Recommendations
1 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 | 28.10 | ADD — QUEUED @27.06 | 27.06 | 21.52 | 30.39 (+12.3%) | 38.15 (+41.0%) | 40.04 (+48.0%) | We are holding all 4,000 Lenovo shares and adding to them on a pullback - the one holding in this portfolio that is working. The shares jumped 8.3% today to a fresh 52-week high on nearly double the usual volume, and the model's rating recovered to a buy on a breakout to new highs - the strongest trend reading in the book. Behind the price move is a real business: AI-related revenue grew 84% over the past year and now makes up roughly 38% of the group, the datacentre division set a record quarter, and management points to a pipeline of around billion. On our own valuation work Lenovo is the second-highest-quality company of the twenty-one we follow, and the twelve-week projection of HKstill sits below our discounted-cash-flow anchor of HK - so the price is full but not detached. We are NOT buying at today's close of HK. That would be chasing a stock that has already run past the level our model wants to pay, at the top of its 52-week range. Instead we have placed a standing order to buy 2,000 more shares only if the price eases back to HK, about 3.7% below today's close, and that order lapses on 12 August - the day before Lenovo reports quarterly results - so we are never buying into an earnings surprise. If it fills, we own at an average of HKand still hold over 91% of the portfolio in cash; if it does not, we keep the cash and lose nothing. | Ops: QUEUED gtc_limit, entry band 26.61-27.51 (Entry_Price +/- 0.25*ATR 1.8007). ORDER-TYPE ASSUMPTION - row prints Order_Type 'Buy Stop @ HK' but 27.06 is 3.70% BELOW the 28.10 close; a buy-stop below market fills at market and defeats the no-chase rule, so it is queued as a pullback LIMIT (engine defect, Systemic #1). No-chase basis: Entry_Distance_ATR -0.58, RSI 67.2, 52W-high close 28.10 (High_52W 28.14) on 118.0m sh vs 63.0m. Sizing: Portion 58.23% x 4000 =, size_mod 1.0x (MoS -0.39 inside -1 tier), rounded DOWN to the 2,000-share HKEX board lot = HK; post-fill @ blended 26.10 = 8.4% of book. Risk on the add (27.06-21.52) x 2000 = HK = 0.55% of book; full 6,000 position 1.37%. R:R 2.00 (risk 5.54 / reward 11.09) - HT-3 does NOT apply, this is chandelier_trail not the tuned pullback exits. CAVEAT (Systemic #3): target_4w 38.15 sits 17.1% ABOVE Est_High 32.5813 - L1's documented BreakoutUp_52W overshoot; on the honest 32.58 the R:R falls to ~1.00. target_1w is f=0.30 interpolation (L5 mandatory: calib_1w -0.45, 9 graded buys < 10) and inherits that overshoot - least trustworthy number on the card. target_12w 40.04 is INSIDE DCF_per_share 43.47, no exceedance annotation. VALUATION CAVEAT: fairvalue prices 0992 at 24.74 vs the 28.10 close (13.6% stale basis); marked to today MoS is ~-0.58 -> val_adj -3 -> adj 66, still above the 65 floor, verdict unchanged. Signal is NOT stable-multi-print: BUY76 -> SBUY79 -> WATCH53 -> BUY-ADD69, a one-session recovery - hence limit + partial size, not market + full size. Learnings: L1 (BreakoutUp_52W is a DISCOUNTED family - discount applied to target credibility and sizing, not as a veto), L5, L6 (queue expires 08-12, before the 08-13 Q1 FY27 results - no gap-chase), L8, HT-2 (bias 0; BreakoutUp_52W has n=0 graded history in this book), HT-4 (forward view + definable stop + R:R screened BEFORE the floor - first hktech candidate to clear all three), HT-6 (re-checked CLEAR - no new placement/registration/secondary; June 2026 convertible dilution remains a flagged known risk). Counterfactual no_chase_queue n=4 median wh_1w +5.07% [COST] is CITED and deliberately NOT applied (Step 4.4); this queue makes it n=5 and actionable at the next scoring run. Re-arm/withdraw: auto-withdraw if conviction < 65 or on 2026-08-12; reprice to the live Entry_Price at the next EOD per the provenance rule - never republish @27.06 as a stale limit. HOLD LEG (existing @ 25.62, +9.68% / +HKMV 112,400 = 5.59% of book) - every sell-origination route checked and CLOSED: (a) no Exit_Warning code, so the CLAUDE.md section 9 warning-elevate path has nothing to elevate; (b) NOT PROFIT_ARMED - Gain_ATR 1.38 (needs 3.0), T12_Progress_Pct 70.18 (needs 90; arming requires a close >= 0.90 x 40.04 = 36.04), so ADR-0012 cannot fire and the conviction-collapse tick is unreachable because arming is its precondition; (c) no stop break - close 28.10 vs stop 21.52, 