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 | 31.34 | HOLD | — | 29.12 | n/a | n/a | 44.66 | Hold the position in full and raise the stop. Lenovo has now given back a third of its earnings-gap gain over three sessions, but nothing in the business case has changed: Bank of America upgraded the stock to Buy on a record quarter that showed revenue up 43% to US9bn, AI revenue up 60%, and an AI-server order pipeline that grew 157% in three months to US$54bn. The pullback has taken the shares to 28% below the discounted-cash-flow anchor, so the fall has improved the valuation rather than damaged it. The protective stop moves up to 29.12, which is above the purchase price - from here the position cannot lose money, only give back profit. || Operational: 2nd consecutive NO TRADE cv0, NoTrade_Reason verbatim "Consensus conflict + low conviction"; row publishes no Entry/Stop/Target_Price. Close 31.34 (-5.09%) on 126.5m sh vs 128.6m (flat volume = orderly distribution). Trend 77 UNCHANGED, Momentum (improving off the floor), Volatility, consensus Conflicted, Exit_Strategy none (systemic #3). ADR-0012: FIRE TICK (d) COMPLETED (2nd consecutive NO TRADE) but NOT ARMED - Exit_Warning blank, Gain_ATR 1.64 < 3.0, T12_Progress_Pct 70.17 < 90 and FROZEN A 6TH SNAPSHOT (systemic #1, hand-computes to 22.4%). Arming is a precondition -> trail-raise, not trim; 3rd consecutive disarmed deterioration event (systemic #2). TRAIL RAISED 28.77 -> 29.12 = chandelier 3.0x ATR 2.3469 off unchanged 52W high 36.16; peak_stop re-bases to 29.12, 5.9% above the 27.498 cost, protecting HKHeadroom to trail 14.8% -> 7.1%. L2 does NOT apply (close above stop, 12wk +42.5%). ADR-0020 not triggered (0% erosion; hk carries no trail_erosion config). G1/G3 block a valuation trim anyway. Estimates: 1wk/4wk n/a - no Target_Price for a 3rd session, so the stale 34.75/38.77 pair carried since 08-14 is DROPPED not carried; 12wk 44.66 (+42.5%), exceeds the DCF anchor 43.49. Model band Est_Low 30.71 - Est_High 39.86; close has fallen to the bottom of the band. BREACH RULE: decisive close below 29.12 exits. RE-ARM to ADD only on a base-build in 29.12-31.00 with RSI < 65 (today 63.65, condition cleared for the first time since the gap) AND a bullish actionable label returning - do NOT place a limit today, the row prints no Entry_Price (provenance rule). HT-6 dilution condition absent. · news |
| 9988 | BABA-SW | 126.70 | WATCH | 126.70 | 121.00 | n/a | 136.06 (+7.4%) | 139.80 (+10.3%) | No purchase, and this one was close. Alibaba is the first genuinely actionable buy signal this book has seen in seven sessions, and it clears the conviction bar with a sensible risk-reward: 7% to the four-week target against 4.5% of downside to a definable stop. It is blocked on accounting quality - the company fails the book's fundamental screen, which is a hard veto on any name it does not already own. Two further reasons point the same way: this signal family is the least reliable buy pattern in this book, and the stock is rallying into June-quarter results that land in two sessions, where a single gap would jump straight through the stop. Wait for the print and the pullback. || Operational: WATCH cv53 -> BUY cv69, Buy_Rank 1, 28.39, Entry_Distance_ATR 0.00 (not extended). Fusion: val_adj -3 (MoS -0.687) + bias 0 -> adj cv 66, CLEARS the 65 floor. HT-4 PASSES: T4 136.06 (+7.4%) and T12 139.80 (+10.3%) both positive, stop 121.00 = session low, R:R 9.36/5.70 = 1.64 >= 1.5. BLOCKED by Quality_Pass FALSE (F-Score 3) = hard quality-gate block on a -0 new entry - binding and non-discretionary. Independent supports: (i) L1/HT-2 - FailedBreakdown_Up_20D is the discounted-chase family here, standalone scorecard n=2 hit_1w 0.50 with 50% stop-first, mirror-era +10 edge retired as a regime artifact (D1); (ii) BABA Q2 results 2026-08-20 pre-US-open, HK reaction 08-21 - pre-earnings run-up on analyst-briefing reports of narrowing instant-commerce losses and ~45% cloud growth; a 4.5% stop does not survive an earnings gap (L6 spirit applied on the way in). Structure: close 126.70 still 0.35% BELOW EMA200 127.15; Volatility pillar 21 (Bearish). Logged as withheld so the counterfactual grader prices the refusal - if BABA prints well and runs, review this honestly at the next learnings review (systemic #5). · news |
