Hong Kong — End-of-Day Verdict
0 ACQUIRE + 2 DISPOSE + 1 ANTICIPATE. ACQUIRE: NO ACTION - the only bullish actionable row was 3690 MEITUAN (adj 68 = 63 + 0 + 5), declined on no-chase (-1.26 ATR past entry) and a -9.2% 12wk view; 0316 OOIL (adj 66) faded to WATCH intraday. DISPOSE: 1347 HUA HONG SELL (50%) @133.30 market-on-open (ADR-0012 armed + 2 ticks - dead 12wk 135.60 < 135.97 and 5 consecutive NO TRADE; valuation amplify suppressed by G2 on a blank DCF; NoTrade|sell +5; HK); 2888 STANCHART SELL (29%) @234.60 market-on-open (armed + 1 tick - dead 12wk, T12_Progress 103.59%, every forward estimate negative; HK - the displaced arm from 08-10 finally takes a slot). ANTICIPATE: 0992 LENOVO BUY-STOP 2,000 @29.29 exp 08-12 (adj 65 = 68 - 3 + 0; Quality_Pass True, FAIR_VALUE, powerful uptrend, but unproven signal family and a flat-to-negative Est_* view - fills only on strength, dies before the 13-Aug print; HK). : sells free HKagainst HKof contingent buying, on top of HKdry - comfortably. THE BIG NON-ACTION: 0005 HSBC is armed with a live tick and would be a HKtrim, withheld solely because HK-11 retired that exit to a 159.50 watch line that has not fired.
Recommendations
1 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 |
|---|---|---|---|---|---|---|---|---|---|
| 1347 | Hua Hong Grace Semiconductor | 133.30 | SELL 50% | 133.30 | 118.00 | 132.71 (-0.4%) | 144.34 (+8.3%) | 135.60 (+1.7%) | Cutting half of Hua Hong because the position has stopped working and the model no longer sees a reason to own it: the stock is 16.7% below our cost, sits below both its 20-day and 50-day averages with RSI 42, and the engine has refused to trade it on five straight readings. Its own 12-week target (135.60) is now within 2% of the price, so the upside the thesis rested on has gone. We keep half the position because the 4-week estimate is still positive and foundry demand could turn. OPERATIONAL: ADR-0012 profit-take fired - PROFIT_ARMED (T12_Progress 98.3%) + 2 ticks: (c) dead 12wk (135.60 < 1.02x133.30=135.97) and (d) 5 consecutive NO TRADE prints (T0..T4). Tier = profit_take_2 (50%) on 2+ ticks. Valuation amplify SUPPRESSED by G2: MoS -17.57 but DCF_per_share is blank and Quality_Pass False - unreliable anchor, so no bump. Scorecard cell NoTrade|sell +5 (n=7, hit_1w.75, ACTIONABLE). HK-9: treat as a 1-week de-risk, not a 4-week thesis - Est_4W +8.3% is the acknowledged cost. Board lot 500 assumed - 1500 =; verify before placing. Remaining 1500 trail 118.00 (2.0xATR chandelier 113.02 is looser - not lowered). L8: executes at the 08-12 open. · news |
| 2888 | STANCHART | 234.60 | SELL 28.7% | 234.60 | 224.83 | 236.08 (+0.6%) | 232.44 (-0.9%) | 226.48 (-3.5%) | Taking a small profit off Standard Chartered after a strong run. The shares have pushed 3.6% above the model's own 12-week fair target and every forward estimate now points slightly down (-0.9% at 4 weeks, -3.5% at 12 weeks) even though the recent news - a 10% first-quarter profit rise and a new institutional crypto partnership with FalconX - is constructive. We are trimming roughly a quarter of the holding and keeping the rest, letting the balance run with a raised stop. OPERATIONAL: ADR-0012 profit-take fired - PROFIT_ARMED (T12_Progress 103.59%) + 1 tick: (c) dead 12wk (226.48 < 1.02x234.60=239.29). Tier = profit_take_1 (30%); 348 x 30% = on the 50-share board lot (assumed; verify). Tick (b) NOT counted - conviction is -19, one point short of the 20-pt bar. Valuation amplify SUPPRESSED: NA_CCY row, and the -7.54 MoS is internally inconsistent (DCF 243.54 sits ABOVE the 234.60 close) - currency-mismatch artifact, val_adj 0. Cell FailedBreakdown_Up_20D|sell is report-only (n=4) - no bias applied. Was the DISPLACED FIRED ARM on 08-10; takes a slot today. Remaining 248 trail 224.83. L8: executes at the 08-12 open. · news |
