Recommendations
2 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 |
|---|---|---|---|---|---|---|---|---|---|
| 2888 | STANCHART | 242.60 | ADD | 242.60 | 232.95 | 240.31 (-0.9%) | 236.23 (-2.6%) | 235.56 (-2.9%) | Standard Chartered broke out to a fresh 52-week high on the strongest conviction reading anywhere in the book, and the trend has been building for six straight sessions while the price climbed from without a losing week. We are adding to a small existing position, sized modestly, with a stop 4% below. The model's flat 12-week projection is not a bearish view — that estimator reads roughly flat-to-spot on every single holding in this book, so it carries no information here. Valuation gives us nothing either way: the fair-value layer is blind on this name. OPERATIONAL: adj 95 (cv95 + 0 val NA_CCY + 0 bias; BreakoutUp_52W|buy n=4 report-only, hit_1w.75 — mentioned, not applied). Order_Type Buy Stop @242.07 sits BELOW the 242.60 close so it triggers at the open = effectively market-at-open; provenance band 242.07 +/- 0.25*ATR(4.56) = 240.93-243.21. EDist -0.12, inside the -0.25 no-chase bar; RSI 63.0. Volume 70.29% x 348 = 244.6 -> board lot 50 -> = HK; line goes 348 -> (2.7% of book); max_position_pct cap and both non-binding. Stop 232.95 = live Stop_Price, ~2.1 ATR, passes HK-12; this also RAISES the standing trail from 232.73. L1 tension stated: L1 warns BreakoutUp_52W 4wk targets overshoot and mean-revert — keep 4wk expectations near zero, treat as a 1-2 week momentum trade. NOT one of HK-1's three condemned chase families. Declined on 09-07 and 09-08 on the 'negative forward pair'; that reasoning is retired today on evidence (T12_Progress_Pct 102.99; all 82.8-103.0) and the name is +2.2% since the 09-04 STRONG BUY at 237.40. Yahoo 503'd twice; a domain-restricted search returned 178.20/+5.44%, irreconcilable with the 9-session tape — discarded as an unreliable scrape. Event risk: 09-11 US PPI, 09-14 US CPI, 09-17 FOMC+HKMA all touch 2888. |
| 0857 | PETROCHINA CO LTD H SHS | 10.43 | BUY — QUEUED @10.50 | 10.50 | 9.74 | 10.42 (-0.8%) | 10.47 (-0.3%) | 10.41 (-0.9%) | PetroChina is the only name in this 31-stock book that is simultaneously cheap on intrinsic value, passes our balance-sheet and cash-flow quality screen, and is printing a live buy signal. It trades on about 10x earnings against peers at 11x and the wider oil and gas industry at 13x, even after a 43% year. We are placing a buy order that only executes if the stock breaks out to a new one-month high — so we pay nothing unless the move confirms itself. Stop sits 7% below. OPERATIONAL: adj 78 (cv76 + 4 val on MoS +0.2008 with Quality_Pass True, -2 on the live BreakoutUp_20D|buy cell, n=13 hit_1w.385 — an HK-1 chase family, haircut applied and stated). RE-QUEUE, not a stale carry: the anchor's live Entry_Price and Order_Type are still exactly 'Buy Stop @ HK', identical to the 09-08 card, so this is a reprice-to-identical satisfying the provenance rule on the current snapshot's own terms; band 10.50 +/- 0.25*ATR(0.30) = 10.425-10.575. WARNING: today's high was 10.51 — the 09-08 queue's trigger WAS TOUCHED by one cent. Holdings still show and monitor trades_detected is empty, so it is NOT booked as filled per ground truth, but a live fill at 10.50 would sit -0.67% under the 10.43 close. ADR-0019 reconciliation gap, not a model failure. EDist +0.23: the entry