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 |
|---|---|---|---|---|---|---|---|---|---|
| KLAC | KLA Corporation | 181.57 | SELL 30% | 181.57 | 184.00⚠ | 182.92 (+0.7%) | 180.15 (-0.8%) | 175.78 (-3.2%) | KLA is a high-quality process-control franchise whose stock has fallen 14% in a month even as management RAISED its outlook - wafer-fab equipment to the low $150bn range for 2026, ~20% second-half growth, a 5bn backlog and advanced packaging guided to 1bn (+70% YoY). We are not selling the business; we are enforcing risk discipline. The protective stop on this position has walked down from to - a 32% retreat over nineteen sessions - while the position size was never once reduced, and the 12-week forward view now sits 3% BELOW the current price. Taking 30% off restores the risk balance and keeps in a thesis that is intact. OPERATIONAL: ADR-0020 Trail Erosion, first fire on the 272.00 watermark, VETO-IMMUNE to US-17/US-18/US-34/US-22/US-30. US-37 stalemate - identical card written 07-28, 07-29, 08-01, 08-20, 08-21, none filled; not executing since 08-21 cost Held at the 30% ADR-0020 floor: MoS -4.24 would bump but G2 SUPPRESSES (Est_12W 175.78 / DCF 61.80 = 2.84x). ADR-0012 adds no leg - armed via T12_Progress 103.29 so US-33 blocks tick (c); conviction rose (no b); AVOID not NO TRADE (no d); no second Exit_Warning (no a). 4th consecutive settled close below 184.00; below EMA10/20/50. Retained, trail 184.00, watermark re-bases to 184.00 ONLY on the observed fill. Realised approx +vs 175.75 cost. · news |
| OKTA | Okta Inc. | 130.97 | SELL 30% | 130.97 | 140.55⚠ | 131.65 (+0.5%) | 137.86 (+5.3%) | 146.77 (+12.1%) | Okta has broken its risk line and stayed broken. The shares have now closed below our protective level for four straight sessions, are 6.8% beneath it, and finished today in the bottom tenth of the day's range - below every meaningful moving average. Earnings land in two sessions with a plausible 10% move in either direction. Nothing in the business has cracked - guidance is for 9% revenue growth with a 26% operating margin and the full year was raised - so this is a risk cut on a broken line ahead of a binary event, not a verdict on the company. We keep. OPERATIONAL: basis is the invalidation-limb break, NOT the label (US-17 satisfied by the line). 08-20 pre-committed trigger fired 08-21, card unexecuted (US-37 stalemate session 2, cost). US-27 re-verify at each settled close - NO reclaim, a further -3.09% tonight. Valuation tier-bump is live (G2 does NOT suppress at Est_12W 146.77 / DCF 107.59 = 1.36x) but DECLINED on three counts: US-19 execute the first fire at the committed tier; forward alive (Est_4W +5.3%, Est_12W +12.1%) so L2's negative-12wk limb is unmet; the 08-21 upgrade tape (Citizens Market Outperform Wells Fargo Overweight) is unrepudiated. Earnings 2026-08-26 AMC, +-10% band, Event Horizon stance tighten-trail; CLAUDE.md catalyst-aware hold permits a trim on a decisive stop-loss break, which is what fired. Retained, trail 140.55. Realised approx -vs 139.80 cost. · news |
| NVDA | NVIDIA Corporation | 208.48 | SELL 30% | 208.48 | 217.07⚠ | 209.37 (+0.4%) | 210.74 (+1.1%) | 220.54 (+5.8%) | Nvidia has closed below our protective level three sessions running and is now 4% beneath it, two days before the single largest event in this portfolio - Q2 results on 26 August, where the plausible move is more than three times the stock's normal daily range and where the read-through touches four other holdings. The news into the print is genuinely good: guidance of $91bn against consensus near $92bn, earnings expected to more than double year over year, gross margins held around 75%. We are not calling