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 | 203.72 | SELL 30% | 203.72 | 193.38 | 204.69 (+0.5%) | 210.28 (+3.2%) | 205.37 (+0.8%) | Cutting 30% of KLA purely as risk discipline: the protective stop on this position has drifted 28.9% below its own high-water mark (272.00 down to 193.38) while the holding has never once been reduced, and a risk line walking that far away from an uncut position is a risk failure on its own. The business is not the problem — Q2 revenue rose 15.2% to 66bn on solid AI-driven demand — but the shares are priced roughly 3.3x above their intrinsic-value anchor and the model's own twelve-week view is only +0.8% from here, so very little upside is being surrendered for a meaningful reduction in exposure. stay long the thesis. OPS: ADR-0020 first fire, session 13, tier fixed at the 30% floor; US-36 veto-immune (US-17/18/34/22/30 do not apply). ADR-0012 tick (c) fired and was DECLINED under US-33 — armed on the T12_Progress limb (99.2) and tick (c) reads the same metric; US-26 passes (Gain_ATR 2.73, +15.91%). Valuation amplify to 50% SUPPRESSED by G2 (T12 205.37 vs DCF 61.80 = 3.32x). Trail: 1.5 ATR 188.37 is looser — stop 193.38 STANDS; model Stop_Price 224.18 sits above the close (S1 defect, US-29) and is not used. Dividend ex-date 17-Aug = execution date, so the seller keeps it. Re-price of an unexecuted card, not a US-19 escalation. · news |
| NBIS | Nebius Group N.V. | 277.68 | SELL 30% | 277.68 | 235.69 | 281.18 (+1.3%) | 302.05 (+8.8%) | 340.81 (+22.7%) | Banking a third of a position that has run 43.7% in three sessions and has never been trimmed, on risk discipline rather than on any doubt about the company. Nebius' Q2 was exceptional — revenue up 454%, AI-cloud up 514% to a 0bn run rate at a 50% adjusted-EBITDA margin, backlog above $40bn — and Bank of America raised its target to on the print, so the remaining stay fully invested in that story. The reason to cut is that the stop protecting this holding had fallen 20% below its high-water mark with no size ever taken off, and a name whose twelve-week range still spans roughly -46% to +168% should not be carried at full weight on hope alone. OPS: ADR-0020 first fire, session 3, pre-committed unconditionally 08-12; tier held at the 30% floor. IMPORTANT — fires on the CARRIED stop 215.37 (-20.23%); after tonight's raise to 235.69 erosion is -12.71%, INSIDE the 15% threshold, so this is the LAST session ADR-0020 fires on the 270.00 watermark absent fresh erosion (systemic #2, spec gap). ADR-0012 unavailable, not merely un-fired (Gain_ATR 2.36, T12_Progress 81.48, no warning code) — not claimed as a basis. Valuation amplify to 50% SUPPRESSED by G2 (DCF NA, Rev_Growth +855%). TRAIL RAISE 215.37 -> 235.69, stop-locking +11.4% over cost. PRE-COMMITMENT re-based: a settled close below 235.69, or a 2nd consecutive settled STRONG SELL - REDUCE, fires a mandatory 50% trim (105 of the remaining 210). · news |
| GOOGL | Alphabet Inc. Class A | 345.90 | SELL 30% | 345.90 | 332.73 | 346.92 (+0.3%) | 355.25 (+2.7%) | 385.09 (+11.3%) | Taking 30% off the book's biggest winner — up 181% on cost — after a fourth straight close below the line we committed to in advance, with the shares now sitting under both their 20- and 50-day averages and 15% below the 52-week high. Alphabet's business is doing fine: revenue grew 24% last quarter, Google Cloud accelerated to 82% growth on a $514bn backlog, and stay to own that. But the price has stopped confirming the story, the profit is very large, and the model has now gone quiet on the name for two consecutive sessions, which is exactly the pattern that historically precedes giving gains back. OPS: TWO independent bases at the same 30% tier — (1) 4th consecutive settled close below the pre-committed 350.66 (343.80/343.54/346.36/345.90); (2) NEW ADR-0012 armed on the Gain_ATR limb (25.39, Avg_Cost 123.00, US-26 clean, US-33 does not bind) + tick (d) = 2 consecutive NO TRADE prints. One tick -> profit_take_1 30%, no escalation, US-19 satisfied. Ticks (a)/(b)/(c) decline (T12 385.09 = 1.113x Close, alive). US-27 RE-VERIFICATION: a settled close >= 350.66 WITHDRAWS this card. Valuation val_adj -3, too shallow to amplify; G2 would suppress (T12 vs DCF 145.29 = 2.65x). TRAIL RAISE 332.13 -> 332.73. · news |
| MU | Micron Technology, Inc. | 971.66 | HOLD | 971.66 | 882.92 | 983.63 (+1.2%) | 1040.42 (+7.1%) | 1110.28 (+14.3%) | Holding all and raising the protective stop to. Micron broke a twelve-session quiet streak with the model turning constructive again, the shares are 13% above cost and trading well above every major moving average, and the twelve-week view still points 14% higher. Raising the stop banks the risk reduction without giving up the upside. OPS: NO TRADE x12 -> HOLD cv69. Not armed (Gain_ATR 1.90, T12_Progress 87.51, no warning code). TRAIL RAISE 851.72 -> 882.92 (1.5 ATR), which sets a NEW Peak Stop and re-bases ADR-0020 erosion to 0%. HOLD is not a buy action so MU does not enter the buy pool. Quality_Pass True, Q_Rank, F 8. SHARED with ustech/ussemicon — execute once (L7). |
