HK Financials — End-of-Day Verdict
1 DISPOSE x 0 ACQUIRE - TRIM 3968 by 50% at the 08-13 open
Recommendations
0 to acquire · 1 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 |
|---|---|---|---|---|---|---|---|---|---|
| 3968 | China Merchants Bank | 48.02 | SELL 50% | 48.02 | 48.10⚠ | 47.99 (-0.1%) | 48.20 (+0.4%) | 49.04 (+2.1%) | We are selling half our China Merchants Bank position. The stock closed at HK, just below the HKstop-loss we set to protect the position -- the first time any holding in this book has broken its risk line since we started -- and for a second straight day the model can find no coherent case for owning it, reading the evidence as outright conflicting rather than merely weak. We committed in writing yesterday to cut on exactly this second signal failure, and we are honouring that today rather than waiting for confirmation that may never be cheap. This is not a verdict on the bank: it came through today's fundamental refresh as the highest-quality company we own, and the three-month projection is still modestly positive. It is a verdict on carrying full size in a name whose signal has gone dark and whose stop has been taken out. We keep the other half, because the break was wafer-thin and China's July credit data lands in two sessions and could resolve this either way. OPERATIONAL: signal-degradation trim, explicitly NOT a profit-take. Two triggers, both pre-committed on 08-11: (1) ADR-0012 tick (d) / CLAUDE.md §9 'sustained NoTrade' -- 2nd consecutive NO TRADE print, cv0, 'Consensus conflict + low conviction' verbatim; (2) stop breach, close 48.02 vs recorded stop 48.10 (-0.17% / -0.085 ATR), the book's first live breach since inception. HF-5 checked and DISMISSED -- no ex-dividend event on the horizon calendar and HKis not a dividend adjustment, so the breach is real deterioration. L2 is NOT the basis: its 12wk leg fails (Target_12Week 49.04 is +2.12% ABOVE close). SIZING: pre-committed 30% x 2,000 =, untradeable at 3968's 500-sh board lot; adjacent lots are 500 (25%) and 1,000 (50%); ROUNDED UP to 1,000 on the tie-breaker that the same close broke the stop -- a broken risk line is not a round-down situation. size_mod 1.0x (NA_CCY, no valuation tier scaling either way). Proceeds HK, realised -HKvs the HKcost. PROFIT_ARMED printed and is SUPPRESSED on the ADR-0012 arm criterion itself: Gain_ATR -0.67 (position at a LOSS, deteriorated from -0.30), T12_Progress 97.92 is merely Close/Target_12Week; true progress-from-cost is (48.02-48.65)/(49.04-48.65) = -162%. Neither arm limb met -- booking this as a profit-take would be incoherent. NOT 100%: the pre-commitment reserved a full exit for a DECISIVE break and 0.085 ATR is inside noise (the 47.78 intraday low recovered into the close); the optional_trim counterfactual (n=45, +2.24% median saved @1wk) says this book over-cuts; and Est_1W/4W/12W are flat-to-positive (-0.06% / +0.37% / +2.12%), so there is no bearish forward view, only an absent bullish one. ACTION PRICE PROVENANCE: NO TRADE row carries a blank Entry_Price, so the reference is the anchor's settled close 48.02, range 47.78-48.26 (+-0.25 ATR); Target_Price n/a on a NO TRADE row. L8: executes at the 2026-08-13 open, reprice if the open gaps >~1.5% (outside 47.30-48.74) -- US CPI tonight makes a gap live. Position context: below EMA10 48.85 / EMA20 48.61 / EMA200 48.29, sitting on EMA50 48.02, -2.12% vs VWAP; pillars Trend 36 / Momentum 32 / Volatility 40, consensus Indeterminate, regime Transitional; three consecutive lower highs and lower closes (49.00 -> 48.34 -> 48.02). Catalyst-aware hold does NOT block: the 08-14 China credit print is a macro read-through not a company event, and the rule's escape hatch (confirmed lower-high sequence PLUS a stop-loss break) is met. ADR-0020: stop 48.10 vs peak_stop 48.10 = 0% erosion -- the good case, cut while the stop is at its peak. Valuation NA_CCY -> val_adj 0; per CLAUDE.md §9 valuation could not have originated this sell and did not. Scorecard bias 0 (NoTrade|sell n=1, report-only). RESIDUAL HARD RULES from 08-13: stop stays 48.10 (chandelier 46.14, far below; ratchet never lowered); FULL EXIT of the residual on ANY close below 48.10 -- the 'decisive' qualifier is SPENT now the first breach is on the tape; re-arm to hold only if conviction recovers to >=45 with a non-NO-TRADE label AND the close reclaims 48.10; do NOT re-add on a bounce alone. News: none found 08-11 -> 08-12 on Yahoo Finance / Seeking Alpha (search evidence only -- direct page fetch was unavailable in this unattended run, so this is not a confirmed clean sweep). · news |
