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 |
|---|---|---|---|---|---|---|---|---|---|
| CB | Chubb Ltd | 341.90 | SELL 50% | 341.90 | 331.74 | 339.86 (-0.6%) | 335.09 (-2.0%) | 321.48 (-6.0%) | Chubb has rolled over: the insurer closed at, back below its 50-day average, and the model's 12-week view has turned negative at (-6.0%), so we are taking half the position off while the stock is still only 3% under our cost. The sell signal has now printed in three of the last four sessions on the book's most reliable exit pattern, and the shares failed to join Thursday's broad market rebound. We keep half because the business is not expensive - our own fair-value work still reads Chubb as modestly undervalued, which is why this is a half-sale rather than an exit. OPS: engine portion 88.52% cut one tier to 50% on the UNDERVALUED valuation read (CLAUDE.md s9 amplify-only, argue-down). Scorecard cell FailedBreakout_Down_20D|sell +1 (n=55, hit_1w.528) -> adj 66, clears the 65 floor. The EOD 09-16 re-arm (close <) did NOT fire; this is originated by the live printed signal, not the rung. Stop re-based to (model exit stop) and peak_stop re-based on the cut. Next rung: close below -> sell the balance. · news |
| AAPL | Apple Inc | 337.00 | SELL 30% | 337.00 | 322.23 | 349.98 (+3.9%) | 380.26 (+12.8%) | 369.09 (+9.5%) | Apple has run from to in the five sessions since the iPhone 18 Pro keynote - a +6.9% move that leaves the position 9.2% above our cost and stretched nearly 5% above its volume-weighted average price. The model's confidence in the move halved overnight even as the price held, which is the classic signature of a rally running out of buyers. We bank roughly a third of the gain into the retail launch weekend, where Apple has historically 'sold the news', and keep 70% of the shares because the 12-week view is still constructive at OPS: ADR-0012 profit-take - ARMED (Gain_ATR 4.02, T12_Progress 91.3, T12 > close) + FIRED on tick (b), conviction -26 vs the prior EOD snapshot. G1 softening does not apply (conviction 43 < 80) and ticks (b)/(c)/(d) fire regardless. Tier held at profit_take_1 = 30%: the MoS_FV -3.06 one-tier bump is SUPPRESSED by G2, the DCF-reliability gate (DCF anchor vs a close and a 12-week target is unreliable; FV_per_share is plainly broken). Never 100% from a profit-take. Trail raised ->. Catalyst note: iPhone 18 Pro first retail weekend 18-Sep and quarterly triple witching the same session - 70% stays exposed through both. The 4wk/12wk estimates are UP, not down: this is a de-risk into strength, not an exit. · news |
| NVR | NVR Inc | 6242.23 | SELL ALL | 6242.23 | — | 6179.24 (-1.0%) | 6032.27 (-3.4%) | 5612.34 (-10.1%) | NVR has printed a strong-sell in four straight sessions and sits below both its 50-day and 200-day averages with a 12-week view of (-10.1%). The homebuilder barely participated in Thursday's rally, up 0.9% against the S&P's 1.1%, and is fighting a 10-year Treasury above 5% after the Fed's hike - the wrong backdrop for mortgage-sensitive names. The holding is a-share, stub left from the seed and there is no case for keeping a deteriorating position at a size too small to matter - we close it out. OPS: engine portion 88.52% of = 2.66, rounds to at the US lot of 1, so this is a full exit. Signal exit (stable REDUCE x4, Sell_Rank #), NOT a profit-take, so the never-100% rule does not apply. Adj 76 clears the floor. Ranked into slot on dollar bleed with loss-prevention precedence over NUE's displaced profit-take. 12wk annotation: exceeds the DCF anchor of - projection assumes growth beyond priced-in. Position removed from state on fill. |
| GOOG | Alphabet Inc Cl C | 343.68 | NO ACTION | 343.68 | 328.74 | 340.41 (-1.0%) | 332.77 (-3.2%) | 310.95 (-9.5%) | The model flagged Alphabet's C shares for a sale, but we are not taking it. The signal fired on a day the stock ROSE 1.3% into a broad market rebound, it had disappeared for two sessions before reappearing, and the far larger Class A holding of the same company did not print a sell at all - selling one share class and keeping the other five times bigger one is arbitrary. On top of that, the sealed portion of Judge Brinkema's ad-tech remedies opinion is due to be unsealed within days; Alphabet already avoided a forced breakup, and we would rather learn the terms than sell blind into them. We raise the protective stop instead. OPS: news OVERRIDE demotes a technically-qualifying DISPOSE (adj 66 = 65 raw +1 cell bias, would have been slot on bleed). Late-flipped REDUCE, prints. Catalyst-aware hold: regulatory unsealing <=5 sessions, per Step 7 trim only on a confirmed lower-high or decisive stop break - neither present (close vs stop). Trail raised ->; peak_stop unchanged at (erosion 5.3%, ADR-0020 clear). RE-ARM: a close below with the 12wk still negative -> SELL 50%, and the same treatment extends to GOOGL if its row flips to STRONG SELL - REDUCE. · news |
