Recommendations
0 to acquire · 0 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 |
|---|---|---|---|---|---|---|---|---|---|
| 0992 | Lenovo Group Limited | 28.54 | HOLD | 29.29 | 24.33 | 31.74 (+8.4%) | 39.19 (+33.8%) | 40.67 (+38.9%) | HOLD the book's only position: Lenovo is the HK-listed AI-server bellwether and the thesis is intact - revenue growth is not in doubt (+161.8% YTD, +114.8% over 30 days on the datacentre mix shift), the 12-week projection of 40.67 sits inside the 43.49 DCF anchor, and the stock trades far above every moving average. The engine has simply stepped back from offering more: the label went BUY - ADD cv69 to HOLD cv66, the signal rotated BreakoutUp_52W to AnticipationUp_VCP, is 0.0, and the re-arm entry of 29.29 now sits ABOVE the 28.54 close - the breakout has gone back into its base. Valuation is a measured headwind, not a veto (FAIR_VALUE, MoS -0.6056 on the freshly regenerated snapshot, val_adj -3, adjusted conviction 63 - which would have failed the 65 floor had this printed as a buy); the EPV anchor at 17.78 drives that negative reading while the DCF anchor at 43.49 would read +0.52, and on a name growing revenue 20% the truth is bracketed by them. Quality is genuinely strong - F-Score 7, ROE 27.9%, Quality_Rank, Bargain True. SELL CLASSES ALL CLOSED: (a) no Exit_Warning code, so the section 9 warning-elevate path has nothing to elevate; (b) NOT PROFIT_ARMED - true Gain_ATR 0.63 on the blended 27.498 cost (needs 3.0, i.e. a close >= 32.45) and T12_Progress_Pct 70.17 (needs 90, i.e. a close >= 36.60), so ADR-0012 cannot fire and ticks b/c/d are unreachable because arming is their precondition; (c) no stop break - 14.7% of headroom vs the 24.33 stop, so L2 does not apply; (d) 12-week view +42.5%, so L2's second leg fails independently; (e) label is not bearish; (f) price above EMA5/10/20/50/100 and far above EMA200 15.99. ADR-0020 trail erosion 0% - peak stop ROSE for a fourth straight session (21.31 -> 21.52 -> 22.73 -> 24.33); new peak recorded at 24.33. TRAIL-RAISE (not armed, so not the ADR-0012 mandatory kind - this is the model's own chandelier_trail): stop 22.73 -> 24.33, +7.0%. Risk if stopped = (27.498 - 24.33) x 10,000 = HK = 1.58% of book; the stop still sits 11.5% BELOW the blended cost and becomes a break-even line near a close of 31.5. CATALYST-AWARE HOLD APPLIES WITH FORCE: Lenovo Q1 FY2026/27 results 2026-08-13 is TWO trading sessions out (<=5), two-sided band +/-6.0% / +/-0.94 ATR, confidence high, horizon stance queue-below. Standing instruction is HOLD/trail THROUGH the event - a trim would require a confirmed lower-high or a decisive stop break and neither exists. Do NOT pre-trim into 08-13, and do NOT add before it (L6). On the +/-6% band the position is worth HK-302,500 the morning after, a HKtwo-sided swing = 0.85% of book either way, survivable at this stop. NEWS confirms with a louder counterweight than last session: no placement/registration/secondary found, so HT-6 is re-checked and CLEAR; but ISG booked US29bn revenue (+35.8% y/y) and STILL posted an US operating loss, group gross margin 14.7% (management: 'north of 17%' ex-CSP), with memory/component cost inflation a stated 2026 headwind - that ISG margin line is exactly what the 08-13 print will be judged on. *** EXECUTION DISCREPANCY RECONCILED HERE (no separate ledger line is possible - one card per date/ticker): the 08-10 queued ADD FILLED today at 28.75, and the band-top placement is why it filled at all (session low was 28.36; a limit at the model entry of 28.30 would have missed by HK). BUT THE FILL WAS 3x THE AUTHORISED SIZE. The card sized ADD = HK -> post-execution target 6,000. The ledger recorded as that post-execution TARGET; the auto-executor read it as the quantity to BUY and filled 6,000, leaving the book with and HKspent - HKmore capital than authorised, committed two sessions ahead of a two-sided earnings print. Position is now 14.20% of book, not the ~8.6% the card wrote. This is a field-semantics defect between the ledger schema and the executor, not a judgement error, and it is filed as Systemic #1 with the fix (executor should compute add = - qty_at_decision, or the schema should gain an explicit qty_add). It is also the strongest argument for today's NO ACTION: the book already carries more risk into 08-13 than it chose to. *** ALSO NOTE the anchor row's position columns are STALE - the 17:00:15 scan ran 11 seconds BEFORE the 17:00:26 fill, so it prints Avg_Cost 25.62 / Gain_ATR 1.77 / Unreal_PnL_Pct 11.4 from the 4,000-share era; true figures are 27.498 / 0.63 / +3.79% (Systemic #2). Benign today because arming fails on either number. RE-ARM TRIGGER for a future ACQUIRE: a BUY - ADD label with > 0, mechanically a close/trigger through 29.29 - but not before 08-13 per L6. EXIT TRIGGERS: decisive close below 24.33; OR any Exit_Warning code; OR a SELL - REDUCE label; OR two consecutive NO TRADE/sub-40 prints with Target_12Week rolling under the close (L2). Estimates use f=0.30 (L5 mandatory - calib_1w -0.45, 9 graded buys < 10). CAVEAT on the 4wk figure: Target_Price 39.19 sits 20.3% ABOVE Est_High 32.57, and this overshoot has now been measured three sessions running at 17.1% -> 17.1% -> 20.3%, i.e. persistent AND widening (Systemic #3). 32.57 is the honest upper end of the 4-week range. The 12wk 40.67 carries NO DCF-exceedance annotation - it sits inside the 43.49 DCF anchor. · news |
| 3750 | CATL | 651.50 | WATCH | 651.50 | 641.00 | 668.00 (+2.5%) | 706.49 (+8.4%) | 583.33 (-10.5%) | NO BUY on CATL despite it being the highest-conviction non-held name in the book and the one row with a genuinely attractive printed reward/risk. The battery-and-storage growth case is real and the freshly regenerated valuation is respectable - FAIR_VALUE, Quality_Pass True, F-Score 6, ROE 21.9%, revenue growth 41.4%, Quality_Rank - but the model is not offering a trade and the forward view contradicts the setup, so the book stays out. Rejected on three independent grounds, any one sufficient: (1) WATCH is not an actionable label - 0.0, there is no trade to size; (2) val_adj -6 on MoS_FV -1.1621 takes conviction 63 to an adjusted 57, eight points below the 65 floor; (3) HT-4's forward-view screen fails outright - Target_12Week 583.33 sits 10.5% BELOW the 651.5 close, which is the exact structure HT-4 exists to catch and the same structure that cost 9618 its BUY label two sessions after this book declined it on 08-10. Additionally the printed stop of 641.0 is only 1.6% below entry - a 0.38 ATR stop on a name whose ATR is 27.50 - which is not a risk-definable level but noise, so the headline R:R of 5.24 is not a real number. Signal family FailedBreakdown_Up_20D is L1-discounted here and HT-2 keeps scorecard_bias at 0 (n=2, below the n>=5 floor). Re-arm trigger: an actionable BUY label with Target_12Week above the close and a stop at least 1 ATR away. |
| — | AI - HK Tech | — | NO ACTION | — | — | n/a | n/a | n/a | NO ACTION both sides - an affirmative verdict, not an absence of one. The buy pool is EMPTY: not one of the 21 rows in the anchor snapshot carries a bullish actionable Action (no STRONG BUY, no STRONG BUY - ADD, no BUY, no BUY - ADD), and the book's own holding printed HOLD with 0.0 and a re-arm entry of 29.29 sitting ABOVE the 28.54 close. The sell pool is EMPTY: the only holding is 0992 and all three sell-origination classes (signal, ADR-0020 risk-discipline, amplified valuation) are checked and closed. No anticipate slot - no name has printed BUY - ANTICIPATE in any of the last four snapshots. It was a broad HK-tech down session: conviction fell on nearly every name (9618, 0100, 0148, 0068, 2513 and 6869, 1211 and 0700, and eight of 21 rows printed FailedBreakout_Down_20D falling knives with 12-week projections 26-43% below spot. HK (85.80%) fully retained; HKcommitted; nothing to execute at the 2026-08-12 open. Book value HK = +0.52% vs the 2,000,000 seed, 14.20% - a level the book did not choose (see the 3x executor over-fill on the 0992 card, Systemic #1) and will not add to ahead of the 08-13 Lenovo print. Valuation was STALE at 8 days and was REGENERATED this run (FetchedAt 2026-08-11 17:17:15), which confirmed the 08-10 card's own predicted MoS to the decimal. Scorecard: graded 9, calib_1w -0.45, f=0.30 mandatory per L5, every cell below the n>=5 floor so scorecard_bias is 0 everywhere per HT-2; the no_chase_queue counterfactual FLIPPED from n=5 median +5.07% [COST] to n=6 median -3.09% [SAVED], which argues to keep the no-chase bar strict rather than to loosen it. |
Outcome & track record
Accuracy at the time of this record.
9 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 →
-0.450 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.
50 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.