Recommendations
0 to acquire · 0 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 |
|---|---|---|---|---|---|---|---|---|---|
| 0992 | Lenovo Group Limited | 30.12 | HOLD | — | 29.40 | n/a | n/a | 42.92 | Hold Lenovo in full and raise the trailing stop to HK. The four-session pullback since the record Q1 print is profit-taking into a stretched multiple, not a thesis break — infrastructure revenue nearly doubled to US5bn at a 9.1% operating margin, the AI-server pipeline stands at US$54bn (+157% QoQ), and a Saudi AI infrastructure deal has since landed; with the shares now 34% below their refreshed DCF anchor of 45.65 and the 12-week view at 42.92 (+42.5%), there is no case for selling into weakness. The stop does the risk work: at 29.40 it sits 6.9% above the 27.498 cost, so the downside from here is a locked-in gain. OPERATIONAL: third consecutive NO TRADE cv0 (NoTrade_Reason verbatim 'Consensus conflict + low conviction'); row publishes no Entry/Stop/Target. ADR-0012 fire tick (d) COMPLETED for a 2nd session but DISARMED — Gain_ATR 1.16 < 3.0 and T12_Progress_Pct 70.18 < 90 FROZEN for a 7th snapshot (hand-computed 17.0%; Systemic #1). Trail = chandelier 3.0x ATR 2.2536 off the unchanged 52W high 36.16 = 29.40, up from 29.12; peak_stop re-bases to 29.40, protecting HKHeadroom only 2.39% (close 30.12) after the session low printed 29.60 — one more down session and the position exits on the stop, which is the intended design. L2 does NOT apply (close above stop, 12wk emphatically positive) — a break tomorrow is a trail exit, not an L2 exit. ADR-0020 not triggered (0% erosion, watermark re-based up a 4th session). G1/G3 would block any valuation/divergence trim anyway (Trend 77). Valuation refreshed today: Quality_Pass TRUE, F-Score 7, DCF anchor rose 43.49 -> 45.65 so the standing 'exceeds DCF anchor' annotation on the 12wk target is RETIRED. No Target_Price for a 4th session so 1wk/4wk are n/a and the stale 34.75/38.77 pair stays dropped. HT-6 override condition absent (no placement/dilution headline). RE-ARM to ADD only on a base-build in 29.40-31.50 with RSI < 65 (today 59.14, clear) AND a bullish actionable label returning; no limit placed today per the provenance rule. · news |
| 1024 | KUAISHOU-W | 37.80 | WATCH | 37.80 | — | n/a | n/a | 25.41 (-32.8%) | Not bought despite being the cheapest quality name in the book. Kuaishou screens as genuinely undervalued — margin of safety +0.34, the strongest accounting-quality score in the universe (F-Score 9), and a DCF anchor of 54.88 against a 37.80 close — but the tape is emphatic in the other direction: it printed the highest conviction in the book today on a BREAKDOWN signal, the model's own 12-week view is 25.41 (-32.8%), and the shares sit 28% below their 200-day trend with RSI at 30.5. Cheap plus falling is a value trap, not an entry. OPERATIONAL: AVOID cv86 on BreakdownDown_20D — bearish family, long-only book, not buyable in any reading. HT-4 kill on the forward view; note-B falling-knife pattern. Cheapness cannot originate a buy (CLAUDE.md 9 — valuation amplifies candidates in the technical pool, never originates). Q2 + interim results after today's close, reaction 2026-08-20. Logged so the counterfactual grader prices the refusal and so a naive conviction sort never surfaces it as a long. |
| 2382 | Sunny Optical Technology | 58.40 | WATCH | 58.40 | 57.69 | n/a | 63.13 (+8.1%) | 47.06 (-19.4%) | Not bought. Sunny Optical is the second of three names the refreshed valuation flags as undervalued with a passing quality screen (margin of safety +0.17, F-Score 7, DCF 83.12 vs a 58.40 close), but the technical read gives no entry: the row is a WATCH, not a buy label, conviction sits 10 points under the floor, and the model's 12-week view of 47.06 is 19.4% BELOW today's price with the shares 10% under their 200-day trend. A cheap name the model expects to fall is not an acquisition. OPERATIONAL: WATCH cv55, adj cv 55 (val_adj 0 — a positive MoS cannot lift a non-actionable row), 10 pts sub-floor. HT-4 kill on the forward view. FailedBreakdown_Up_20D is this book's discounted-chase family (L1/HT-2, scorecard n=2 hit_1w 0.50 with 50% stop-first, bias 0). Re-assess only if a bullish actionable label returns with T12 above the close. |