| 0981 | SMIC | 75.85 | WATCH | 75.85 | 73.60 | n/a | 83.97 (+10.7%) | 69.11 (-8.9%) | No purchase. SMIC's own twelve-week model view sits 9% below today's price - the setup asks the book to buy something the model expects to be worth less by November - and the valuation screen puts it at the far end of expensive on weak accounting quality. A watch-list flicker is not an entry. || Operational: AVOID cv63 -> WATCH cv63 (not actionable) -> adj cv 53 after val_adj -10 (clamped; MoS -4.658), 12 pts sub-floor. HT-4 kill: T12 69.11 vs close 75.85 = -8.9% forward view. Quality_Pass FALSE. Closed +0.93% on a session where the prior session’s bounce leaders reversed hard. |
| 0100 | MINIMAX-W | 318.80 | WATCH | 318.80 | 286.60 | n/a | 405.32 (+27.1%) | 181.72 (-43.0%) | No purchase. MiniMax was the closest miss on conviction alone, but the model expects the shares to be worth 43% less in three months, and the stock swung 16% between its high and low today. That is not a risk this book gets paid to take, and interim results on 26 August add a binary event on top. || Operational: NO TRADE cv0 -> WATCH cv59 (not actionable) -> adj cv 59, val_adj 0 (NA valuation, quality gate off), 6 pts sub-floor - closest near-miss on conviction. HT-4 kill is decisive regardless: T12 181.72 vs close 318.80 = -43.0%. Session range 336.60/286.60 = 15.7% intraday swing on a -4.49% close. Interim results 2026-08-26 (horizon). |
| 0772 | CHINA LIT | 20.96 | WATCH | 20.96 | 20.50 | n/a | 23.41 (+11.7%) | 19.44 (-7.3%) | No purchase, second session running. China Literature's twelve-week model view remains below today's price and the valuation screen is at the extreme end of expensive. Nothing here to own. || Operational: WATCH cv55 (flat, not actionable) -> adj cv 45 after val_adj -10 (clamped; MoS -6.422), 20 pts sub-floor. HT-4 kill: T12 19.44 vs close 20.96 = -7.3%. Quality_Pass FALSE. Closed -2.42%. |
| — | hktech | — | NO ACTION | — | — | n/a | n/a | n/a | No trades today. Yesterday's broad Hong Kong tech bounce failed within a session - the names that led it reversed 8-13% - which is exactly why the book refused to chase them, and that refusal paid for itself in one day. The one actionable buy signal to appear in seven sessions, Alibaba, was blocked on accounting quality and sits two days from earnings. The single holding, Lenovo, fell 5% in a third session of profit-taking after its record quarter, but its trend is intact, Bank of America upgraded it this week, and the shares now trade 28% below the cash-flow valuation anchor - so it is held in full with the stop ratcheted up to 29.12, which is above the purchase price. The book stays 85% in cash. || Operational: ACQUIRE = NO ACTION - 9988 BUY cv69 -> adj 66 CLEARED the floor and PASSED HT-4 (R:R 1.64, T4 +7.4%, T12 +10.3%) but is hard-blocked by Quality_Pass FALSE on a -0 entry, and independently fails L1 (FailedBreakdown chase, 50% stop-first here) plus the 08-20 earnings-gap test; anticipate slot empty. DISPOSE = NO ACTION - sole holding prints NO TRADE with fire tick (d) COMPLETED but UNARMED (Gain_ATR 1.64 < 3.0, T12_Progress frozen 70.17 < 90), no Exit_Warning, +42.5% 12wk. Only change: 0992 trail 28.77 -> 29.12 (no order). HKfully retained (14th session). Book HK = +1.92% vs seed, 15.37% / cash 84.63%. Scorecard refreshed: entries 98, graded 10, calib_1w +0.129 but f = 0.30 retained per L5 (metric swung 0.475 on one observation); no actionable cells, all bias 0. |
Outcome & track record
Accuracy at the time of this record.
10 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 →
0.129 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.