the print. We are taking a third off a position whose risk line has been broken for three sessions, and letting the remaining carry the event. OPERATIONAL: 08-21 pre-committed TIGHTENED trigger, BOTH limbs fired - 3rd consecutive settled close below 217.07 AND a close below the 214.50 low; 'catalyst notwithstanding' as written (US-35). Invalidation limb satisfies US-18/US-34, not a lone flip. This is NOT a profit-take: armed via T12_Progress 94.53 but US-26 UNVERIFIABLE (Avg_Cost blank) and ZERO ADR-0012 ticks fire - conviction rose (no b), Est_12W +5.79% alive and US-33 blocks the T12 limb (no c), label HOLD not NO TRADE (no d), no second Exit_Warning (no a). MoS -2.28 is above the -3 threshold so no tier bump. Cell FailedBreakdown_Up_20D|sell +2 (n=11,.60/.625). Displaced at slot 4 on 08-21 on dollar rank; takes a slot tonight on merit. Retained, trail 217.07 held (model prints 197.00, a loosening, refused under US-29). Realised n/a - Avg_Cost blank, FILL IT. · news |
| VIST | Vista Energy, S.A.B. de C.V. | 70.66 | BUY | 70.66 | 68.38 | 70.96 (+0.4%) | 72.32 (+2.3%) | 76.53 (+8.3%) | Vista Energy is the only name in this universe that is both cheap and financially clean - it is the sole holding-eligible stock trading below our estimate of intrinsic value, it passes the quality screen at rank with a Piotroski score of 7, and revenue is growing over 100%. The technical trigger fired today at exactly the model entry price, so we are buying at the level rather than chasing an extension. Peter Thiel's fund has just disclosed a roughly 1% stake worth about - now its second-largest position - and the company is buying back its own shares; the consensus target of sits 26% above today's close. The counterweight is real and is why this is a partial rather than a full position: the last earnings report was a miss and the sell-side is openly debating the story. Cheap, clean, at the entry, with a marquee holder and an imperfect print. OPERATIONAL: 08-21 re-arm trigger FIRED - live Entry_Price + Quality_Pass True + adj >= 65, all three met, third arm cycle (08-19 @69.74 and 08-20 @70.50 both withdrew unfilled). Trail NO TRADE cv0 -> NO TRADE cv0 -> WATCH cv54 -> BUY cv76, late-emerging on the settled close (L4/US-7/US-11); SINGLE buy print, stability confidence LOW. Chase-family bar cleared by one point: FailedBreakdown_Up_20D is a -5 cell needing raw >= 75 under US-10/US-16, prints 76 - the honest weak point of this card. Adj 76 + 1 - 5 = 72. NO CHASE: Entry_Distance_ATR -0.0, close IS the entry; RSI 56.44; above EMA10/20/50/200 and VWAP. Not a queue - sidesteps US-21's unfilled-limit problem entirely. Sizing 0.4924 x 200k x 20% = x 1.25 (UNDERVALUED + Quality_Pass) = -> x 70.66 =; inside the 60% cap and the. Entry range 69.90-71.42 (70.66 +- 0.25 x ATR 3.03). R:R approx 2.66. NO DCF warning - Est_12W 76.53 sits BELOW the 86.44 DCF anchor, the only card tonight. · news |
| GOOGL | Alphabet, Inc.- Class A | 348.06 | HOLD | 348.06 | 333.68 | 350.80 (+0.8%) | 348.48 (+0.1%) | 360.32 (+3.5%) | Alphabet printed the day's only strong sell signal and we are declining it. This is a single flip out of three quiet sessions, the price is nowhere near our protective level, conviction rose rather than collapsed, and the 12-week view is still positive. It was also the one holding that rose today while the rest of the book fell 2.2%. We raise the trailing stop instead. OPERATIONAL: 88.52% exit DECLINED under the US-18/US-34 lone-flip bar - trail A21C NO TRADE cv0 -> AVOID cv61 x2 -> STRONG SELL cv75 Sell_Rank 1; close 348.06 NOT below the 332.73 invalidation; Est_1W +0.79%, Est_4W +0.12%, Est_12W +3.52% all alive. US-34's cited precedent is this same