Recommendations
3 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 |
|---|---|---|---|---|---|---|---|---|---|
| PLTR | Palantir Technologies Inc. | 185.93 | BUY — QUEUED @165.23 | 165.23 | 155.72 | 185.83 (+12.5%) | 192.22 (+16.3%) | 204.23 (+23.6%) | Palantir is the strongest technical setup in the book right now - top-ranked buy at maximum conviction (100) on the pullback family that has been this portfolio's only reliable buy signal, backed by a Q2 that grew revenue 93% year-on-year and a Pentagon Maven contract now a formal program of record worth up to 3bn over five years. It passes the quality screen (F-Score 8, quality rank), but at the shares sit 2.5 ATR above the model's entry - richly priced against intrinsic value - so we bid on a pullback rather than chase. Operational: L1-favoured cell (+1 bias); no-chase bar breached at Entry_Distance_ATR -2.52 so QUEUED as a GTC limit, not bought; size halved to 0.5x on MoS -13.08; auto-withdraw on a family flip off PullbackBuy_VWAP, conviction <65, Quality_Pass turning False, or 2026-09-01 passing unfilled. If PLTR gaps below 165.23 do NOT fill at the gap - reprice to the live Entry_Price (US-9). · news |
| OKTA | Okta Inc. | 172.91 | ADD — QUEUED @156.65 | 156.65 | 147.61 | 173.80 (+10.9%) | 180.48 (+15.2%) | 194.98 (+24.5%) | Okta re-rated hard on a beat-and-raise that reframed it as the identity layer for AI agents - revenue (+11%), EPS vs, FY guidance lifted, and the stock closed +28.6% at a 52-week high with Needham taking its target to from We already own at a 24% gain and want more, but not at the top of a one-day 28% gap; the order is a limit back at the model entry, which is where a normal post-gap consolidation would take it. Operational: held-name ADD so the quality gate (Quality_Pass False, F=7) does not block it; L6/US-9 bar chasing the gap so the printed Buy Stop @156.65 is placed as a GTC LIMIT (a stop below market would fill at market = the chase L6 forbids); size halved on MoS -4.34; the standing trail is raised 140.55 -> 147.61 (2.0x ATR chandelier off the 174.85 high) and governs the position, not the model's 114.29 stop. Auto-withdraw on conviction <65, a family flip, or 2026-09-01 unfilled. · news |
| CRM | Salesforce Inc | 252.05 | BUY — QUEUED @213.47 | 213.47 | 165.24 | 251.98 (+18.0%) | 254.70 (+19.3%) | 290.59 (+36.1%) | Salesforce broke a long slump with the cleanest print of the week: adjusted EPS, revenue 35bn (+11%), Agentforce plus Data Cloud ARR near 9bn (+210%), a full-year EPS guide raised roughly 18.5% at the midpoint, and a widened Anthropic partnership (Claudeforce). Uniquely in this book it is the one name our discounted-cash-flow anchor values ABOVE the market price (vs) and the only Bargain flag on a quality-passing name. But it closed +22% on the day at RSI 80 - so we bid at the model entry, roughly a full gap-fill lower, and get paid to wait. Operational: chase-family cell (BreakoutUp_20D bias -4, hit_1w.286) so raw conviction 90 must carry it, which it does; Entry_Distance_ATR -2.49 breaches the no-chase bar so QUEUED, not bought; the printed Buy Stop @213.47 is placed as a GTC LIMIT for the same reason as OKTA; size 0.75x on MoS -2.51. Note the model stop 165.24 is 22.6% below the limit - unusually wide; tighten to a 2.0x ATR line on fill. Auto-withdraw on conviction <65, family flip, or 2026-09-01 unfilled. · news |
| KLAC | KLA Corporation | 183.77 | SELL 30% | 183.77 | 184.00⚠ | 184.56 (+0.4%) | 186.66 (+1.6%) | 183.03 (-0.4%) | This is a risk-discipline trim, not a view on KLA's business. The protective stop on this position has been walked down from to - a 32% retreat - across 21 sessions while the position size never moved once, which is a risk-management failure regardless of the fundamentals. Those fundamentals remain sound (Q1 FY27 guided to 0bn revenue, 5bn backlog, advanced packaging up ~70%, zero sell ratings among 29 analysts), which is precisely why we cut 30% and keep in the semicap cycle rather than exiting. The 12-week model view of against a close is the only negative forward in the book. Operational: ADR-0020/US-36 Trail Erosion, FIRST fire on the 272.00 watermark - the position has still never been cut, so the held-name sell veto stack (US-17, US-18/34, US-22, US-30) does NOT apply, by rule. Note the US-18/34 signal leg has LAPSED tonight (label AVOID cv61, not a 3rd settled STRONG SELL - REDUCE) - under US-32 the card de-escalates to its ADR-0020 floor of 30% and does not escalate. Valuation amplify (MoS -4.24 would bump SUPPRESSED by G2: Est_12W is 2.96x the DCF anchor, so the anchor is unreliable. Model Stop_Price 197.97 prints ABOVE the close - the US-29 bearish-family artifact - refused; trail holds 184.00. On execution the watermark re-bases 272.00 -> 184.00 and future fires face the veto stack again. THIRD consecutive session this card has been written unexecuted - ADR-0019 Disposition owed. · news |