| 9618 | JD-SW | 112.60 | WATCH | 112.60 | — | n/a | n/a | 105.93 (-5.9%) | Not bought. JD is the third undervalued, quality-passing name surfaced by today's valuation refresh (margin of safety +0.12, F-Score 6, DCF 158.90 against a 112.60 close) and it was one of only four names to close green in a broad HK-tech washout — but it prints a bearish failed-breakout signal, sits 4.7% below its 200-day trend, and the model's 12-week view of 105.93 is still 5.9% below today's price. One green session inside a downtrend is not a base. OPERATIONAL: AVOID cv63 on FailedBreakout_Down_20D — bearish family, long-only book. HT-4 kill on the forward view. Valuation cannot originate a buy (CLAUDE.md 9). Third of the three UNDERVALUED + Quality_Pass names all sitting in bearish families with negative 12wk views — the value-trap alignment note-B and HT-4 were written for. Re-assess on a bullish actionable label with a positive T12. |
| 9988 | BABA-SW | 124.20 | WATCH | 124.20 | 122.87 | n/a | 133.09 (+7.2%) | 137.09 (+10.4%) | Still not bought, and yesterday's refusal is already paying. Alibaba's buy signal collapsed within a single session — conviction fell and the shares gave back 2.0% to 124.20 — leaving it 16 points under the entry bar even before the accounting-quality screen it also fails. The forward view remains constructive (4-week 133.09, 12-week 137.09) but June-quarter results land tomorrow, and a binary earnings event is not something a 4.5%-wide stop can underwrite. Waiting for the print costs nothing; being wrong through it costs a full stop. OPERATIONAL: BUY cv69 -> WATCH cv52; val_adj -3 (GROWTH_PRICED_IN, MoS -0.657) -> adj cv 49, 16 pts sub-floor — it now fails on conviction alone before the quality gate (Quality_Pass FALSE, F-Score 3) is needed. The 2026-08-18 withhold at 126.70 is currently +1.97% saved; carried forward, not revised. L6 spirit applies on the way in. L1/HT-2: FailedBreakdown_Up_20D carries no bias here. Results 2026-08-20 pre-US-open ~19:30 HK, HK reaction Fri 08-21 — re-assess after the print, and only on a live Entry_Price (provenance rule), never at the stale 126.70. · news |
| — | AI · HK Tech | — | NO ACTION | — | — | n/a | n/a | n/a | No trades today, on either side. The Hong Kong tech tape had a broad clear-out — names closed lower, several down 8-12% — and for the first time in this window not a single name in the universe generated a buy signal; the three cheapest quality names the refreshed valuation surfaced are all in downtrends the model expects to continue, and cheapness alone is never a reason to buy. On the other side, the one position held (Lenovo) still shows a strongly positive forward view and trades above its stop, so there is nothing to sell either. Cash stays at HK85% of the book, and the only change is a tighter protective stop on Lenovo. OPERATIONAL: ACQUIRE 0 / DISPOSE 0. Buy pool EMPTY — 8 AVOID, 7 WATCH, 3 NO TRADE, zero bullish actionable labels, zero BUY - ANTICIPATE (anticipate slot empty). Top three convictions (1024 cv86, 1347 cv75, 1810 cv72) are all bearish families. 1024/2382/9618 all UNDERVALUED + Quality_Pass and all HT-4-killed on negative 12wk views. Sell pool empty: 0992 NO TRADE, blank Exit_Warning, unarmed profit-take, close above trail, +42.5% 12wk. Only book-keeping change: 0992 trail 29.12 -> 29.40. Fairvalue REGENERATED today (crossed the 7-day bar). Scorecard entries 104 graded 10 calib_1w +0.129, no actionable cells, f=0.30 retained per L5. Overlay 45 days stale and inert. Nothing pending for the 2026-08-20 open. |
Outcome & track record
Accuracy at the time of this record.
10 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 →
0.129 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.