| NVDA | NVIDIA Corporation | 225.16 | HOLD | 225.16 | 216.38 | 226.56 (+0.6%) | 230.59 (+2.4%) | 241.20 (+7.1%) | Holding all and tightening the stop to ahead of earnings on 26-Aug. The model went quiet on the name after several strong sessions, which is a reason to protect the position rather than to exit it — the shares are still above every major moving average and the twelve-week view is intact. OPS: HOLD cv96 -> NO TRADE cv0 is a LABEL FLIP, not a within-label decay, so it is not ADR-0012 tick (b) — declined consistently with GOOGL/KLAC/LLY. Tick (d) needs 2 consecutive NO TRADE prints; this is the first. PROFIT_ARMED prints with T12_Progress 93.35 but US-26 DECLINES THE ARM: Avg_Cost is blank in input/usportfolio.json, so Gain_ATR and Unreal_PnL_Pct are unverifiable — SESSION 6 of this defect. FILL Cost (systemic #3). TRAIL RAISE 215.45 -> 216.38 (1.5 ATR). Earnings 08-26, 7 sessions out, outside the 5-session catalyst window. SHARED with ustech/ussemicon — execute once (L7). |
| LLY | Eli Lilly & Co | 1180.16 | HOLD | 1180.16 | 1159.23 | 1186.20 (+0.5%) | 1208.58 (+2.4%) | 1244.86 (+5.5%) | Holding all on a stop of, which is now only 1.8% below the price — one of the two tightest lines in the book. Lilly slipped 2.4% and closed under its 20-day average for the first time in this run, though it remains above its longer-term averages with a twelve-week view about 5% higher. The stop, not a decision, will settle this one. OPS: 2nd consecutive NO TRADE cv0 (NoTrade_Reason verbatim: 'Consensus conflict + low conviction') = ADR-0012 TICK (d) IN TEXTBOOK FORM on a PROFIT_ARMED row with T12_Progress 94.80 — and US-26 declines the arm because Avg_Cost is BLANK, so there is no armed position for the tick to fire against. Had Cost been filled this is a mandatory 30% trim = That is the quantified price of one missing field (systemic #3). Trail: 1.5 ATR 1119.68 is looser — stop 1159.23 STANDS. Watch the EMA20 1185.55 break. |
| OKTA | Okta Inc. | 147.43 | HOLD | 147.43 | 144.84 | 147.81 (+0.3%) | 155.88 (+5.7%) | 169.12 (+14.7%) | Holding all on a stop of — just 1.8% below the price after Okta gave back 4.9% and surrendered most of the prior session's gain. The position is still 5.5% ahead of cost and the twelve-week view points ~15% higher, so this is not a name to exit on a single bad day; but it is the most likely stop event in the book next week and it is being watched on that basis. OPS: HOLD cv45 -> NO TRADE cv0 is a LABEL FLIP, not a within-label decay — not tick (b), same discipline as NVDA/LLY. NOT ARMED at all (Gain_ATR 1.04, T12_Progress 87.17, no warning code), so the profit-take path is unavailable regardless. Trail: 1.5 ATR 136.46 is looser — stop 144.84 STANDS. Quality_Pass False is irrelevant to a hold (the gate only blocks new entries). A stop break would be the L2 combo ONLY if the 12wk also turns negative, which it has not. Earnings 08-26. |
| NET | Cloudflare Inc | 315.78 | HOLD | 315.78 | 301.75 | 315.54 (-0.1%) | 319.01 (+1.0%) | 332.40 (+5.3%) | Holding all on a stop after Cloudflare fell 4.6% and the model dropped the buy signal it had carried for three sessions. No add is being made — the shares are still 8% above their 20-day average, which is not where we want to be buying — and the position keeps its stop about 4.6% below the market. OPS: BUY cv65 -> NO TRADE cv0 (NoTrade_Reason verbatim: 'Consensus conflict + low conviction'), so there is no buy row to evaluate. The fairvalue refresh FIXED the EPV artifact on this name: FV_per_share ->, MoS_FV -943.71 -> -9.70 (still deeply growth-priced, Quality_Pass False, F 4, ROE -12.7%). ADD RE-ARM carried, re-based to the live EMA20 292.12 +/- 0.25 ATR (287.41 - 296.83) with conviction >= 65. Trail: 1.5 ATR 287.53 is looser — stop 301.75 STANDS. WARNING 14TH SESSION: duplicate NET rows (150 / 0) in input/usportfolio.json make a position read Held=N to the scan — DE-DUPLICATE. |
| SPCX | Space Exploration Tech Corp | — | HOLD | 141.29 | 129.00 | n/a | n/a | n/a | Holding all 160 SpaceX shares on a manual stop. The share-lockup expiry that was the main risk to this position has passed without damage — the stock rose 6.1% the day insiders were first allowed to sell and 15.8% the day after, absorbing a doubling of the freely traded float, and Morgan Stanley calls the fundamentals largely unchanged at an attractive entry point. The holding has been re-marked to a close, its first observable public price in our records, which is 9.5% above the stop. OPS: permanently no_close in the scan (private-market row), so MUTED on signals — news-assessed only, classification confirms-hold. Price sourced from Alphabet's SpaceX-stake disclosure (551.2M shares, 9bn at 141.29 on Thu 08-13); re-marks the holding -> Fairvalue un-blinded this row (Val_Signal NA -> GROWTH_PRICED_IN, MoS_FV -63.6) — changes nothing, valuation never originates a sell. Cost still BLANK — fill it (systemic #3). AUTO-UNMUTE on the first scan row with a non-null Close. Re-arm on a settled close below 129.00. · news |
Outcome & track record
Accuracy at the time of this record.
134 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.