| 0992 | Lenovo Group Limited | 28.54 | BUY-STOP — QUEUED @29.29 | 29.29 | 24.33 | 29.31 (+0.1%) | 28.97 (-1.1%) | 28.40 (-3.0%) | Buying Lenovo only if it breaks higher. The stock is in a powerful uptrend - well above its 20-, 50- and 200-day averages - passes our quality screen, and screens as fairly rather than expensively valued, so we want exposure; but it has already run 14% above its one-month average, so we will not chase it here. Instead we place a buy order that only triggers if the shares push through HK, with a stop at HK. The order is deliberately cancelled before Lenovo's 13-Aug quarterly results so we are not holding a live order into an unknown print. OPERATIONAL: ANTICIPATE slot (ADR-0012), 4th slot - separate from the <=3 ACQUIRE. Eligible: BUY - ANTICIPATE printed in 4 of the last 4 snapshots (>= required); adj 65 = 68 + (-3 val_adj, MoS -0.605 x 5) + 0 bias >= 55 floor. Quality_Pass True. SIZE DEVIATION FLAGGED: slot calls for 0.5x = HKbut Lenovo's 2,000-share board lot (assumed; verify) forces one lot = HK = 1.0x - taken at one lot rather than skipped, not silently up-sized. AnticipationUp_VCP has NO graded buy cell (|sell n=1, report-only) - unproven family, bias 0. Est_* forward view is flat-to-negative (-3.0% at 12wk) and diverges sharply from the model Target_Price 39.19 (+33.8%); the wide 4wk band [17.94, 46.60] is near-uninformative - flagged, not resolved. SUPERSEDES the 08-10 GTC @28.57, which TRIGGERED today (low 28.36) and was not executed: an execution miss of only -HKvs the 28.54 close. Repriced to the live 29.29 buy-stop per the provenance rule - not rolled. L6: expires 08-12, before the 13-Aug print. · news |
| 0939 | CCB - H SHS | 8.82 | HOLD | — | 8.46 | 8.82 | 8.80 | 8.71 | Dropping the repeated plan to trim CCB. We have proposed this sale three sessions running and it has not been acted on; rather than reissue it a fourth time we are retiring it to a watch line and will only revisit on fresh evidence. The bank still holds above its moving averages. OPERATIONAL: HK-11 escalation complete - proposal #3 (08-07 25,000 @8.79; 08-10 15,000 @8.855, already market-at-open) unexecuted; retire, never re-issue. RE-ARM TRIGGER: a fresh SELL - REDUCE print, OR a decisive close below 8.46. The 08-10 MOO sell would have filled ~8.89 vs the 8.825 close - inaction COST HKStop raised 8.43 -> 8.46 (model). Still PROFIT_ARMED with tick (c) live - the arm is acknowledged and deliberately not acted on under HK-11. |
| 3750 | CATL | 651.50 | HOLD | — | 625.50 | 648.54 | 636.80 | 616.45 | Dropping the repeated plan to sell down CATL. This trim has been proposed on three separate sessions without being executed, and meanwhile the stock has recovered - the model has moved it back from no-trade to hold and conviction is rising. We retire the sale to a watch line and keep the full. OPERATIONAL: HK-11 escalation complete - 07-28, 08-07, 08-10 all unexecuted; retire, never re-issue. RE-ARM TRIGGER: a decisive close below 625.50, OR a fresh SELL - REDUCE print. The 08-10 MOO sell would have filled 641.00 vs the 651.50 close - inaction SAVED HKthe largest single item this session. Signal improving NO TRADE(0) -> HOLD(63) across the day supports retiring rather than escalating. Stop held at 625.50 (model 596.51 and the 2.0xATR chandelier 596.50 are both looser - not lowered). |