ticker on 08-01 (lone cv73, declined, closed +6.73%). SEPARATELY the 30% profit-take carried since session 8 is WITHDRAWN under US-32, not carried stale: it fired on tick (d), 8 consecutive NO TRADE, and that streak is BROKEN. Re-checked from scratch - arm VERIFIED per US-26 (Gain_ATR +31.30 AND Unreal_PnL +182.98%, both positive) but ZERO ticks fire: conviction rose (no b), Est_12W 360.32 > 1.02 x 348.06 = 355.02 (no c), NO TRADE streak broken (no d), PROFIT_ARMED is the only warning (no a). Armed but not fired -> MANDATORY TRAIL-RAISE. Model prints 333.68, ABOVE the standing 332.73 - US-29 permits raises, taken. Not executing the 08-21 card saved PRE-COMMITTED TRIGGER (US-35): a SECOND consecutive settled STRONG SELL - REDUCE, OR any settled close below 333.68, fires SELL (30%) immediately. |
| LLY | Eli Lilly & Co | 1246.93 | HOLD | 1246.93 | 1168.46 | 1260.54 (+1.1%) | 1277.27 (+2.4%) | 1379.89 (+10.7%) | Lilly is holding up well - down under 1% on a broadly red day, above every rising moving average, with a 12-week view still about 10% higher. A profit-taking rule technically triggered tonight, but only because the signal engine went quiet and prints a zero score when it does; nothing about the position actually deteriorated. We hold and keep the trailing stop where it is. OPERATIONAL: PROFIT_ARMED via T12_Progress 90.36. Tick (b) mechanically fires - conviction vs the prior EOD, far past the 20-pt bar - and is DECLINED as a US-29 structural artifact: the collapse is ENTIRELY a label transition into NoTrade cv0 ('Consensus conflict + low conviction') on a flush session. Nothing else deteriorated: -0.67% to 1246.93, above EMA10 1229.15 / EMA20 1211.11 / EMA50 1172.67 / EMA200 1045.75 all rising, RSI 57.82, Est_4W +2.43% and Est_12W +10.66% alive, no second Exit_Warning, ~78 pts clear of the trail. US-26 ALSO cannot be completed - Avg_Cost blank - so the arm is unverifiable by hand. Armed but not fired -> trail-raise; model prints no stop on a NoTrade row, so 1168.46 STANDS. See systemic finding #10 - tick (b) should not read a NoTrade transition as a conviction collapse. FILL Avg_Cost in input/usportfolio.json. |
| MU | Micron Technology, Inc. | 910.43 | HOLD | 910.43 | 929.07⚠ | 920.24 (+1.1%) | 986.47 (+8.4%) | 1056.62 (+16.1%) | Micron fell nearly 6% today and slipped below our protective level for the first time, but the cause is a policy headline that hit the entire memory sector, not anything specific to Micron: a report that the administration will let Apple buy memory chips from Chinese suppliers CXMT and YMTC. Analysts covering the space call the reaction overdone - CXMT cannot qualify for iPhones, YMTC has committed its newest capacity to domestic customers, and any China-sourced chips would only go into Apple products sold in China. Micron remains the highest-quality name in this book, is still 6% above our cost, and the 12-week view is the strongest we hold at +16%. One bad session on a sector headline is not an exit. OPERATIONAL: NO TRADE cv0 - US-17, a bare NO TRADE alone is never a dispose basis. L2's negative-12wk limb is badly unmet (Est_4W +8.35%, Est_12W +16.06%). SINGLE-session break of 929.07 (-2.01%), not confirmed. US-12/US-22 sector catalyst with the company thesis intact; MU traded as low as 897.86, SanDisk -9%, WDC -5/-7%. Quality_Rank 1/52, F=8, +5.93% over cost. Intraday rows printed Stop_Price 782.77/778.48 - ~16% loosenings, REFUSED under US-29. PRE-COMMITTED TRIGGER (US-35): a SECOND consecutive settled close below 929.07, OR any settled close below tonight's 887.60 low, fires SELL (30%) immediately. NVDA 08-26 carries direct HBM read-through into this name. · news |