| GOOGL | Alphabet, Inc.- Class A | 340.65 | SELL 30% | 340.65 | 333.68 | 342.22 (+0.5%) | 339.49 (-0.3%) | 362.01 (+6.3%) | Banking part of a 177% winner while its near-term outlook has gone flat and a real catalyst has landed. Chief Scientist Jeff Dean is leaving after 27 years to found his own startup, taking three senior researchers with him, and the flagship Gemini model still has not shipped after a planned June launch - shares fell ~4-5% on the news and now sit below both their 20- and 50-day averages. The four-week model view is -0.34% while the twelve-week is still +6.27%, which is exactly why this is a 30% trim and not an exit: Cloud backlog above $514bn and the broader franchise are intact, so stay. Operational: ADR-0012 armed + FIRED on tick (d), THREE consecutive settled NO TRADE prints (08-25/26/27). US-26 VERIFIED GENUINE - Gain_ATR +35.99, Unreal_PnL_Pct +176.95%, both positive, so this is not the armed-loser artifact. Ticks (a)/(b)/(c) all declined and stated: (a) PROFIT_ARMED is the only warning code, (b) conviction no change, (c) Est_12W 362.01 > 1.02 x 340.65 = 347.46 and US-33 blocks the limb on a T12-armed name anyway. profit_take_1 = 30%, never 100% from a profit-take. US-19: this is the SAME first fire carried from 08-26 unexecuted, NOT an escalation - portion held at 30%. No valuation amplify (MoS -0.69 above the -3 threshold). Ex-dividend 09-04 is immaterial (~6bps). Trail holds 333.68; watermark re-bases 358.00 -> 333.68 on execution. SECOND consecutive session unexecuted - ADR-0019 Disposition owed. · news |
| NVDA | NVIDIA Corporation | 227.98 | HOLD | 228.78 | 217.07 | 228.75 (-0.0%) | 233.48 (+2.1%) | 238.02 (+4.0%) | Hold and trail. Nvidia's Q2 delivered record revenue of 2bn (+106% year-on-year), EPS (+120%), a Q3 guide near $108bn and a CFO forecast of 70% growth in fiscal 2028 against a 44% street estimate - the stock rallied 8% to 227.98 and dragged the whole tape with it. Conviction has climbed -> across four snapshots and the twelve-week view is alive at +4.4%; nothing here says sell. Operational: the 08-26 pre-committed trigger (a settled close below 217.07 that is also below the post-print session's low) is UNTRIGGERED - the close came in 5.03% ABOVE the line - and is now DISSOLVED, replaced below. ADR-0012 is armed (PROFIT_ARMED, T12_Progress 95.78) and tick (a) is technically present (RSI_DIV), but US-26 CANNOT BE COMPLETED - Avg_Cost is blank in usportfolio.json, so Gain_ATR and Unreal_PnL_Pct are unverifiable - and US-26 requires the profit-take be declined by rule in that case. Armed-but-not-fired therefore resolves to a mandatory trail-raise; both candidate raises LOOSEN the line (model Stop_Price 211.56, 2.0x ATR chandelier 213.25) so US-29 refuses them and the trail HOLDS 217.07. New pre-committed trigger (US-35): two consecutive settled closes below 217.07 fires SELL (30%) at the next EOD without re-litigation. FIX THE COST FIELD - fill NVDA's Cost in usportfolio.json. · news |
| OKTA | Okta Inc. | 172.91 | HOLD | 156.65 | 147.61 | 173.80 (+10.9%) | 180.48 (+15.2%) | 194.98 (+24.5%) | The existing are held and the protective line is raised. Okta closed +28.6% on a beat-and-raise; the trail moves 140.55 -> 147.61 (2.0x ATR chandelier off the 174.85 session high), locking in a stop above the pre-print price for the first time. Operational: US-27 satisfied in the position's favour - the catalyst-hold survived the print and the six-session trim thread is now formally closed, not merely withdrawn. Model Stop_Price 114.29 refused as a loosening (US-29). Peak Stop re-bases upward 144.84 -> 147.61; ADR-0020 erosion 0.00%. Watch line: a settled close back below the 156.50 gap-day low is the gap-failure signal and re-opens the trim case. Companion ADD card queued at 156.65 - the position can add on a retrace while the trail protects it. |