| 7709 | XL2CSOPHYNIX | 28.10 | HOLD | — | 28.00 | 29.12 | 29.40 | 29.67 | Dropping the repeated plan to halve this leveraged SK Hynix ETF. We have proposed the cut four times without execution and the product has since stabilised - all three forward estimates now sit above the current price. We retire the sale to a watch line at HK, just below the close, so any renewed breakdown re-triggers it immediately. OPERATIONAL: HK-11 escalation complete - 07-27, 07-28, 08-06, 08-10 all unexecuted; retire, never re-issue. RE-ARM TRIGGER: a decisive close below 28.00 (the close is 28.10, only 0.4% above - expect the monitor to test this line quickly; that fire IS the re-arm). The 08-10 MOO sell would have filled 27.54 vs the 28.10 close - inaction SAVED HKHK-12: both the model Stop_Price 39.95 (42% ABOVE the close) and the 2.0xATR chandelier (11.26, -60%) are degenerate and REJECTED; 28.00 is a hand-set percentage line. Model Target_Price is -19.38, negative and unusable. Position is -75.5% from cost. |
| 0005 | HSBC HDGS | 161.70 | HOLD | — | 159.50 | 160.98 | 161.31 | 162.77 | Holding all 12,000 HSBC shares. The forward view is flat - roughly HK-163 across one, four and twelve weeks against a HKclose - which is uninspiring but not a reason to sell the book's largest position, and the shares go ex-dividend on 14-Aug. We keep the existing HKwatch line and will only revisit the trim if that breaks. OPERATIONAL: PROFIT_ARMED (T12_Progress 99.34%) with tick (c) live (162.77 < 1.02x161.70 = 164.93), which would fire a 30% trim = HKthe largest candidate in the book. NOT TAKEN: this exit was already escalated and RETIRED to a WATCH line on 08-06 with re-arm 'close below 159.50'; the close is 161.70, so the re-arm has NOT fired and HK-11 forbids re-issuing it. Model stop 156.17 is looser than the 159.50 watch line - not lowered. 14-Aug ex-dividend (second interim) is a MECHANICAL gap and must NOT be read as a stop break by the monitor. |
| 0700 | TENCENT | 470.80 | HOLD | — | 466.94 | 466.55 | 466.48 | 490.19 | Holding Tencent through tomorrow's results. The engine has gone neutral on the stock, but second-quarter numbers land after the 12-Aug close - one session away - and we do not trim into an earnings print. The 12-week view is still the most positive in the book at +4.1%. OPERATIONAL: PROFIT_ARMED (T12_Progress 96.04%) + tick (d) (4 consecutive NO TRADE) fires a 30% profit-take, OVERRIDDEN by the catalyst-aware hold - earnings <=5 trading days, so HOLD/trail through the event; trim only on a confirmed lower-high or decisive stop break. Tick (c) does NOT fire (490.19 > 1.02x470.80 = 480.22). Stop held at 466.94, but it sits only 0.8% below the close: a post-earnings GAP through it is NOT an automatic exit - require a confirmed close below on the session AFTER the print. Avg_Cost 1.00 in the portfolio JSON is a data artifact, so /Gain_ATR figures on this row are meaningless. |
| 2513 | KNOWLEDGE ATLAS | 1192.00 | HOLD | — | 975.86 | 1189.74 | 1164.02 | 1249.79 | Holding Knowledge Atlas and raising the stop sharply. The stock has rallied hard - it traded as high as HKtoday against a HKclose at the end of July - and is a candidate for inclusion in the Hang Seng index at the 21-Aug quarterly review, so we ride it through the event rather than trimming into it. We lift the protective stop from HKto HKto lock in more of the move. OPERATIONAL: PROFIT_ARMED (T12_Progress 95.38%) but NO tick fired - (c) does not fire (1249.79 > 1.02x1192 = 1215.84), conviction is -13 (< 20), label is AVOID not NO TRADE. Armed-but-not-fired => MANDATORY trail-raise, not a trim. New stop = 2.0xATR chandelier off today's 1298 high (1298 - 2x161.07). HK-12: model Stop_Price 1514.13 is 27% ABOVE the close - degenerate, REJECTED. Catalyst-aware hold also applies (HSI review <=5 sessions). Position -27.1% from the 1635 cost - see HK-9 on why it was re-entered there. |
