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 |
|---|---|---|---|---|---|---|---|---|---|
| 0992 | Lenovo Group Limited | 30.60 | BUY | 30.60 | 29.56 | 30.73 (+0.4%) | 32.05 (+4.7%) | 33.13 (+8.3%) | Lenovo has printed four consecutive strengthening buy signals (conviction while trading right at the model's entry level with RSI at 59 - a confirmed setup rather than a chase, which is exactly what this book demands from this signal family before acting. Shares screen at fair value with a discounted-cash-flow anchor of HKversus a HKprice and the company passes the quality screen, so the 4-week (+4.7%) and 12-week (+8.3%) model paths sit comfortably below intrinsic value rather than assuming a re-rating. No company news on the day either way. Operational: re-issue of the unexecuted 2026-09-01 card, repriced from 30.58 to the live entry 30.60 (no stale limit carried). Size deliberately held at HKversus a model of HK - HK-1 forbids up-sizing this family and adj 68 is only 3 pts over the floor. Stop 29.56 accepted as printed (sane ATR width, unlike the book's HK-12 rows). L7: family already holds across hktech/hksemicon - execute once at parent level. L8: executes at the 2026-09-03 open; reprice if the open gaps >1.5%. · news |
| 3750 | CATL | 564.50 | SELL 49.87% | 564.50 | 556.49 | 571.21 (+1.2%) | 564.21 (-0.1%) | 565.02 (+0.1%) | CATL has now flagged a breakdown for three straight sessions with the signal strengthening into a 4.65% one-day fall, and the model sees no recovery worth waiting for - the 4-week and 12-week price paths are both flat (-0.05% and +0.1%), so trimming a third of the holding gives up essentially no expected return while removing risk from a position already 25% below cost. Shares screen at fair value, so valuation neither argues for holding on cheapness nor for cutting deeper. No dated company news on the day explains the fall, which itself argues for respecting the tape. Operational: 2nd session of this card (2026-09-01 @592 unexecuted); HK-11 escalation counter. The 09-01 @592 limit is WITHDRAWN, not carried. Provenance: the model Entry_Price 585.32 prints 3.7% ABOVE market (Entry_Distance_ATR 1.05) and is unfillable as a sell limit - action price is the anchor close 564.50, executed market-at-open. Volume: 49.87% x 300 = 149.6 -> nearest board lot (100) =; retains 200. Realised -HKvs cost 751.80. Remaining carry stop 556.49. L8: executes at the 2026-09-03 open. · news |
| 0700 | TENCENT | 438.20 | SELL 85.46% | 438.20 | 430.00 | 431.96 (-1.4%) | 435.88 (-0.5%) | 438.39 (+0.0%) | Tencent is rallying inside a downtrend - RSI 42 with five of the last six daily reads flagging a reduce - and this is the most reliable exit signal this book has, right 63% of the time over 21 graded trades. The model sees the shares essentially flat to lower over the next month (1-week -1.4%, 4-week -0.5%, 12-week +0.04%), so halving the position surrenders no expected return. Shares screen at fair value against a HKcash-flow anchor, so this is a trend call rather than a valuation call, and no company news on the allowed sources changed the picture. Operational: fresh card - the prior 0700 proposal was written off 2026-08-01 among ten stale unacted cards, so the HK-11 counter resets to 1. Volume: 85.46% x 200 =; rounded DOWN to 100 (50%) because rounding to the nearest lot (200) would convert a REDUCE label into a full exit it does not authorise. Avg_Cost recorded as 1.00 (corrupt, prints +43,720%) so realised is NOT computable - backfill the. Remaining hold stop 430.00, which the stale 2026-07-05 overlay independently names as the 0700 exit line. FOLLOW-UP: re-check the H200 chip-access thesis on a datable source - it is a genuine bull case if current. L8: executes at the 2026-09-03 open. · news |
| 0823 | LINK REIT | 38.54 | HOLD | — | 38.70⚠ | 38.49 | 38.57 | 37.96 | Link REIT slipped a fraction below its protective level for a second day, but the shares actually rose on the session and are recovering toward it, and the model sees the price broadly flat over the next month - so this is drift, not a breakdown, and on a high-yielding property trust a slightly lower three-month price path is what the dividend itself does to the share price. Cutting here would bank nothing and give up the income. The protective level is left unchanged rather than lowered. Operational: the 09-01 pre-committed re-arm (second consecutive close below 38.70 with a negative 12wk) fired on COUNT - 38.20 then 38.54, Est_12W 37.96 (-1.5%) - but fails L2's DECISIVE qualifier at a -0.41% breach with price up 0.9% on the day, and the engine prints NO TRADE cv0 which cannot clear the 65 floor. Stop HELD at 38.70, never lowered. Re-arm SHARPENED: a close below 38.20 (prior session's low-water close) with a negative 12wk. |
| 2513 | KNOWLEDGE ATLAS | 1117.00 | HOLD | 1117.00 | 980.00 | 1121.93 (+0.4%) | 1165.38 (+4.3%) | 1266.44 (+13.4%) | Knowledge Atlas fell 5% and is flagging a sell, but the reading carries very low confidence and the same signal family is right only about half the time on a large sample - meanwhile the model still projects the shares meaningfully higher over one and three months (+4.3% and +13.4%). This is the position where holding through the index event on 28 August was rewarded with a 9.6% one-day gain, and the same discipline applies: hold, with a defined level below which the decision is revisited. Operational: STRONG SELL-REDUCE cv37, Sell_Rank #3, stable across all 4 intraday prints, price (-5.26%, -HK). adj = 37 (FailedBreakout_Down_20D|sell bias 0, n=65 hit_1w.516; valuation NA) - 28 pts BELOW the 65 floor, so not a DISPOSE. WATCH line 980.00 stands (1.17x ATR below close). HK-12: the model Stop_Price 1116.99 against an 1117.00 close is a zero-width stop, REJECTED. September HK IPO lock-up expiry window [-8,+4] stance tighten-trail touches this name class - the reason the WATCH line is not widened. |
| 1347 | Hua Hong Semiconductor | 117.50 | HOLD | 117.50 | 105.00 | 124.48 (+5.9%) | 130.93 (+11.4%) | 128.50 (+9.4%) | Hua Hong joins the Hang Seng Index in two days, which mechanically forces index-tracking funds to buy the shares - a well-evidenced upward catalyst worth between 2% and 9%. Selling a depressed holding immediately before a forced-buying event is the wrong side of a known flow, so the position is held through it with a defined level below which that judgement is abandoned. The model also projects the shares higher over one and three months. Operational: AVOID cv53, -26.56% underwater. Catalyst-aware hold binds; HK-11 has already retired this exit to a WATCH line at 105.00 (close 117.50). Exit RE-ARMS at the 2026-09-07 EOD - the post-inclusion unwind session - or earlier on a close below 105.00. HK-12: model Stop_Price 133.04 prints 13.2% ABOVE the close (inverted), rejected as a risk control. |
Outcome & track record
Accuracy at the time of this record.
192 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.