| NBIS | Nebius Group, N.V. | 210.91 | HOLD | 210.91 | 213.00⚠ | 212.37 (+0.7%) | 232.67 (+10.3%) | 241.92 (+14.7%) | Nebius closed a hair under our protective level - less than 1% below it - on a day the whole AI complex de-risked ahead of Nvidia's results. That is a marginal breach, not a decisive one, the signal actually strengthened, and the 12-week view is still roughly 15% higher. We hold the core and leave the stop unchanged. OPERATIONAL: HOLD cv42, up from 35. -0.98% below the 213.00 trail - marginal, US-22 says hold not trim. Est_4W +10.3%, Est_12W +14.7% alive; no Exit_Warning; -0.33% vs cost; ADR-0020 erosion -9.63%, below the 15% threshold. Model prints Stop_Price 200.16 - a loosening, REFUSED under US-29. Standing overlay override ('keep the core, no adds until the Meta question clears') honoured as a RESTRAINT - NBIS is excluded from any post-print add on the 08-26 NVDA read-through - while noting the overlay is 51 days stale and under US-31 has lost origination force. |
| NET | Cloudflare, Inc. | 280.35 | HOLD | 280.35 | 283.72 | 281.68 (+0.5%) | 279.40 (-0.3%) | 289.89 (+3.4%) | Cloudflare gave back yesterday's recovery and slipped just over 1% below our protective level - a marginal break, one session after it had reclaimed that same line. The signal is a watch, not a sell, and the 12-week view remains modestly positive, so we hold. But the heal-and-break cycle has now run twice in five sessions and does not get unlimited free passes, so the exit condition is written down in advance. OPERATIONAL: WATCH cv52 on PullbackBuy_VWAP. -1.19% below 283.72 - marginal, not decisive (US-22); label is a WATCH not a sell (US-17); Est_12W 289.89 = +3.4% alive so L2's second limb is unmet. Model prints 280.83 - a loosening, REFUSED under US-29. Peak Stop stays 301.75, un-rebased, no cut has occurred. PRE-COMMITTED TRIGGER (US-35): a SECOND consecutive settled close below 283.72, OR any settled close below tonight's 280.18 low, fires SELL (30%) immediately. CRM reports 08-26 AMC - enterprise-SaaS read-through lands on NET and OKTA together. WARNING: NET still reads in input/usportfolio.json on the 20-session duplicate-row defect, so portfolio_monitor.py MISSED this breach entirely (reported 5, actual 6); treated as @ 274.25 per the transaction ledger. |
| SPCX | Space Exploration Tech Corp | — | HOLD | — | 129.00 | n/a | n/a | n/a | SpaceX has no tradable quote in our data, so it carries no signal and no price targets - we assess it on news alone. The staggered share-unlock schedule is proceeding and the market has absorbed every stage so far: the first release in early August roughly doubled the freely traded float and the stock still rose 6% that day and 35% over the following week. Another 7% unlock is due around 10 September, with much larger releases near November earnings and in December. Morgan Stanley's view that the fundamentals are largely unchanged still stands. Classification: watch, not concern. We hold on the manual stop. OPERATIONAL: permanent no_close, MUTED on signals - no Action, conviction, entry or target synthesized, never guessed. 08-06 tranche freed ~911.5m shares (-116bn), float 4.9% -> 11.8%, stock +6.1% that day and +35% in five sessions (~+$500bn mkt cap); 319m more on 08-12; a 7% unlock around 08-21. Next: ~7% around 2026-09-10, then ~1.3bn near Q3 earnings in early November, then the 180-day expiry in December; Musk's 6.42bn shares locked until June 2027. Manual stop 129.00, ~4.4% below the last quoted level - a live line. No trim, no adds. AUTO-UNMUTES the moment it prints a non-null Close. · news |
Outcome & track record
Accuracy at the time of this record.
157 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.