| 0981 | SMIC | 65.30 | HOLD | — | 58.86 | 65.95 | 67.09 | 68.62 | Holding SMIC. The repeated trim was retired to a watch line on 08-10 and the trigger has not been hit - the shares closed at HK, comfortably above the HKline - and all three forward estimates are modestly higher than the current price. No action. OPERATIONAL: HK-11 retirement holds; re-arm = a fresh SELL - REDUCE print OR a decisive close below 58.86. PROFIT_ARMED but tick (c) does NOT fire (68.62 > 1.02x65.30 = 66.61) and conviction is flat at 60 across all five prints. HK-12: model Stop_Price 72.18 is ABOVE the 65.30 close - degenerate, REJECTED; 58.86 retained. |
| 0522 | ASMPT | 166.20 | HOLD | — | 145.50 | 169.46 | 180.85 | 190.49 | Holding ASMPT. It is the one position the profit-take framework does NOT flag, and the forward estimates are the most positive in the book - roughly +9% at four weeks and +15% at twelve against a HKclose - so despite being 15.4% underwater there is no case to cut here. Stop unchanged at HK. OPERATIONAL: NOT PROFIT_ARMED - T12_Progress 87.25% is below the 90% arm threshold and Gain_ATR is -2.86, so no ADR-0012 candidacy. 5 consecutive NO TRADE prints noted but there is no arm to fire against. 2.0xATR chandelier 145.14 is looser than 145.50 - not lowered. |
| 3690 | MEITUAN-W | 93.05 | WATCH | 89.04 | 85.44 | 92.41 (+3.8%) | 91.55 (+2.8%) | 84.45 (-5.2%) | Passing on Meituan despite it ranking first on our buy screen. Two problems: the shares have already run 4.5% past the price we wanted to pay, so buying here is chasing; and the model's own twelve-week view is 9% BELOW today's close, meaning the projected path is down. We would rather miss it than pay up for a name whose forward estimate is negative. OPERATIONAL: adj 68 = 63 + 0 (Val_Signal NA) + 5 (PullbackBuy_VWAP|buy, n=11, hit_1w.75, ACTIONABLE) - clears the 65 floor but fails the ACT-NOW bar on Entry_Distance_ATR -1.26 (no-chase) and on HK learnings section B (negative-12wk buys are not buyable: Est_12W 84.45 is -5.2% even against the 89.04 entry). Quality_Pass is False, but Val_Signal 'NA' makes the hard quality gate AMBIGUOUS - 3rd consecutive run; logged as a systemic finding rather than used as the deciding reason, since the two grounds above are sufficient. Deliberately NOT queued at 89.04. |
| 0316 | OOIL | 149.50 | WATCH | 149.50 | 148.20 | 150.95 (+1.0%) | 154.08 (+3.1%) | 170.87 (+14.3%) | Passing on OOIL again. It was the strongest name in the book at yesterday's close but faded through today's session from a strong buy to a watch rating, so there is no live buy trigger to act on. It remains the most attractively valued name we track and stays on the watchlist. OPERATIONAL: adj 66 = 63 + 3 (UNDERVALUED, MoS +0.299 -> +6 halved because Quality_Pass is False) + 0; clears the floor on paper but the anchor Action is WATCH, not a bullish actionable label, so it is not in the buy pool. 4th consecutive session blocked or faded. Action trail today: STRONG BUY(86) -> BUY(60) -> BUY(63) -> WATCH(63) -> WATCH(63). RE-ARM: a fresh BUY/STRONG BUY print at the settled close with Entry_Distance_ATR > -0.25. |
Outcome & track record
Accuracy at the time of this record.
157 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.