sits ABOVE the market, correct for a buy-stop — self-limiting, cannot be a chase. Volume 71.55% x 600,000 x 20% = HKnotional; /10.50 = 8,177 -> board lot 2,000 -> = HK; max_position_pct 60% x 600k cap and both non-binding. Size modulator held at 1.0x, NOT the 1.25x that UNDERVALUED + Quality_Pass permits, because the model's forward estimates are dead flat (1wk -0.1%, 4wk +0.4%, 12wk -0.2%) — Target_Price 12.01 (+14.4%) is the real objective. Stop 9.74 = live Stop_Price, ~2.5 ATR, passes HK-12. L6 clear: H1 results moved it on 09-08, today is +0.77% follow-through not a gap, and a buy-stop cannot round-trip an unfilled order. Fifth BUY-family session in nine on a price rising 10.19 -> 10.43. Event note: 09-09 China CPI/PPI touched this name and is now behind us; next event 09-15 China activity data. expires_on 2026-09-14. · news |
| 2513 | KNOWLEDGE ATLAS | 913.50 | SELL 85.46% | 913.50 | 847.45 | 965.06 (+5.6%) | 1034.34 (+13.2%) | 1103.38 (+20.8%) | Z.AI broke decisively below the risk line we had set for it and has not recovered, while carrying the strongest sell reading in the book for a second straight session. It is our largest liquid holding at roughly 17% of the portfolio and sits 44% below cost, so we are cutting 80% of it and keeping a small stub. We are honest that the model's own 12-week projection points higher from here — but our sell signals on this market have a strong one-week record and essentially no four-week record, so we treat this as a tactical de-risk, not a call that the story is over. OPERATIONAL: adj 73 (cv69 + 0 val NA + 4 on RallySell_BearTrend|sell, n=25 hit_1w.682 vs hit_4w.200 — the book's best exit cell). Session 2 of the 09-08 triggered retirement (WATCH 980.00 broken by 6.53%); stable 4/4 66-66-69-69, Sell_Rank #1. Action price 913.50 = anchor close = anchor Entry_Price, Order_Type Market, corroborated TO THE CENT by Yahoo (913.500 / -0.27% / prev 916.000, 'At close 4:08:04 PM GMT+8') — the only name independently verified this run (Yahoo 503'd on attempts). Band 913.50 +/- 0.25*ATR(92.10) = 890.47-936.53. Volume 85.46% x 1,000 = 854.6 -> lot 100 ->, leaves 200, proceeds ~HKNOTE this row is the ONLY holding whose forward view is NOT degenerate: T12_Progress_Pct 82.79 vs 90-103 everywhere else, so the +20.8% 12wk really is a projection above spot, not an echo of it — the exit is justified on HK-9 (1-week instrument) and explicitly NOT on the 4wk view per HK-9(b). PROFIT_ARMED absent, so no ADR-0012 question. Residual stop RAISED 810.00 -> 847.45 (live Stop_Price, ~0.7 ATR, binding). News: no headlines dated 09-07..09-09; nearest is Moonshot AI's confidential HK IPO filing (~5d) — a mild sector negative on AI-listing supply, a read-through not a catalyst. Event risk: 09-11 Oracle Q1 FY2027 read-through and 09-14 US CPI both touch this name — de-risking before them, not into them. Non-execution of the 09-08 card cost HKHK-11 ESCALATION NOTICE: this is session 2; if unexecuted at the 09-10 open it is escalated ONCE more then RETIRED to a WATCH line at 847.45 — never a fourth proposal. · news |