| MCO | Moody's Corp | 461.88 | NO ACTION | 461.88 | 450.30 | 456.86 (-1.1%) | 445.11 (-3.6%) | 411.58 (-10.9%) | Moody's turned in a fresh strong-sell and the chart is genuinely broken - sits well below both its 50-day and 200-day averages with a 12-week view of (-10.9%) - but we are holding for now. The specific pattern that triggered it has been the single worst-performing exit signal in this book's history, right about one time in eight, and this is only its first print after two quieter sessions. We arm a price trigger rather than sell into a signal with that record. OPS: RallySell_BearTrend|sell bias -8 (ACTIONABLE, n=18, hit_1w.125) -> adj 55, below the 65 floor; the strongly-negative-cell rule requires raw conviction >=75 and MCO is 63. L2 does not admit it either: close is above the stop, so there is no stop break. Largest displaced dollar bleed in the book at. RUNG: a close below with the 12wk still negative is an L2 stop-break basis that BYPASSES the cell haircut -> SELL 50% on the next EOD. |
| NUE | Nucor Corp | 265.14 | NO ACTION | 265.14 | 253.64 | 269.69 (+1.7%) | 280.29 (+5.7%) | 279.20 (+5.3%) | Nucor is our best-performing holding at +20% and it has started showing a momentum divergence - the price keeps making highs while the underlying momentum gauge does not. That normally argues for taking some profit. We are not acting today because the position is only, the model's conviction is actually RISING, and the trend is intact well above both moving averages. Two larger de-risks take priority. OPS: ADR-0012 ARMED (Gain_ATR 5.86, T12 > close) + FIRED on tick (a) RSI_DIV. 1 tick -> profit_take_1 30% = / DISPLACED by the <= cap; ranked below NVR on loss-prevention precedence. Ledgered optional so the counterfactual grader prices what not acting cost. G1's letter does not cover the FailedBreakdown_Up_20D family but its spirit (NBIS protection: divergence in a strong uptrend usually resolves up) applies at conviction 85. Trail held at - the model's chandelier is LOWER and a trail is never lowered. If the trim is still live next EOD it takes a slot. |
| KO | Coca-Cola Co | 88.06 | HOLD | 88.06 | 87.69 | 88.78 (+0.8%) | 90.46 (+2.7%) | 90.10 (+2.3%) | Coca-Cola is the book's largest position at and it recovered this session - the model's conviction bounced from 49 back to 65 and the stock closed up at. Last night's profit-take trigger has therefore un-fired, so we simply hold and keep the protective stop tight, 0.4% under the close. OPS: ADR-0012 ARMED (Gain_ATR 3.26, T12_Progress 97.7, T12 > close - a genuine arm, not the T12_Progress artifact) but NOT FIRED: tick (b) cleared on the conviction recovery, no second Exit_Warning code, T12 > 1.02x close so the forward view is not dead, and no repeated NO TRADE. Armed-but-not-fired -> mandatory trail-raise, not a trim. Model Stop_Price is LOOSER than the standing, so the tighter existing trail is retained. A close below is tick-equivalent and fires the 30% ( /) on the next EOD. |
| ALLY | Ally Financial Inc | 40.60 | NO ACTION | 40.60 | 39.44 | 39.86 (-1.8%) | 38.14 (-6.1%) | 33.23 (-18.2%) | We already halved Ally at the open on Thursday at, and the remaining stub still carries the worst forward view in the book at -18%. The stock held above our protective line today and the model's conviction is fading rather than building, so we let the existing trigger do the work instead of selling twice into the same weakness. OPS: SOLD 09-17 open @ (fill confirmed, txn usbh-20260918-0001). Third consecutive STRONG SELL - REDUCE but conviction decaying -> 69; RallySell_BearTrend|sell bias -8 -> adj 61, below the 65 floor, and raw 69 < the 75 required on a strongly-negative cell. No stop break (close >) so L2 does not admit it. RUNG STANDS: a close below -> sell 50% of the balance. |
Outcome & track record
Accuracy at the time of this record.
182 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 →
-0.078 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.
84 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.