| MU | Micron Technology, Inc. | 935.39 | HOLD | — | 929.07 | 943.71 | 1027.82 | 1106.34 | Hold. Micron is the highest-quality name in the book (quality rank, F-Score 8), sits 8.8% above cost, and carries the strongest forward view we have: +9.9% at four weeks and +18.3% at twelve. It traded down to 906.89 intraday - below our 929.07 stop - but closed at 935.39, back above it. Operational: L4 governs, the settled close is authoritative, so the intraday breach is NOT a break and no trim case opens. Label went NO TRADE cv0 (one settled print, not two) - a bare NO TRADE is never a dispose basis (US-17). Not profit-armed (T12_Progress 84.55, no Exit_Warning). ADR-0020 erosion 0.00% - the cleanest risk line in the book. AVGO reports 09-02 as the next HBM read-through. |
| NBIS | Nebius Group, N.V. | 218.48 | HOLD | 218.48 | 213.00 | 221.54 (+1.4%) | 236.19 (+8.1%) | 270.16 (+23.7%) | Hold. Nebius recovered +2.13% to 218.48, a comfortable 2.57% clear of the 213.00 line, with a twelve-week view of +23.7% and no warning codes. Operational: model Stop_Price 210.05 refused as a loosening (US-29). The 2026-07-05 overlay note (keep the core, no adds until the Meta question clears) is 54 days stale - under US-31 it survives as restraint only and cannot originate. Event Horizon: NVDA's ~9.3% stake lock-up window opens around 09-08 - a structural overhang to watch, NOT a sell basis. Pre-committed trigger stands: two consecutive settled closes below 213.00 fires SELL (30%). |
| LLY | Eli Lilly & Co | 1176.10 | HOLD | 1176.10 | 1168.46 | 1181.84 (+0.5%) | 1193.34 (+1.5%) | 1192.71 (+1.4%) | Hold, but this is now the tightest risk line in the book - the 1176.10 close sits just 0.65% above the 1168.46 trail after a third straight down session. The defensive-anchor thesis is intact (quality rank, F-Score 8) and the forward view is flat-to-positive, so there is no sell case; there is also very little room left. Operational: conviction faded -> and the BUY - ADD print of 08-26 did not repeat, so the US-35 re-arm (a SECOND consecutive settled BUY - ADD on PullbackBuy_VWAP at raw >= 76) is NOT triggered and lapses. ADR-0012 armed (T12_Progress 98.61) but US-26 CANNOT BE COMPLETED - Avg_Cost blank - and tick (c) is blocked by US-33 on a T12-armed name; declined by rule. Model Stop_Price 1099.95 refused as a loosening (US-29). A settled close below 1168.46 fires the standard 30% trim review at the next EOD. FIX THE COST FIELD - fill LLY's Cost in usportfolio.json. |
| NET | Cloudflare Inc | 308.23 | HOLD | 301.90 | 283.72 | 309.48 (+2.5%) | 312.02 (+3.4%) | 336.20 (+11.4%) | Hold. Cloudflare gained 8.19% to 308.23 on the Salesforce-led enterprise-software re-rating and flipped to a BUY print at conviction 74 - full vindication of the 08-26 decision to withdraw its trim on the settled-close reclaim (US-22). The position is +12.4% over cost and sits 8.6% clear of its trail. Operational: the ADD scored adj 65, exactly at the floor, and ranked 7th - outside the three ACQUIRE slots, so no add tonight. Trail HOLDS 283.72: both a model Stop_Price of 281.24 and a 2.0x ATR chandelier at 272.57 would LOOSEN the line, refused under US-29. Known defect: NET prints Held=N for a 23rd session because usportfolio.json carries a duplicate NET row with - this blinds Avg_Cost, the ADR-0012 arm test and the monitor's stop check on a position. One-line fix, outstanding three weeks. |
| SPCX | Space Exploration Tech Corp | — | HOLD | — | 129.00 | n/a | n/a | n/a | Hold on the manual stop; muted on signals and news-assessed instead. SpaceX remains absent from the scan (permanent no_close - no public quote feed), so no action, conviction, entry or target is synthesized for it. The news read is unchanged at WATCH: the staggered unlock keeps being absorbed - the 08-06 tranche freed ~911.5m shares and the stock rose 35% in five sessions, and the ~7% 08-21 tranche passed with only a mild give-back. Next pressure points are a further ~7% unlock around 09-10, a ~1.3bn-share tranche near Q3 earnings in early November, and the 180-day expiry in December. Operational: MSCI review effective 08-31 and the S&P DJI September announcement 09-04 are both potential index-demand catalysts. Manual stop 129.00 stands; no trim, no adds. Auto-unmutes and resumes full signal + fusion treatment on the first non-null Close in the scan CSV. · news |
Outcome & track record
Accuracy at the time of this record.
166 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 →
-0.191 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.