23.4% headroom, L2 inapplicable; (d) Target_12Week +42.5% vs close, so L2's second leg fails independently; (e) label is BUY - ADD, not SELL - REDUCE; (f) price above EMA5 26.20 / EMA10 25.22 / EMA20 24.31 / EMA50 22.23 / EMA100 19.06 / EMA200 15.74. ADR-0020 trail erosion 0% - Peak Stop ROSE 21.31 -> 21.52, new watermark recorded. Not armed, so no mandatory trail-raise; the model's own chandelier_trail lifted the stop and that engine line (21.52) is carried rather than a hand-tightened one (a raw 3.0xATR chandelier off the close computes 22.70 - deliberately not used). Held-share estimates from the 28.10 close at f=0.30: 1wk 31.12 (+10.7%), 4wk 38.15 (+35.8%), 12wk 40.04 (+42.5%). Exit triggers: decisive close below 21.52; OR any Exit_Warning code; OR a SELL - REDUCE label; OR two consecutive NO TRADE / sub-40 prints WITH Target_12Week rolling under the close (L2). Q1 FY27 results 2026-08-13 (4 sessions, single-sourced, absent from the parent horizon file) - the hold-through-earnings decision belongs to the 08-12 run. · news |
| 9618 | JD-SW | 127.50 | WATCH | 127.50 | 126.20 | n/a | 134.06 (+5.1%) | 154.89 (+21.5%) | JD.com carries the second-best twelve-week outlook in this portfolio - our model projects HKagainst today's HK, about 22% higher - and it is fairly valued rather than expensive. But for the sixth session running the short-term signal will not confirm: today's reading recovered from nothing to a weak watch, which is progress but still well short of a buy rating. We are not paying for a forecast the day-to-day trend refuses to endorse, so JD stays on the watch list with no money committed. | Ops: no actionable label (WATCH cv53, trail NOTRADE0 -> WATCH15 -> NOTRADE0 -> WATCH53); floor is 65 on adj conviction and this is 53 with val_adj 0 and bias 0. Family is FailedBreakdown_Up_20D - L1-discounted here, and the local cell is n=2 hit_1w 0.00 (report-only, HT-2). Constructive change worth recording: cv0 -> cv53 on a rising RSI 65.1, close 127.50 (+0.16%), Trending regime, stop definable at 126.20. Watch, do not queue. Re-arm: a BUY/STRONG BUY label at conviction >= 65 with Target_12Week intact. |
| PORTFOLIO | hktech | — | WATCH | — | — | n/a | n/a | n/a | One decision today: add to Lenovo, but only on a pullback. The shares rose 8.3% to a 52-week high and are now the portfolio's only holding and its only genuine winner, up 9.7% since we bought them two sessions ago. Rather than chase the price, we have left a standing order to buy 2,000 more shares at HK - roughly 3.7% below today's close - which expires on 12 August, the day before Lenovo reports results. Nothing is being sold. Of the other twenty names we follow, eight remain in outright decline with projections 30-43% below current prices, and the rest give no buy signal at all, so the portfolio stays over 91% in cash whether or not the Lenovo order fills. The book is worth HKagainst its HKstarting capital. | Ops: 1 ACQUIRE (queued, 0992 ADD @ 27.06 gtc_limit, expires 08-12) / 0 DISPOSE / 1 HOLD / ANTICIPATE slot empty (zero BUY - ANTICIPATE prints across the last three snapshots). Book value HK = 112,400 (5.59%) + cash 1,897,520 (94.41%); +0.50% vs the 2,000,000 seed. : buys need HKIF filled, sells free HKself-financing from declared (2.85% of it), leaving HKScorecard entries 70, graded 9, calib_1w -0.45 -> f = 0.30 (L5 mandatory); ALL cells below n>=5 so scorecard_bias 0 everywhere (HT-2). Valuation FRESH (FetchedAt 2026-08-03, 4 days) but its 24.74 price basis for 0992 is 13.6% below the 28.10 close - flagged on the card. Overlay 33 days stale and inert. Event Horizon: parent hk fallback, 0992 absent for the FOURTH consecutive session (Systemic #2); 0992 Q1 FY27 results 2026-08-13 drove the queue-expiry design. Systemic findings: #1 Order_Type prints 'Buy Stop' on a below-market entry (would execute as a chase); #2 parent horizon misses hktech names; #3 BreakoutUp_52W 4wk target 17.1% above Est_High, which inflates every R:R computed from it (2.00 -> ~1.00 on the honest figure); #4 7709 prints WATCH with Stop_Price -3.40; #5 2318 identity defect (Ping An, not AAC) UNFIXED from 08-06; #6 no_chase_queue counterfactual reaches n=5 with today's queue - highest-value item for the next learnings review. Step 8a NOT run (batch mode). |
Outcome & track record
Accuracy at the time of this record.
9 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 →
-0.450 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.