Hong Kong — End-of-Day Verdict
3 DISPOSE + 3 ACQUIRE (queued). Sells free HK (HSBC reduce 1.65M + 2513 full exit 1.18M + 7709 half-trim 50k); buys need HKif all three pullback limits fill - strongly. 0 stop breaches; monitor clean; 2513's 07-21 pending exit never filled (re-issued). HK (unchanged - all pending next open).
Recommendations
3 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 |
|---|---|---|---|---|---|---|---|---|---|
| 0005 | HSBC HDGS | 158.90 | SELL 85.46% | 158.90 | 153.73 | 157.27 (-1.0%) | 160.95 (+1.3%) | 164.28 (+3.4%) | The settled close flipped HSBC to a Strong Sell - Reduce at Sell-Rank 1, conviction 70, on a reversal-from-overbought signal with RSI at 71 after a strong run. That is a signal-based reason to bank the bulk of a bank position that has stretched, so the model's 85% de-risk stands. Valuation is currency-blind here (NA_CCY) so it neither originates nor amplifies. Operational nuance: this is a REVERSAL sell, not a profit-take, so the trim-discipline gates (G1-G3) do not apply - they guard momentum breakouts, and ReversalDown_Overbought is the valid-exit contrast case. Honest caveat: the model's own forward view is only mildly positive (12wk +3.4%), so this is a risk-management reduce on the reversal + RSI extension, not a dead-view exit; the specific ReversalDown cell is report-only (n=1) but the broader HK SELL family is reliable (HK-3). L4: acted on the settled close, not an intraday whipsaw. L8: executes MARKET at the next open 07-23; reprice if the open gaps >~1.5%. Reduce (HSBC 400-lot), leaving 1,600 on a 153.73 stop. · news |
| 2513 | Knowledge Atlas | 1175.00 | SELL ALL | 1175.00 | — | 1213.53 (+3.3%) | 1164.04 (-0.9%) | 1236.47 (+5.2%) | Re-issue of the standing full exit: yesterday's pending exit at ~1219 never filled (no trade recorded, still held), and the name has faded to 1,175. It prints a bearish FailedBreakout_Down_20D for a third straight session on the book's actionable down-trend sell cell (bias +2), sits ~28% underwater against a 1,635 cost, and has zero fair-value coverage to lean on - there is no floor and no forward case (4wk -0.9%, and the +5.2% 12wk sits inside a -18%/+72% band, i.e. noise). Operational nuance: L3/HK-5 - this de-risk has been flagged for six sessions; execute it promptly at MARKET rather than re-recommend indefinitely. The label softened from Strong Sell-Reduce to AVOID but the exit thesis (persistent bearish, no valuation floor, deeply underwater) is unchanged. Extreme variance: >15% intraday moves on four of the last five sessions - stay MARKET and expect slippage. L7: check hktech before executing - execute once. Frees HKthe book's largest single risk line. · news |
| 7709 | XL2CSOP HYNIX 2x | 55.76 | SELL 50% | 55.76 | 44.00 | 58.46 (+4.8%) | 55.22 (-1.0%) | 56.45 (+1.2%) | De-risk a deeply underwater (-51% vs 114.56 cost) 2x leveraged ETF that has printed a bearish breakdown-family signal for a third straight session, conviction 61, with a dead forward view (4wk -1.0%, 12wk +1.2%) - the underlying's SK Hynix catalyst has now passed and the ETF fell into it. Operational nuance: the leveraged-ETF doctrine says trim 50% on a low-conviction (cv<70) reversal/breakdown rather than full-dump at an oversold low (RSI 38.7) - the 7747 full-exit that missed a +5.8% bounce is the lesson this respects. So bank half (, ETF 100-lot) and hold 900 on the manual 44.00 trail (unbreached; the scan's 84.35 stop is above spot and invalid for a long). Escalates to full exit only on conv>=70 + price below the manual trail. L8: MARKET at the 07-23 open. Frees HK. · news |
| 3690 | Meituan | 83.65 | BUY — QUEUED @77.22 | 77.22 | 72.61 | 83.76 (+8.5%) | 82.47 (+6.8%) | 82.66 (+7.0%) | Top-ranked acquire: a fresh PullbackBuy_VWAP BUY at conviction 73 - the book's single most reliable buy cell (HK-2, bias +3) and the one setup the learnings say to prefer over any breakout/momentum chase. Adjusted conviction 76 clears the floor comfortably. Valuation is blind (NA) so it neither helps nor hurts and the quality gate is off. Operational nuance: the model entry 77.22 is ~8% below the 83.65 close, so this is a no-chase GTC buy-limit that fills only on a pullback to VWAP - queued, not bought at the close (HK-4: no-chase stays strict). Size (100-lot) ~HK at the book's 20% new-buy weight. Expires in 3 sessions (07-27) or on conviction falling below 65; stop 72.61. · news |
