Recommendations
0 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 |
|---|---|---|---|---|---|---|---|---|---|
| 2513 | Z.AI (Knowledge Atlas) | 1059.00 | SELL 30% | 1059.00 | 975.00 | 1096.84 (+3.6%) | 1109.79 (+4.8%) | 1244.10 (+17.5%) | Trimming 30% of our largest AI holding after the stock made another lower high and closed below both its 20- and 50-day trend lines, with the engine flagging its strongest sell signal of the day. The position is down 35% from cost and still represents 19% of the book, so cutting a third reduces concentration risk while keeping for the MSCI index inclusion that takes effect on 31 August. Valuation gives no read here (the company has negative earnings power on the model), so this is purely a trend-and-risk decision. OPERATIONAL: engine asked for 85%; deliberately down-sized to the 30% floor tier under the catalyst-aware-hold clause (MSCI ACWI effective 31-Aug, direction up) and HK-9 (hk sell signals are a 1-week instrument). REPRICE of the unexecuted 08-21 card from 1,129.00 per the provenance rule; non-execution of that card cost HKHK-11: this is re-issue #2 of the chain - the next EOD must escalate once then RETIRE to a WATCH line, never a fourth issue. Model Stop_Price 1,058.99 rejected as zero-width (HK-12); residual trails at 975.00. Counter-case acknowledged: strong intraday reversal (low 971 -> close 1,059 near the 1,065 high) and all three Est_* horizons positive - discounted because Est_Model_Ver prints (stale) and calib_1w is -0.261. · news |
| 3750 | CATL | 621.00 | SELL 33.3% | 621.00 | 600.00 | 618.60 (-0.4%) | 618.02 (-0.5%) | 640.22 (+3.1%) | Taking a third off CATL because the stop line we committed to a few sessions ago has been broken: the shares closed at 621 against our 625.50 trail, having slipped below both the 20- and 50-day averages on two straight down days, and the signal engine's confidence in the name collapsed from 63 to zero during the session. The position is 17% underwater and the model sees the shares going nowhere over the next month. We are keeping two thirds because the longer-term valuation still looks reasonable - the discounted cash-flow anchor sits 16% above today's price - and there is no company news driving the weakness. OPERATIONAL: origination is the pre-committed 08-21 re-arm 'close < 625.50 OR a fresh SELL-REDUCE' (price leg fired) plus CLAUDE.md §9 sustained NoTrade/Conflicted (3 consecutive NT prints into the close, Pillar_Consensus Indeterminate). L2 does NOT fire (Est_12W +3.09%, not negative) - that is why this is the minimum tier and not the engine's historic 67%. Breach is only -0.21xATR, stated as marginal. PROFIT_ARMED on this row is SPURIOUS (Gain_ATR -6.19) - ADR-0012 is not the basis. Row prints no Entry_Price/Stop_Price, so the settled close is the reference. Board-lot caveat: 30% = but no HK lot table exists, so (33.3%) is taken as safe under either a 10- or 100-share lot. Residual 200 trails at 600.00 (-0.99xATR, below today's 618 low). · news |
| 0981 | SMIC | 67.85 | HOLD | 67.85 | 69.63⚠ | 68.41 (+0.8%) | 70.81 (+4.4%) | 74.18 (+9.3%) | Holding SMIC through Nvidia's results tomorrow despite the shares slipping 2.6% below our stop line and under all three major moving averages. Nvidia reports after the US close on 26 August and SMIC trades directly off that read-through, so selling into the event rather than after it would be trading blind; the model also still sees the shares 9% higher over twelve weeks. But the reprieve is conditional, not open-ended. OPERATIONAL: STOP BROKEN - close 67.85 vs the 69.63 trail (-2.56%, -0.46xATR), below EMA20 69.99 / EMA50 71.34 / EMA200 68.83. L2 does NOT fire (Est_12W +9.34%). Model Stop_Price 75.66 is 11.5% ABOVE the close - inverted, rejected (HK-12). Deferred ONE event only under the catalyst-aware-hold clause; -0.46xATR is not the 'decisive break' that clause requires to pre-trim. BINDING PRE-COMMITMENT: if the 27-Aug settled close is below 69.63, a 30% trim is MANDATORY at the 28-Aug open - no further re-evaluation and no lowering of the line. Stand down only on a settled close at or above 69.63. Trail deliberately NOT re-based. PROFIT_ARMED but Avg_Cost blank -> no-guess blocks ADR-0012. |
| 0700 | Tencent | 442.00 | HOLD | 442.00 | 430.00 | 439.28 (-0.6%) | 432.92 (-2.1%) | 438.75 (-0.7%) | Tencent again screens as one of the weaker names we own - trading well below its 200-day average and below its volume-weighted average price - but we are not selling a 1.6% position into that weakness. The exit was already retired to a hard line at HKand the shares closed at 442, so nothing has triggered. OPERATIONAL: prints the book's #2 sell (STRONG SELL - REDUCE cv70, Sell_Rank 2, 85.46, adj 71) and is deliberately NOT acted on - HK-11 binds (exit retired 08-17 to the 430.00 WATCH line, also the overlay's standing override; never re-issue). Close 442.00 is 2.8% above the line - the closest since retirement. Model Stop_Price 418.35 is BELOW 430.00 -> do not lower. Avg_Cost = 1.00 is a DATA ERROR (Unreal_PnL_Pct +44,100%), so PROFIT_ARMED is meaningless on this row. |