| 2888 | STANCHART | 233.80 | HOLD | — | 223.61 | 235.37 | 232.24 | 225.69 | We are holding Standard Chartered. On a day when the book cut half its other position, this one moved the other way: the strength of the signal RECOVERED after yesterday's slide, the stock is up 2.3% for us, and it trades HKclear of our HKprotective stop with the healthiest chart in the book -- above every moving average we track. Every one of the four conditions we wrote down yesterday for trimming it went unmet. US inflation data tonight is the near-term risk, and the position has enough cushion to absorb it. OPERATIONAL: all four 08-11 pre-commitments checked, NONE fired -- (1) conviction ROSE, away from the <=20 trim trigger not toward it; (2) no non-PROFIT_ARMED Exit_Warning; (3) no NO TRADE print, so the two-consecutive counter never started; (4) no formal SELL/REDUCE. PROFIT_ARMED again SUPPRESSED under ACTIVE HF-2, 13th consecutive session: the arm comes purely from an inverted 12-week target (225.69 < close 233.80, -3.47%) driving T12_Progress to 103.59, while Gain_ATR is 0.99 against a 3.0 threshold. ADR-0012 ticks: (a) none; (b) conviction rose; (c) reads fired but is mathematically implied by the same inversion that created the arm (systemic finding #2); (d) no NO TRADE. L2 fails its stop-break leg (233.80 vs 223.61, 1.96 ATR cushion). TRAIL: no raise -- the 2.0-ATR chandelier computes to 223.39, HKBELOW the 223.61 ratchet floor, and the ratchet is never lowered; next raise when close > ~234.02. ADR-0020: stop 223.61 vs peak_stop 223.61 = 0% erosion. Stop-risk HK (0.56% of book). Pillars Trend 77 / Momentum 32 / Volatility 40, consensus Consolidation_Bullish; above EMA10 233.15 / EMA20 229.92 / EMA50 220.43 / EMA200 191.58 and +1.90% vs VWAP. Forward curve is up-then-down (1wk +0.67%, 4wk -0.67%, 12wk -3.47%) -- which is why this is a hold, not an add. Valuation NA_CCY -> val_adj 0. US July CPI 08-12 20:30 HKT is tagged to 2888 and lands on the 08-13 session; HK Q2 GDP final ~08-14 and FOMC minutes 08-19 also tagged. News: no dated 08-11/08-12 headline; background only -- the HKdividend went ex 2026-08-06 (past), Q1 profit +10% ahead of forecasts, FalconX institutional-crypto partnership, new financial-sponsors coverage team. PRE-COMMITMENT next session (re-based to cv46): trim 30% if conviction falls >=20 from 46 (i.e. <=26), OR any non-PROFIT_ARMED Exit_Warning prints, OR two consecutive NO TRADE, OR a fresh formal SELL/REDUCE at cv>=65; full exit on a decisive close below 223.61; raise the stop the moment the chandelier clears 223.61 (close > ~234.02). · news |
| 2388 | BOC Hong Kong Holdings | 49.82 | NO ACTION | 50.16 | 49.18 | 49.94 (-0.4%) | 52.64 (+4.9%) | 49.91 (-0.5%) | BOC Hong Kong stopped being a buy candidate today. The signal that had it as the book's strongest name yesterday withdrew outright, with the model's confidence falling further than any other stock in the portfolio. The irony is that the pullback we said we wanted finally arrived -- the price is now below the level we would have paid -- but the case for buying left with it. The bank still screens expensive against its own book value, still fails our quality test, and its three-month projection is still statistically identical to today's price. Nothing to do. OPERATIONAL: label is WATCH, not a buy print -- 2388 never enters the buy pool. Conviction, the largest single-session decay in the book. Even if it were actionable: 46 + val_adj -1 + bias 0 = adjusted 45 vs the 65 floor, short by 20; Quality_Pass FALSE after today's fairvalue refresh (F-Score 4) is a HARD gate on a new entry; Target_12Week 49.91 vs close 49.82 = +0.18%, a rounding error, HF-3's spirit. The 08-11 re-arm trigger is UNMET on three of four legs -- 12wk needs > +2% (is +0.18%), raw conviction needs >= 66 (is 46), Quality_Pass needs TRUE (is FALSE); only the fourth, Entry_Distance_ATR > -0.25, is now satisfied (+0.35, no longer chased), which is exactly the wrong leg to have cleared. Fifth consecutive session withheld. RE-ARM unchanged. Scorecard cell PullbackBuy_VWAP|buy is n=2 with ZERO graded outcomes -- L1 favours this family but this book still has no evidence either way, so no bias applied. News: none found. · news |
| 0388 | HKEX | 406.60 | NO ACTION | 406.60 | 400.60 | 407.55 (+0.2%) | 421.28 (+3.6%) | 420.26 (+3.4%) | HKEX was one of only two stocks in the book to rise today, but its buy signal withdrew and we remain on the sidelines. The three-month view is genuinely attractive at roughly +3.4%, and this is the name we are tracking most closely -- but the exchange still screens as expensively priced against what its own business earns, it still fails our fundamental quality screen, and it now reports interim results in five trading sessions. Buying a quality-gate failure directly into an earnings print is not a risk-defined entry at any price. We wait for the results to pass. OPERATIONAL: label reverted to WATCH -- no buy print, so not in the pool. Signal rotated PullbackBuy_VWAP -> FailedBreakdown_Up_20D, i.e. from the L1-favoured family to the L1-disfavoured one. Conviction; 57 + val_adj -6 + bias 0 = adjusted 51 vs the 65 floor, short by 14. Quality_Pass FALSE after today's refresh (F-Score 4, MoS_FV -1.111, OVERVALUED_VS_RI against the exchange's own residual-income anchor) is a HARD gate on a new entry. HF-3 does NOT fire -- Target_12Week 420.26 > close 406.60 (+3.36%). Interim results + interim dividend 2026-08-19 are now 5 sessions out, INSIDE the catalyst window, which independently bars queueing an entry. RE-ARM TRIGGER unchanged (all four, on a settled close): raw conviction >= 71 AND Quality_Pass TRUE at the next fairvalue refresh AND Target_12Week > Close AND the 08-19 results PAST, not pending. Two of the four legs continue to trend the right way -- keep tracking. News: none found. · news |
Outcome & track record
Accuracy at the time of this record.
6 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 →
0.184 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.