Recommendations
0 to acquire · 1 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 |
|---|---|---|---|---|---|---|---|---|---|
| S63 | ST Engineering | 10.34 | SELL 50% | 10.34 | 10.01 | 10.33 (-0.1%) | 10.34 (+0.0%) | 10.26 (-0.8%) | We are selling half our ST Engineering holding. This is the one position on the book where the evidence has stopped being ambiguous: the shares have now closed below our protective line for a fifth straight session, the break has deepened every step of the way, and all three of the model's forward views have turned flat-to-negative. We set an explicit, public trigger four sessions ago -- a settled close at 10.39 or below -- and today it closed at 10.34. On our own rules that must be acted on now rather than deferred a sixth time. The business is sound and we are keeping half, but the shares are also priced far above what the cash flows support, so we take the risk off in the half we sell. OPERATIONAL: L2 BOTH LEGS MET AND THE STANDING RE-ARM HAS FIRED. Leg 1: settled close 10.34 vs the carried 10.51 stop = -1.62%, 1.02x ATR -- the DEEPEST of the sequence (09-02 -0.19%, 09-03 -0.76%, 09-04 -0.29%, 09-07 -1.05%, tonight -1.62%) and a 5th consecutive session below stop. Leg 2: Est_1W 10.33 (-0.10%), Est_4W 10.34 (0.00%), Est_12W 10.26 (-0.77%) -- negative. The 09-07 card's re-arm was written UNAMBIGUOUS AND FINAL: 'a settled close at 10.39 or below with Est_12W still negative arms the L2 exit -- and per L3/SG-1 it must then be ACTED ON at the next open, not deferred a fifth time.' 10.34 clears it by 5 cents; the session HIGH 10.38 never touched the trigger level, so there is no goalpost ambiguity. Engine still asks for nothing (NO TRADE cv0, Portion 0, no Sell_Rank, NoTrade_Reason verbatim 'Consensus conflict + low conviction') -- this is a CLAUDE.md SS9 / L2 signal-based exit, not a label-driven one, exactly as SG-2 requires (stop-break + negative-12wk, never the label alone). PORTION: base 30% per this book's L2 de-risk convention (U11 08-18/08-26/08-28, B61 08-21), AMPLIFIED ONE TIER to 50% under SG-3 -- MoS_FV -9.6455 is far past the -3 sell-amplify trigger and the 09-04 and 09-07 cards both pre-committed to exactly this amplification 'the moment a qualifying sell exists'. Quality_Pass TRUE, so this is a de-risk of an expensive good business, NOT a quality exit -- hence 50% and never 100%. Est_12W 10.26 exceeds DCF_per_share 6.83 -- annotated, not modified. Scorecard NoTrade|sell -1 (n=11, hit_1w 0.455, ACTIONABLE) noted as a haircut; it does not gate a SS9 exit. Close sits below EMA20/EMA50 and is now converging on EMA200 10.02. Proceeds ~Son a Sposition. RESIDUAL : stop RE-BASED to 10.01 -- the 2.0x ATR chandelier (10.34 - 0.333 = 10.007) sitting in confluence with EMA200 10.02. This is a deliberate LOWERING of the 10.51 line and it is only legitimate because we are ACTING on the break in the same breath; the old line is unworkable for a residual once price is 1.6% through it. News: none found on Yahoo/SeekingAlpha for today or the prior session -- no catalyst override; the August dividend was paid 04-Sep and went ex weeks earlier, so this is clean price decay, not a payout artefact. Horizon: 09-11 US CPI and 09-16 FOMC land inside the window but neither names S63. L8: executes at the 09-09 open; reprice if it gaps >~1.5%; horizons and the scorecard grade from the execution date/price. · news |
