Recommendations
0 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 |
|---|---|---|---|---|---|---|---|---|---|
| 3988 | Bank of China | 5.97 | SELL 30% | 5.97 | 5.63 | 6.05 (+1.3%) | 5.98 (+0.1%) | 5.93 (-0.8%) | Bank of China has done its work - the position is up 9.8% and the shares are now stretched, with RSI at 78 and the price already past the level the model itself would buy at. Decisively, the engine's own twelve-week projection has dropped BELOW today's price: it sees 5.93 against a 5.975 close. We bank 30% of a genuine gain into that strength and keep 70% of the position. Valuation is silent here (the H-share reports in RMB against an HKD quote), so this rests on the signal and the realised profit. OPERATIONAL: this is a true ADR-0012 profit-take, not the inverted PROFIT_ARMED artifact affecting hk and sg - the arm is Gain_ATR 3.57 >= 3.0 computed off a REAL Avg_Cost of 5.44. Fired by tick (c), dead forward view: Target_12Week 5.93 < 1.02 x 5.975. G1 softening applies to tick (a) only and does NOT block (b)/(c)/(d); G2/G3 govern the valuation-elevate path, not in play at val_adj 0. One tick -> profit_take_1 = 30%, no valuation amplify, never 100% from a profit-take. Proceeds HKrealised +HKResidual, stop raised 5.60 -> 5.63. THE STRONG BUY - ADD ON THE SAME ROW IS DECLINED: Entry_Distance_ATR -0.30 puts the close past the model entry and RSI 77.97 is above the 75 extension line - you do not add to a name whose 12-week view just went negative. L8 gap-guard asymmetric on a de-risk: sell into a gap up (better fill) or a gap down (more urgent), reprice to the actual open. · news |
| 3328 | Bank of Communications | 7.88 | SELL 43% | 7.88 | 7.48 | 7.91 (+0.4%) | 7.92 (+0.5%) | 7.90 (+0.3%) | Bank of Communications printed the single loudest sell signal in the book tonight - top-ranked, maximum conviction - having been an equally emphatic buy 24 hours earlier. Behind the whipsaw sits a simple fact: RSI is at 73 and the model's twelve-week projection offers just +0.25% from here. The bank re-rating has been paid for. We take 43% off at a 5.1% profit and keep the rest. Valuation is silent (currency-mismatched statements). OPERATIONAL: this is a LABEL-originated REDUCE, not a profit-take - Gain_ATR 1.91 is below the 3.0 ADR-0012 arm. Engine was 84.59%; scaled to the 50% tier and lot-rounded to 3,000 because the label is a SINGLE print that reversed a cv90 buy the previous session, and the originating cell ReversalDown_Overbought|sell is n=1, report-only - no evidence base for the engine's size. Proceeds HKrealised +HKagainst cost 7.50. Residual, stop raised 7.42 -> 7.48. Supported by this book's optional_trim counterfactual: n=68, median +5.87%/4wk 'saved' - the largest such sample in the HK family, and it says the trims this book declines would have helped. |
| 0939 | CCB - H SHS | 9.60 | SELL 33% | 9.60 | 9.16 | 9.64 (+0.4%) | 9.78 (+1.9%) | 9.75 (+1.6%) | China Construction Bank is the third member of the same overbought bank sleeve to flip to a sell tonight - RSI 71.6, second-ranked exit, stable across three consecutive prints - and like its peers the model's twelve-week view has flattened to +1.6%, which is not enough forward return to justify carrying full size through it. We take a third off at a 4.3% profit and keep two-thirds for the re-rating. Valuation is silent (currency-mismatched statements). OPERATIONAL: label-originated REDUCE, not a profit-take (Gain_ATR 1.81, below the 3.0 arm). Engine 84.59% cut to the 30% tier, matching the tier the parent hk book applies to the same name this session. Proceeds HKrealised +HKagainst cost 9.20. Residual, stop raised 8.96 -> 9.16. L7 CROSS-BOOK: the parent hk book trims 0939 the same session ( @9.60) - separate share ledgers, same security, / HKfamily-wide. Execute both or neither; do not net silently. |
| 3968 | China Merchants Bank | 51.45 | NO ACTION | 51.40 | 49.01 | 51.30 (-0.2%) | 51.80 (+0.8%) | 51.93 (+1.0%) | China Merchants Bank is the one bank in the book that still looks buyable on the tape - not extended, the coolest RSI in the sleeve at 69, and a conviction that clears our bar comfortably. We are still not adding, because the model's own forward view for it is the same flat reading that is causing us to sell three of its peers tonight: +0.8% over four weeks, +1.0% over twelve. Adding to a sector in the same breath as trimming three of its members, on identical evidence, is incoherent. We hold what we have and lift the risk line. OPERATIONAL: adj 74 CLEARS the 65 floor - this is a coherence decline, not a floor failure. Entry_Distance_ATR -0.04 (not extended), val_adj 0 (NA_CCY). Intra-row contradiction flagged: Target_Price 57.96 (+12.8%) vs Est_4W 51.80 (+0.8%) on the same row and horizon. Stop HELD at 49.01, above the 48.85 model level - never lowered. RE-ARM: a BUY - ADD reprint with Est_4W >= +4%. Ledgered optional/direction:none so the counterfactual grader prices the decline. |
| 0966 | China Taiping Insurance | 21.96 | NO ACTION | 22.25 | 20.13 | 21.98 (-1.2%) | 21.55 (-3.1%) | 20.89 (-6.1%) | China Taiping is the only genuinely cheap name in this book - the intrinsic work puts it deeply undervalued, a rare reading in a sleeve where most rows cannot be valued at all. We are not buying it. It fails the quality screen outright, which blocks any new position, and the model's own twelve-week projection of 20.89 sits 4.9% BELOW the 21.96 close - the engine is forecasting a lower price while labelling it a buy. Blocked on two independent counts. OPERATIONAL: Quality_Pass=False hard-blocks new entries (never ADDs, but this is unheld). val_adj +5 = half of the +10 the MoS_FV +0.786 would earn, halved because the quality screen failed. Entry is a buy-stop 0.35 ATR above the close. Ledgered optional/direction:none so the grader prices the gate block. |
| 2888 | STANCHART | 228.60 | HOLD | — | 228.47 | 227.65 | 228.97 | 222.68 | Standard Chartered is flat on cost and sitting just above its risk line, with a weak signal but no exit trigger. The twelve-week view is mildly negative but the price has not broken the stop, so the two-part exit test fails on its first leg and we hold. OPERATIONAL: close 228.60 vs stop 228.47 - above, holds; L2 requires stop-break AND negative 12wk and only the second leg is satisfied. Stop HELD at 228.47, never lowered. The 07-05 overlay override 'hold 0005/2888 unless below the recorded stop' stands and points the same way (overlay is 58 days stale and heavily down-weighted). |
Outcome & track record
Accuracy at the time of this record.
20 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 →
3.000 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.