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 | 10.37 | SELL 50% | 10.37 | 10.01 | 10.36 (-0.1%) | 10.38 (+0.1%) | 10.27 (-1.0%) | We are halving a stalled position in an excellent business that has simply become too expensive. ST Engineering's operations are strong - first-half revenue up 11.1% and net profit up 27.1% on a record S7bn order book - but the shares sit roughly 9.6 margin-of-safety units beyond our intrinsic-value anchor, the trend has flattened, and our own twelve-week projection is negative at 10.27 against a 10.37 close. Taking half off the table banks the stall and leaves the rest to run. OPERATIONAL: this is the 09-08 card CARRIED, not a new decision - it was ordered for the 09-09 open and holdings still read 3,800, so under SG-1 it escalates to an UNCONDITIONAL MARKET ORDER at the 09-10 open (limit removed). Repriced 10.34 -> 10.37 per the action-price provenance rule. CLAUDE.md SS9 / L2 basis: leg 2 (Est_12W -0.96%) holds and the 09-07 standing re-arm (settled close <=10.39 + negative Est_12W) is still satisfied at 10.37; leg 1 does not print tonight ONLY because the residual stop was deliberately re-based 10.51 -> 10.01 last night on the explicit condition that we act on the break now - acting is what makes that re-basing legitimate. Engine asks for nothing (NO TRADE cv0, NoTrade_Reason 'Consensus conflict + low conviction') per SG-2. Portion 30% base amplified one tier to 50% under SG-3 on MoS_FV -9.6455; Quality_Pass TRUE so never 100%. Residual stop HELD 10.01 (2.0x ATR chandelier 10.036, EMA200 10.025). Est_12W 10.27 exceeds DCF_per_share 6.831 - annotated, not modified. Scorecard NoTrade|sell -1 (n=12) noted, does not gate a SS9 exit. L8: grade from the 09-10 execution price; flag if the open gaps >~1.5%. · news |
| Z74 | Singtel | 4.56 | WATCH | 4.56 | 4.49 | 4.56 (+0.0%) | 4.59 (+0.7%) | 4.49 (-1.5%) | Singtel is the strongest buy candidate this book has produced in weeks and we are still standing aside, because the model's own twelve-week projection points down. The chart is clean - top-ranked, not extended, entry right at the close - and the quality screen passes, but the projected path is 4.49 in twelve weeks against 4.56 today, and the four-week view reaches barely a fifth of the trade's own 4.71 objective. Paying up for a stock the model expects to be lower two quarters out is not a trade. OPERATIONAL: adj 70 = raw 73 + val_adj -3 (MoS -0.5987 x 5) + bias 0 (FailedBreakdown_Up_20D|buy n=1, report-only - stated, not applied); clears the 65 floor, Quality_Pass TRUE so the hard gate does not bind, Entry_Distance_ATR 0.00 so no-chase passes outright. Refused on Est_12W -1.54% - BN4 was refused last night at a MILDER -0.67%, so buying here would be incoherent. Second independent refusal: L1 discounts FailedBreakdown chases and this book has acted on that family three times with zero graded outcomes and one same-session signal death (V03). Label is one print old (WATCH53 -> NT -> NT -> WATCH56 -> BUY73). Would-be size: 69.18% x 200k x 20% = Smod 1.0x, capped by declared to @ 4.56 = S - essentially the entire CASH row, which must not be committed while holdings are unreconciled. Est_12W 4.49 exceeds DCF_per_share 1.9528 - annotated, not modified. RE-ARM (numeric, fixed): a settled-close BUY reprint with Est_12W >= +1.0% vs the entry arms the acquire at the next open, at the then-live Entry_Price - no carried limit. Counterfactual logged and overruled: no_chase_queue n=43 median +4.43%/1wk 'cost' and SG-6 both argue this book under-adds, but they reach deep-pullback QUEUES, not an at-market entry with a negative forward view. · news |
| U11 | United Overseas Bank | 41.18 | HOLD | 41.18 | 41.27⚠ | 41.13 (-0.1%) | 40.91 (-0.7%) | 40.45 (-1.8%) | UOB slipped nine cents below our risk line today with a mildly negative twelve-week projection - technically an exit trigger, but nine cents on a forty-one dollar stock is noise, not a break, and this is the largest position in the book. We are holding and setting a hard line instead: if it settles at or below 40.85 with the forward view still negative, we cut. OPERATIONAL: L2 BOTH legs technically met - close 41.18 vs stop 41.27 (-0.22%) and Est_12W 40.45 (-1.77%) - and DELIBERATELY REFUSED as drift at 0.16x ATR (0.5476). Consistency basis: this book called -0.29% and -0.76% 'drift around the stop, not a decisive break' on S63 (09-03/09-04) and refused V03 at 0.28x ATR on 09-08; cutting Son session one at 0.16x ATR would contradict all three. Engine prints NO TRADE cv0 ('Consensus conflict + low conviction'), no Sell_Rank, Portion 0. RE-ARM (numeric, fixed): settled close <= 40.85 (~1.0%, ~0.77x ATR below the stop - same geometry as the S63 10.39 trigger) with Est_12W still negative arms the L2 exit at the next open. Stop HELD 41.27, not lowered. Horizon: 09-11 US CPI bank leg sits in the window. |