| BN4 | Keppel | 11.88 | WATCH | 11.80 | 11.13 | 11.91 (+0.9%) | 11.91 (+0.9%) | 11.80 (+0.0%) | Keppel was the only genuine buy candidate on the book tonight and we are not taking it. The chart is the best on the book -- a clean 20-day breakout on heavy volume from the most reliable buy pattern we have -- but the company fails our financial quality screen, and that screen is a hard block on any brand-new position. We would rather miss a good chart than start a position in a business that cannot clear our own bar. OPERATIONAL: THE FLOOR IS CLEARED AND THE GATE STILL BLOCKS -- adj 73 = raw 73 + val_adj -10 (MoS_FV -3.0931 x 5 = -15.5, clamped to -10) + scorecard bias +10 (BreakoutUp_20D|buy, n=6, hit_1w 1.000, hit_4w 1.000, ACTIONABLE -- the book's single best cell). 73 >= the 65 floor. BLOCKED SOLELY on the Step 6 hard quality gate: Quality_Pass FALSE (F_Score 6, Composite_Rank 13.0) on a NEW ENTRY (, unheld) -- the gate blocks new entries only, never adds. This is now the SECOND consecutive cycle BN4 is refused, but the reason has NARROWED: on 09-07 it failed BOTH the gate AND the floor (adj 52); tonight it fails ONLY the gate. Verified against the FRESHLY REGENERATED fairvalue (FetchedAt 2026-09-08 18:03) -- not carried from the stale file, because the 09-04 entry's Z74 read error is exactly the mistake this re-check exists to prevent. Quality_Pass is False in both the old and new file; the gate is real. TWO INDEPENDENT SECONDARY REFUSALS, either of which would have demoted this to a queue anyway: (1) NO-CHASE -- Entry_Distance_ATR -0.38 is past the -0.25 bar, close 11.88 sat at the session high 11.91, so at best this was a QUEUED GTC at the 11.80 model entry, never a market buy; (2) the 12-week view is NEGATIVE -- Est_12W 11.80 = -0.67% vs the 11.88 close, with Est_1W and Est_4W both only +0.25%, so the model pays us nothing beyond a week. Est_12W 11.80 also exceeds DCF_per_share 1.797 by ~6.6x -- annotated, not modified; the DCF anchor on BN4 is not credible and is NOT used to amplify anything. Signal trail intraday was WATCH cv35 -> WATCH cv53 -> WATCH cv53 -> BUY cv73, i.e. the BUY label is ONE print old and appeared only at the settled close -- L4 says the settled close is authoritative, but a one-print label is not the multi-snapshot stability this book requires. News: none found for today or the prior session. RE-ARM: BN4 becomes actionable only if Quality_Pass flips TRUE on a future fairvalue refresh; until then no BN4 chart, however good, opens a position. Note 09-21 FTSE ST index review touches BN4 on weight only. · news |
| V03 | Venture Corporation | 16.66 | WATCH | — | 16.74 | 16.67 | 16.70 | 16.51 | Our Venture order filled at the open and the shares closed lower on the day, so the position is about 1.4% underwater on its first session and the buy signal that justified it has already gone. We are not selling on day one -- the shares are barely through our risk line and the business itself is doing well, with first-half profit up 5.6% and the interim dividend raised 20%. But we have set a firm level: another leg down and we cut it. OPERATIONAL: QUEUE FILLED, THEN THE SIGNAL DIED SAME-SESSION. The 09-07 gtc_limit @16.91 filled at the 09-08 OPEN of 16.90 (open = high = 16.90, session low 16.51), ~= SClosed 16.66 = -1.42% vs the fill. Action trail 15:00 BUY cv60 -> 16:00 BUY cv60 -> 16:30 NO TRADE cv0 -> 18:00 NO TRADE cv0, NoTrade_Reason verbatim 'Consensus conflict + low conviction' -- the label collapsed intraday on the very session it filled. ALL THREE 09-07 CAVEATS NOW VINDICATED ON THE TAPE: the horizons were flat-to-negative and the price went the flat-to-negative way; the identical FailedBreakdown_Up_20D setup