| 0981 | SMIC | 65.00 | SELL 40% | 65.00 | 62.78 | 65.41 (+0.6%) | 66.59 (+2.4%) | 69.49 (+6.9%) | SMIC has printed a sell signal for three straight sessions and the conviction behind it has recovered rather than faded, so we are trimming 40% of the position and keeping the majority. Renewed US-China trade tension and Chinese rare-earth export controls are pressuring Chinese chipmakers directly. The counterweight is real — second-quarter profit more than tripled on AI demand and the company is expanding capacity — and that is exactly why this is a partial trim and not an exit. The shares also screen as expensive against our intrinsic-value estimate. OPERATIONAL: adj 72 (cv68 + 0 val + 4 on RallySell_BearTrend|sell). SESSION 3 = the HK-11 FINAL ESCALATION, pre-committed in the 09-08 entry. Unlike 09-08 (which was an explicit sub-floor carry at adj 58) this card CLEARS THE FLOOR ON ITS OWN: raw cv recovered, stable 4/4 65-65-65-68, Sell_Rank #2 — a re-originated card, not a carry. Action price 65.00 = anchor close = anchor Entry_Price, Order_Type Market; band 65.00 +/- 0.25*ATR(2.697) = 64.33-65.67. Volume HELD at the carried card's, NOT the model's 85.46% — escalating size on a final escalation would move two variables at once; effective 40%, leaves 1,500, proceeds ~HKValuation amplify WITHHELD: MoS_FV -4.66 <= -3 qualifies for the tier bump but HK-9(b) forbids justifying a larger exit on the 4wk view. Est 1wk +0.6% / 4wk +2.4% / 12wk +6.9%, all inside the noise band; T12_Progress 93.54 = near-degenerate; the 12wk band spans 47-104, a +/-44% spread with no information. PROFIT_ARMED prints but Avg_Cost is blank so arming rests on the degenerate T12 branch — per the standing rule ADR-0012 is honoured on this book ONLY via Gain_ATR >= 3.0 on a real Avg_Cost; NOT a profit-take. Residual stop HELD at 62.78 — live Stop_Price 59.61 would LOWER the risk line, refused under HK-12. Event risk: 09-10 TSMC August revenue, 09-11 Korea 1-10 Sep semi exports and the 09-11 Oracle read-through all touch 0981 within two sessions. Non-execution of the 09-08 card cost HKHK-11 RETIREMENT NOTICE: if unexecuted at the 09-10 open, 0981 is RETIRED to a WATCH line at 62.78 and will NOT be proposed again; re-arms only on a close below 62.78, or a fresh REDUCE at cv >= 65 after a rally back above 69.63. · news |
| 0700 | TENCENT | 434.00 | HOLD | 434.00 | 430.00 | 428.30 (-1.3%) | 430.32 (-0.8%) | 433.96 (-0.0%) | Tencent is flashing a sell signal but sits just above the exit line we committed to, so we hold. The position is small and the move either way is immaterial to the book. OPERATIONAL: STRONG SELL - REDUCE cv65, Sell_Rank #3, adj 69 (+4 RallySell_BearTrend|sell). Retirement line 430.00; close 434.00 is 0.93% above it so the numeric re-arm does NOT fire — NOT re-issued, third consecutive session under HK-11. Also matches the stale overlay's standing 0700 override. Cumulative non-execution since retirement HK (, 442.80 -> 434.00). Position / HK. |
| 7709 | XL2CSOPHYNIX | 44.56 | HOLD | 44.56 | 28.00 | 43.99 (-1.3%) | 48.02 (+7.8%) | 46.60 (+4.6%) | This leveraged ETF printed its first sell signal in nine sessions on the very day it rose 6.9% — our best performer today. We hold. Two separate rules independently say don't sell here, and the last time we cut a leveraged ETF on a weak reversal signal we missed the bounce. OPERATIONAL: STRONG SELL - REDUCE cv64, Sell_Rank #4, adj 68. Retirement line 28.00 sits 59% below the 44.56 close -> re-arm nowhere near firing, NOT re-issued (HK-11). Independently, leveraged-2x-ETF doctrine requires cv >= 70 + 3 prints + price < Stop_Price for a full exit; this is cv64 on a FIRST print. Closed +6.86% (+HK), the book's best mover. Stop_Price 36.11 is non-inverted for the first time in weeks but sits 19% under the market; manual 28.00 remains operative. Position -61.10%, HK. |