| 0857 | PetroChina H | 10.05 | BUY — QUEUED @9.17 | 9.17 | 8.76 | 10.03 (+9.4%) | 10.10 (+10.1%) | 10.26 (+11.9%) | Highest adjusted conviction in the book (78) and a re-arm of the queue withdrawn on 07-21: PetroChina prints PullbackBuy_VWAP BUY again at conviction 70, and it is the book's best-qualified name on valuation - UNDERVALUED with Quality_Pass True (MoS +0.236), which adds +5 and unlocks up to 1.25x sizing. Operational nuance: the model entry 9.17 is ~9% below the 10.05 close, a genuine no-chase pullback limit repriced to the live row (not the withdrawn 9.10). Sized up 1.25x on the valuation edge = (2,000-lot) ~HK. HK-2 anchor confirmed. Expires 07-27 or below cv 65; stop 8.76. · news |
| 0992 | Lenovo | 24.10 | BUY — QUEUED @22.98 | 22.98 | 21.08 | 24.03 (+4.6%) | 25.27 (+10.0%) | 26.40 (+14.9%) | Third acquire: a PullbackBuy_VWAP BUY at conviction 66, adjusted 68 - above the floor and the third of three fresh prints on the book's reliable buy cell (HK-2, bias +3). FAIR_VALUE with Quality_Pass True, a small -1 valuation trim (MoS -0.25) but the DCF (43.5) sits well above spot so the quality gate passes and sizing stays 1.0x. Operational nuance: entry 22.98 is ~5% below the 24.10 close - the tightest of the three pullback limits, so the most likely to fill. (2,000-lot) ~HK. The 12wk estimate is the strongest of the three buys at +14.9%. Expires 07-27 or below cv 65; stop 21.08. · news |
| 0939 | CCB H | 8.55 | HOLD | — | 8.13 | n/a | 8.97 | n/a | Held and strengthening - conviction climbed across the session on a FailedBreakdown_Up recovery, with the close (8.55) back above a rising 8.13 stop. No exit case and no fresh-add case (still below the add threshold), so hold and let the stop trail up. The book's largest share line by count; nothing to do today. · news |
| 2888 | StanChart | 223.60 | HOLD | — | 218.00 | n/a | 234.76 | n/a | Held and firming - conviction into the close (223.60) on a FailedBreakdown_Up, just below the add threshold. No exit case; hold on the 218 trail into H1 results at month-end. Valuation is currency-blind (NA_CCY). Nothing to act on today. · news |
| 0522 | ASMPT | 169.10 | HOLD | — | 156.50 | n/a | 178.15 | 200.98 | Catalyst-aware hold into its own earnings, now estimated for the very next session. No-trade for a fourth straight snapshot, ~-4.7% vs cost, but the 12wk view (200.98, ~+19%) is the most positive in the book and valuation can never originate a sell. Operational nuance: no pre-trim into a binary one session away; the model's stance is add-after-print, not before. Hold on the 156.50 trail (just under recent session lows); only a confirmed lower high or a decisive stop break changes this before the numbers land. · news |
| 1347 | Hua Hong | 161.50 | HOLD | — | 145.00 | n/a | n/a | 184.55 | Held around break-even (~+0.9% vs cost) on repeated no-trade prints - no entry case and no exit case. The spurious PROFIT_ARMED flag persists (systemic finding #2: T12_Progress is Close/Target_12Week, not progress from cost), so the profit-take path stays suppressed - the real gain is a fraction of an ATR, not a harvestable winner. Hold on the 145.00 trail; a fresh breakdown-family SELL print or the 12wk estimate turning negative means dispose immediately. · news |
| 3750 | CATL | 609.50 | HOLD | — | 488.81 | n/a | n/a | n/a | Catalyst-aware hold into H1 results in ~2 sessions. No-trade, ~-1.9% vs cost, no signal either way. FAIR_VALUE on valuation (cannot originate a sell). Hold on the 488.81 stop through the print; no pre-trim into the binary. · news |
Outcome & track record
Accuracy at the time of this record.
112 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 →
-0.283 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.
21 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.