at 16.88 on 08-31 that 'went nowhere in five sessions' has now gone nowhere in six and given back 1.4%; and the family still has ZERO graded samples (FailedBreakdown_Up_20D|buy absent from the scorecard for a THIRD consecutive acted card). L1's discount on FailedBreakdown chases was right and the card overrode it on Buy_Rank -- log that. L2 IS TECHNICALLY MET AND IS DELIBERATELY REFUSED: leg 1 close 16.66 vs the 16.74 stop = -0.48%, but that is only 0.28x ATR (0.284) -- this book has twice called breaches of -0.29% and -0.76% 'drift around the stop, not a decisive break' (S63 09-03, 09-04) and consistency demands the same reading here; leg 2 Est_12W 16.51 = -0.90%, negative. Refusing also on the ground that cutting a position on its FIRST session, 0.28 ATR through a stop that was itself only set last night, is churn, not risk control. RE-ARM, NUMERIC AND FIXED: a settled close at or below 16.60 (~1% and ~0.5x ATR below the 16.74 stop) with Est_12W still negative arms the L2 exit, to be acted on at the next open. Stop HELD 16.74, not lowered. News CONFIRMS the hold: 1HFY26 net profit +5.6% to S, revenue +7.4% to S35bn, interim dividend raised to 30 cents from 25, S11bn cash and no borrowings, ~5% trailing yield; ex-date 31-Aug is past so no ex-div artefact in today's fall. Valuation GROWTH_PRICED_IN, MoS -0.3687, Quality_Pass TRUE. UNRECONCILED in input/sgportfolio.json (V03 still). · news |
| BUOU | FRASER L&C TRUST | 0.90 | HOLD | 0.91 | 0.96⚠ | 0.90 (-1.1%) | 0.91 (+0.0%) | 0.92 (+1.1%) | Frasers Logistics is the biggest position on the book and the model's top-ranked sell, and for the tenth session running we are not selling it. The reason has not changed: the shares are below our risk line, but the model's own 12-week view still points HIGHER, not lower, and this book's rule is that a Singapore sell label alone is not enough -- it fades more than half the time. Selling S of a trust the model thinks recovers, on a label we have measured as a coin-flip, is not risk control. OPERATIONAL: SG-2 LEG 1 PASSES, LEG 2 FAILS, for a 10th consecutive cycle. Leg 1: close 0.900 vs the 0.96 stop = -6.25%, deep and unambiguous, RSI 24.2 (deeply oversold), close below EMA200 0.9409. Leg 2 FAILS: Est_12W 0.92 = +2.2% POSITIVE vs the close. Sell_Rank #1, SELL - REDUCE cv74, Portion 49.19% -- would be / ~S of dollar bleed if taken. Scorecard: BreakdownDown_20D|sell is n=2 (hit 1.0) -- REPORT-ONLY, below the n>=5 floor, so bias 0 stated and NOT adjusted; a 2-sample perfect record is exactly the small-sample artefact SG-5 was written to reject. Valuation FAIR_VALUE, MoS_FV -5.5593 -- past the -3 SG-3 amplify trigger, so this WOULD amplify one tier the moment leg 2 flips, but valuation never ORIGINATES a sell (CLAUDE.md SS9). THE SLEEVE-WIDE FACT, NOW IN ITS 5TH CONSECUTIVE CYCLE: 7 of tonight's 8 stop breaches carry POSITIVE Est_12W -- BUOU +2.2%, A17U +2.95%, J69U +2.35%, N2IU +2.44%, HMN +2.35%, B61 +2.04%, 9CI +0.77% -- and S63 (-0.77%) is the ONLY negative one, and it is the ONLY one being sold. That single fact is the entire load-bearing reason the REIT sleeve is held, and it remains the highest-value open question on this book. Horizon independently corroborates: 09-11 US CPI, 09-16 FOMC and 09-18 BoJ all prescribe TIGHTEN-TRAIL on the S-REIT sleeve, not exit. News: none found. Stop HELD 0.96, not lowered. · news |
Outcome & track record
Accuracy at the time of this record.
127 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 →
0.229 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.