 225.74 | HOLD | 225.74 | 183.00 | 224.32 (-0.6%) | 252.28 (+11.8%) | 309.54 (+37.1%) | Nebius added another 6% and is now above cost, holding every major moving average with second-quarter results due on 12 August, so the position is held with the stop lifted sharply to lock in the recovery. The label reads AVOID rather than a reduce instruction, which on a long-only book means do-not-add, not sell. Yesterday's declined lone STRONG SELL - REDUCE is vindicated: label decayed cv69 SELL -> AVOID cv64 and the name added +6.19% to 225.74, now +6.68% over the 211.60 cost. Not in the sell pool (AVOID is not a REDUCE). Pillars Consolidation_Bullish, Trend 72, above EMA20 202.03/EMA50 205.89/EMA100 187.85/EMA200 155.47, RSI 55.3. Stop_Price prints 225.74 = the close EXACTLY - the US-29 artifact again, rejected. Not armed (T12_Progress 72.93, Gain_ATR 0.50). 12wk band [135.26-643.61] is 4.8x wide and near-uninformative - weight the 1wk. Trail 165.00 -> 183.00 (~1.5 ATR below close on a 28.38 ATR, well above EMA200), which REPAIRS the re-based Trail Erosion gap from -17.5% to -8.5% and takes NBIS off the ADR-0020 watch entirely. CRWV reports 08-11 as the closest comparable read-through; NBIS Q2 08-12. |
| NET | Cloudflare, Inc. | 301.33 | HOLD | 301.33 | 272.00 | 302.61 (+0.4%) | 327.94 (+8.8%) | 335.67 (+11.4%) | Cloudflare reports second-quarter results on Thursday after the close, with consensus at roughly of revenue (up 30% year-on-year) and 27 cents of earnings, and the company has beaten in each of the last four quarters by an average of 12%. The position is 17% ahead of cost at a new 52-week high, so the stop is raised to - above the 20-day average and comfortably above cost - and the shares are held through the print rather than pre-trimmed into it. CATALYST-AWARE HOLD: Q2 is 08-06 AMC (call 5pm EDT), DDOG reports 08-06 BMO and sets the cloud-spend tone hours before; no pre-trim, trim only on a confirmed lower-high or a decisive break. AVOID is a do-not-open label on a long-only book, NOT a sell instruction, and it is not a REDUCE - NET is not in the sell pool. WARNING - THE SCAN STILL CANNOT SEE THIS POSITION, 5th consecutive session: input/usportfolio.json carries TWO NET rows ( Cost 258.0, and twenty-three entries later) and the loader takes the last, so the anchor prints Held=N with Avg_Cost/Gain_ATR/T12_Progress all null - no Exit_Warning, no ADR-0012 arming, no stop monitoring, no Trail Erosion watermark, on a position at +16.80% going into its own earnings. MUST be de-duplicated before the next run. RSI 71.2 at a new 52W high 303.75, pillars Consolidation_Bullish Trend 89, above all four EMAs. Stop raised 260.00 -> 272.00 (just above EMA20 271.11, ~1.70 ATR below close, above the 258.00 cost) so the position is now stopped into a profit. The model's Stop_Price prints 323.51 - ABOVE the close, the US-29 short's-stop artifact on a held long - and is rejected. · news |
| MU | Micron Technology, Inc. | 892.67 | HOLD | 892.67 | 770.00 | 895.62 (+0.3%) | 989.38 (+10.8%) | 1131.45 (+26.7%) | Micron jumped 7.6% and is above cost for the first time in weeks, sitting 16% clear of its stop, so the position simply rides with the risk line unchanged. The engine still declines to take a view - a fifth straight no-trade print - which on its own is never a reason to sell. 5th consecutive NO TRADE cv0 (US-17 bars a bare NO TRADE from originating). Above EMA100 776.18/EMA200 600.41 but still below EMA20 899.29/EMA50 890.23, so the reclaim is incomplete. Not armed (T12_Progress 78.90, Gain_ATR 0.37). No trail-raise: 1.5 ATR below the close is 757.5, BELOW the standing 770, and a risk line is never loosened. Re-based Trail Erosion -18.75% stays live but VETOED (US-17 on the label, US-30 on price now +15.9% above the line and above cost); if MU ever prints an actual sell signal, that eroded watermark AMPLIFIES it. 12wk band [430.18-1886.54] is 4.4x wide - uninformative. AMAT FQ3 08-13 and the AMD -8% after-hours reaction carry the HBM/WFE read-throughs. |
| OKTA | Okta Inc. | 147.82 | HOLD | 147.82 | 140.00 | 148.45 (+0.4%) | 152.73 (+3.3%) | 167.48 (+13.3%) | Okta added 4.4% and is now nearly 6% ahead of cost with every major moving average stacked beneath it, so the stop moves up to - just above the entry price - and the position is left to run. Conviction eased slightly for the first time in this run, which is worth watching but not acting on. 5th consecutive FailedBreakdown_Up_20D print, but conviction DECLINED for the first time in the run -> - noted, not actionable. Pillars Consolidation_Bullish, Trend 69, above EMA20 139.96 with EMA50 128.30/EMA100 113.88/EMA200 102.95 stacked far below, RSI 62.3, 52W high 157.00. Not armed (T12_Progress 88.26, Gain_ATR 1.22). Stop 135.00 -> 140.00 (~1.19 ATR below close, at EMA20, above the 139.80 cost) so OKTA is now stopped into a profit; the model's Stop_Price 140.37 agrees for once. |
Outcome & track record
Accuracy at the time of this record.
114 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.
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.