| 3750 | CATL | 562.00 | HOLD | 562.00 | 556.49 | 562.34 (+0.1%) | 562.34 (+0.1%) | 567.40 (+1.0%) | CATL held above our exit line and the sell signal that looked alarming yesterday has already faded to neutral at an unchanged price. Holding. OPERATIONAL: HOLD cv61 (was STRONG SELL - REDUCE cv82 on 09-08, its strongest sell print ever). Retirement line 556.49; close 562.00 is 0.99% above -> not re-armed, NOT re-issued, fourth session. This directly answers the 09-08 open question about whether HK-11 needs a conviction-escalation clause: the cv82 was noise and the trigger was right — the clause should be DROPPED, not drafted. Forward view dead: Est_1W/4W +0.1%, Est_12W +1.0%. adj 57 (61 - 4 val on MoS -0.88). |
| 1347 | Hua Hong Semiconductor | 111.60 | HOLD | 111.60 | 105.00 | 117.61 (+5.4%) | 122.77 (+10.0%) | 120.33 (+7.8%) | Hua Hong is drifting with no actionable signal and remains above the level that would trigger an exit. Holding, though it is our most expensive name against intrinsic value. OPERATIONAL: AVOID cv60 flat 4/4, 0.0. WATCH 105.00 held, close 111.60 -> not re-armed. adj 50 (60 - 10 val on MoS -19.43, the deepest negative margin of safety in the book). Position -30.25%, HKEvent risk: 09-10 TSMC revenue, 09-11 Oracle read-through and Korea semi exports all touch it. |
| 0005 | HSBC HDGS(USD0.50) HK O SEAS | 166.00 | HOLD | 166.00 | 163.18 | 166.06 (+0.0%) | 164.85 (-0.7%) | 165.98 (-0.0%) | HSBC recovered to a firm hold reading and is our largest position at 37% of the book. We keep the risk line where it is rather than loosening it. OPERATIONAL: HOLD cv73, recovered from yesterday's NO TRADE cv0. Close 166.00, RSI 59.5, 1.70% above the trail. Live Stop_Price 161.34 would LOWER the 163.18 line -> refused under HK-12, stop HELD at 163.18. PROFIT_ARMED present but Avg_Cost is blank -> arming invalid on this book, no trim. HK (36.9%). Event risk: 09-11 US PPI, 09-14 US CPI, 09-17 FOMC + HKMA all touch it. |
| 0939 | CCB - H SHS | 9.61 | HOLD | 9.61 | 9.35 | 9.60 (-0.1%) | 9.75 (+1.5%) | 9.76 (+1.6%) | CCB is the steadiest holding in the book — a firm hold reading, quietly higher again. No change. OPERATIONAL: HOLD cv66, close 9.61, RSI 63.7. Live Stop_Price 9.14 would LOWER the 9.35 line -> refused under HK-12, stop HELD at 9.35. HKEvent risk: 09-10 China money supply/TSF, 09-21 China LPR fixing. |
| 0823 | LINK REIT | 38.86 | HOLD | 38.86 | 38.70 | 38.79 (-0.2%) | 38.71 (-0.4%) | 38.79 (-0.2%) | Link REIT is sitting a fraction above its stop for a fifth straight session with a flat forward view. We hold, but a decisive break of the line triggers an immediate exit. OPERATIONAL: NO TRADE x3 then AVOID cv63. Close 38.86 vs stop 38.70 = 16 cents / 0.41% clearance — the widest in five sessions (two cents on 09-08) but still thin. Stop_Price 39.97 prints ABOVE the close — inverted, rejected under HK-12; stop HELD at 38.70. Est_12W 38.79 (-0.2%), so a decisive break of 38.70 arms the L2 stop-break + negative-12wk exit on the spot. HK. |
Outcome & track record
Accuracy at the time of this record.
206 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 →
-0.241 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.
49 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.