| 7709 | CSOP SK Hynix 2x (XL2CSOPHYNIX) | 37.24 | HOLD | 37.24 | 28.00 | 39.08 (+4.9%) | 37.34 (+0.3%) | 39.13 (+5.1%) | Holding the leveraged SK Hynix tracker for an eighth session. It shows a sell label but at very low conviction, and our standing rule for these 2x products is that only a high-conviction, repeated breakdown justifies exiting - a past exit on a similar product missed a 5.8% bounce. The line that would force action is HKand the fund closed at 37.24. OPERATIONAL: re-arm is 'decisive close < 28.00 OR a breakdown-family signal' - close is 33% above the line and the family is failed_breakout, not breakdown. HK-11 binds. Leveraged-ETF rule independently blocks a full exit (requires conv >= 70 + 3 prints + price < stop; cv39). HK-12 IMPROVEMENT: Stop_Price now prints 26.04, BELOW the close, correcting the 51.89 inverted print of 08-21. Trail held 28.00. Position -67.49% vs Avg_Cost 114.56. |
| 0005 | HSBC Holdings | 163.20 | HOLD | 163.20 | 159.50 | 163.30 (+0.1%) | 162.93 (-0.2%) | 162.66 (-0.3%) | HSBC remains the anchor of the book at 36% of assets and strengthened again through the session, holding 18% above its 200-day average and within touching distance of its 52-week high. The model sees the shares flat from here rather than higher, so we are protecting the gain with the stop rather than adding. OPERATIONAL: conviction climbed -> -> across the six snapshots. PROFIT_ARMED (T12Prog 100.33) with tick (c) firing (Est_12W 162.66 < 1.02 x 163.20 = 166.46), but Avg_Cost is BLANK -> ADR-0012 no-guess rule forbids the profit-take; resolves to a trail action. Model Stop_Price 158.07 is BELOW the standing 159.50 WATCH line -> do not lower. Trail held 159.50. Rate-event exposure is the largest in the book (horizon anchors every macro band to 0005: Jackson Hole +/-2.27xATR, PCE +/-1.94xATR). |
| 0939 | CCB - H Shares | 9.18 | HOLD | 9.18 | 8.88 | 9.16 (-0.2%) | 9.28 (+1.1%) | 9.27 (+1.0%) | Holding CCB into its interim results and dividend on 28 August - the only company in the book with a scheduled earnings date in the next fortnight. The shares faded to a no-trade reading at the close as the model's indicators disagreed, but they still sit 10% above the 200-day average and the dividend is worth waiting for. OPERATIONAL: trail HOLD 64/60/58/61 -> NT0 x2 into the close, Pillar_Consensus Conflicted. PROFIT_ARMED (T12Prog 99.03) with tick (c) firing (Est_12W 9.27 < 1.02 x 9.18 = 9.36) but Avg_Cost is BLANK -> ADR-0012 no-guess rule blocks it. Catalyst-aware hold through the 28-Aug print. Trail held 8.88. Position = HK8.5% of book. |
| 2888 | Standard Chartered | 236.40 | HOLD | 236.40 | 229.36 | 236.24 (-0.1%) | 237.43 (+0.4%) | 226.48 (-4.2%) | Standard Chartered is the strongest holding in the book on the model's reading, strengthening through the whole session and sitting 21% above its 200-day average near the top of its yearly range. We are holding and letting the stop do the work. OPERATIONAL: conviction -> -> on FailedBreakdown_Up_20D, Consolidation_Bullish. PROFIT_ARMED with tick (c) firing (Est_12W 226.48 < 1.02 x 236.40 = 241.13) but DOUBLY blocked - the exit thread was retired 08-13 and Avg_Cost is BLANK (ADR-0012 no-guess). Model Stop_Price 226.48 is below the standing 229.36 -> do not lower. Trail held 229.36. |
| 1347 | Hua Hong Semiconductor | 110.40 | HOLD | 110.40 | 105.00 | 116.31 (+5.4%) | 125.19 (+13.4%) | 129.09 (+16.9%) | Hua Hong has no sell signal today and the model projects the shares 17% higher over twelve weeks, so the retirement of its exit thread continues to look right. The shares trade 15% below their volume-weighted average price - the widest such gap in the book - which usually marks a washed-out rather than a deteriorating position. The line that would force a sell is HKand the close was 110.40. OPERATIONAL: AVOID cv53, Sell_Rank blank, 0.0 - the 08-19 HK-11 retirement holds. Close is below EMA200 115.74 but 5.1% above the WATCH line. All three horizons positive (+5.4% / +13.4% / +16.9%). Model Stop_Price 131.31 prints 19% ABOVE the close - inverted, rejected (HK-12). Trail held 105.00. FLAG FOR NEXT CYCLE: HSI inclusion effective Mon 7-Sep with the characteristic post-implementation give-back carried at direction down, band -6.0..+2.0. |
| 0522 | ASMPT | 162.50 | HOLD | 162.50 | 155.02 | 165.17 (+1.6%) | 169.37 (+4.2%) | 164.69 (+1.3%) | ASMPT carries the best four-week outlook of anything we own (+4.2%) and trades 4% above its volume-weighted average price, the healthiest short-term posture in the book. The signal engine can't make up its mind, so there is no trade, but the earlier stop breach has fully healed and we hold. OPERATIONAL: NO TRADE cv0 on all four of today's prints. NoTrade_Reason verbatim: 'Consensus conflict + low conviction'. Close 162.50 is 4.8% above the 155.02 trail. PROFIT_ARMED is SPURIOUS (-17.24%, Gain_ATR -3.60). Carries the widest single-name horizon exposure in the book (NVDA 26-Aug ecosystem, +/-12.0%). Trail held 155.02. |
Outcome & track record
Accuracy at the time of this record.
179 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 →
-0.261 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.