| BUOU | FRASER L&C TRUST | 0.91 | HOLD | 0.91 | 0.96⚠ | 0.91 (+0.0%) | 0.92 (+1.1%) | 0.93 (+2.2%) | Frasers Logistics keeps flashing a sell and we keep declining it, for the eleventh session running, because the model still projects the price higher over twelve weeks and the trust's own numbers are improving - revenue and net property income both up in the latest half. Selling an oversold industrial REIT at a 32 RSI into a rate-cutting cycle is the wrong side of this trade. OPERATIONAL: representative card for the whole S-REIT sleeve (also J69U adj 84, HMN adj 84, N2IU adj 67, A17U 58, B61 37). Passes SG-2 leg 1 - stop break -5.21%, Sposition, Sof dollar bleed at Portion 80.07, Sell_Rank #3 - and FAILS leg 2 on Est_12W +2.20%. Four sleeve names now clear the 65 floor (up from one on 09-08) yet not one clears leg 2. Signal families carry no usable evidence: FailedBreakout_Down_20D|sell n=73 at hit_1w 0.507 (coin flip, bias 0), BreakdownDown_20D|sell n=2 report-only per SG-5. MoS_FV -5.5593 would amplify a trim one tier IF a sell were ever originated - valuation never originates one (CLAUDE.md SS9). Horizon corroborates trail-not-exit: 09-11 CPI, 09-16 FOMC, 09-18 BoJ all prescribe TIGHTEN-TRAIL on this sleeve. Stop HELD 0.96, not lowered. CAVEAT ESCALATED to systemic finding 1: Est_12W has projected +1.9% to +3.3% on this sleeve for 11 cycles while it actually drifted 1-3% LOWER (BUOU 0.94 -> 0.91 since 19-Aug) - leg 2 may be blocked by an artefact, flagged for the next learnings review. · news |
| O39 | OCBC Bank | 31.35 | HOLD | 31.39 | 31.30 | 31.24 (-0.5%) | 31.37 (-0.1%) | 32.45 (+3.4%) | OCBC is our best holding, up 21% on cost, and it has hit the profit-taking threshold - but nothing about it is deteriorating, so we are letting it run rather than clipping it. The twelve-week view is still higher at 32.45 and conviction is steady. OPERATIONAL: ADR-0012 re-check - ARMED (Gain_ATR 10.24, T12_Progress 96.61, +20.89%, the ONLY holding with a real Avg_Cost 25.93 per SG-7) with ZERO OF FOUR fire ticks: (a) PROFIT_ARMED prints alone, no second Exit_Warning code; (b) conviction flat (dipped 53 intraday, recovered), no >=20pt collapse; (c) Est_12W 32.45 > 1.02 x 31.35 = 31.98, forward view ALIVE; (d) HOLD, not NO TRADE. Armed-without-a-tick means trail-raise, not trim - and the raise is DECLINED because both candidate stops sit below the carried line: model stop 30.33 and 2.0x ATR chandelier 30.29 vs carried 31.30. Stop HELD 31.30, not lowered - note it is only 0.16% under the close, the tightest line on the book, so a routine down-session will trip it. Scorecard PullbackBuy_VWAP|buy -3 (n=6, hit_1w 0.333, recovered from -10) noted; sell-side bias not applicable. |
| V03 | Venture Corporation | 16.74 | HOLD | 16.72 | 16.74 | 16.76 (+0.2%) | 16.75 (+0.2%) | 16.53 (-1.1%) | Venture closed exactly on our stop - not through it - so there is nothing to act on. The business remains sound, with first-half profit up 5.6% and the dividend raised, and we keep the same hard line we set last night. OPERATIONAL: close 16.74 vs stop 16.74 = 0.00x ATR, no break exists; trail NT -> WATCH53 -> BUY60 -> BUY60 -> WATCH53. The 09-08 re-arm is UNCHANGED AND UNFIRED: a settled close at or below 16.60 with Est_12W still negative arms the L2 exit at the next open. Est_12W 16.53 = -1.25%. Stop HELD 16.74. UNRECONCILED (ADR-0019): sg_state.json carries from the 09-08 open fill @16.90 while input/sgportfolio.json still shows - one sync flag per SG-8, not a re-published card. |
Outcome & track record
Accuracy at the time of this record.
129 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 →
0.162 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.