Singapore — End-of-Day Verdict
31st cycle: the acquire side reopens after two empty cycles as S68 SGX prints the board's only bullish actionable row — STRONG BUY - ADD cv99 at BuyRank #1 on a monotonic → conviction ramp — but it is 2 ATR extended, so SG-6's cv≥95 immediate-tranche carve-out is rejected on three grounds and the ADD is queued at 24.27; and O39 becomes the THIRD live decision blocked on the same unreconciled holdings file, leaving the dispose side structurally empty
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 |
|---|---|---|---|---|---|---|---|---|---|
| U11 | United Overseas Bank | 41.13 | SELL 30% | 41.13 | 41.27⚠ | 41.14 (+0.0%) | 41.03 (-0.2%) | 41.06 (-0.2%) | UOB is the book's second-largest position at 26.8% of net worth and it has now closed below its risk line for a fourth straight session while the signal engine has gone completely quiet - the twelve-week view, which had held above the share price through three earlier breaks, finally rolled under it tonight. The fundamentals corroborate the chart rather than contradict it: net interest margin has slid 1.82% to 1.74% to 1.71% exiting July, below management's own 1.75-1.80% guidance for the year, as loans reprice into a lower-rate environment, and first-half profit grew only 3% year-on-year. Taking roughly a third off the position banks the 88c dividend that went ex today, cuts single-name concentration ahead of a dense two-week macro calendar, and leaves the majority of the holding intact for the long-term uptrend that is still in force above the 50- and 200-day averages. OPERATIONAL: L2/CLAUDE.md-9 origination, NOT a signal label - the engine prints NO TRADE cv0 ('Consensus conflict + low conviction') and the stop-break + negative-12wk combination is itself the signal-based sell reason. The 08-14 card pre-registered this exact trigger ('a close below 41.27 after the rebase on a NEGATIVE Est_12W makes U11 the immediate lead dispose') and BOTH limbs fired tonight: close 41.13 vs the 88c-rebased stop 41.27 (-0.34%, 0.21 ATR) and Est_12W 41.06 vs close 41.13 (-0.17%). Persistence substitutes for depth - 4th consecutive break, conviction collapse ->, Est_12W decaying monotonically 42.29 -> 42.09 -> 42.00 -> 41.84 -> 41.06. MINIMUM TIER (30%, not 50%) because both limbs are hairline and NoTrade|sell is the book's worst actionable cell at -5 (n=7, hit_1w 0.25, hit_4w 0.00) - the haircut sizes the trade down, it does not veto an L2 origination. Valuation is amplify-only and MoS -0.55 is not <= -3, so no tier bump. Action price is the settled close: the NO TRADE row prints no Entry_Price, so this is a market order referenced to the anchor close, not a level carried forward. Ex-div honesty: U11's -1.60% today is the 88c going ex; adjusted for it the stock ROSE ~0.5%, which is why the break is shallow - but the stop was rebased by the same quantum so the comparison is apples-to-apples. Residual (S18.7% weight) keeps the 41.27 stop. SG-8 does not apply - this is publication #1 for U11, not a re-publication. · news |
| G07 | Great Eastern | 22.74 | WATCH | 22.96 | 21.35 | 22.69 (-1.2%) | 22.28 (-3.0%) | 22.34 (-2.7%) | Great Eastern was the only buy signal on the entire board tonight and it is being declined. The insurer has run 35% in thirty days and 42% year-to-date, which is exactly why the momentum signal fired - but the underlying business does not support paying up: return on of 11.7% is weak, earnings have gone nowhere over five years, the dividend is not covered by free cash flow, and the shares trade at 17.3x earnings against an 11.2x industry average. Our own model's twelve-week estimate of Sis essentially identical to an independent cash-flow appraisal's Sfair value - meaning the buy-stop at Swould be paying above what the business is worth, with no upside on any horizon. OPERATIONAL: HARD QUALITY GATE - Quality_Pass FALSE (F_Score, Quality_Rank 16/26, negative Earn_Yield -2.26%) blocks NEW entries outright; the gate never blocks ADDs to held names, so this reject does not govern the sgotr book which already holds G07 (L7 cross-book note). Fusion would otherwise have cleared: 71 raw - 3 val_adj (MoS -0.668 x 5) + 6 bias (MomentumContinuation_Up|buy, n=5, hit_1w 0.80, ACTIONABLE) = 74, above the 65 floor. Independently rejected on three more grounds: RSI 81.9 (most overbought row in the universe); ALL THREE estimates sit BELOW the 22.96 buy-stop entry (1wk 22.69 -1.2%, 4wk 22.28 -3.0%, 12wk 22.34 -2.7%) and below the 22.74 close, so the trigger buys into a negatively-projected path; and a 35c dividend goes ex on 08-19, a mechanical -1.5% inside the queue window. Re-arm only on a pullback toward the EMA20 (21.27) with the quality screen repaired. · news |
| O39 | OCBC Bank | 31.55 | HOLD | — | 30.09 | 31.40 | 31.04 | 31.72 | OCBC remains the one position in this book with a real, it is up 21.7% on that cost, and the decision to bank part of the gain still stands - but the trade cannot be published again because the holdings file has not been updated in five weeks and the same card has already been issued three times. Nothing changes until the executed trades are recorded. OPERATIONAL: SG-8 blocking sync flag, cycle 4. The ADR-0012 profit-take still FIRES mechanically on 2 ticks - (a) RSI_DIV present at RSI 78.86, (c) dead 12wk 31.72 < 1.02 x 31.55 = 32.18; tick (b) does NOT fire, conviction ROSE, and (d) does not fire - which maps to profit_take_2 50%. Per SG-8 the card is NOT re-published and the size is NOT re-escalated from the 30% last carried. Genuinely armed (Avg_Cost 25.9325, Gain_ATR 8.22, T12_Progress 99.5%) unlike other rows whose PROFIT_ARMED is spurious. Note for the counterfactual grader: the unexecuted trim has cost nothing so far - O39 was 31.50 when first re-published on 08-13 and is 31.55 tonight, and conviction has since STRENGTHENED to 78. Stop rebased 30.22 -> 30.09 (live model stop) for the 47c interim dividend that went ex today; the tape applied the dividend itself (31.79 -> 31.55 = -0.24 against a 47c ex, so O39 rose ~0.7% ex-div), so no phantom breach. Unblocks the moment input/sgportfolio.json is synced. |
| B61 | BT SEMBAWANG EST | 4.49 | HOLD | — | 4.62⚠ | 4.48 | 4.49 | 4.54 | Bukit Sembawang has spent seven straight sessions below its risk line but the exit decision is stuck behind the same unrecorded holdings file as the rest of the book's sells, and the evidence says the exit was never a strong one anyway - the stock screens as the cheapest name here on an asset basis and the twelve-week view still points slightly higher. OPERATIONAL: SG-8 blocking sync flag, cycle 3. 7th consecutive close below 4.62 (4.49, -2.81%) on a Sposition. NoTrade|sell is ACTIONABLE at -5 (n=7, hit_1w 0.25, hit_4w 0.00) and Est_12W 4.54 is +1.1% ABOVE close, so the L2 combo does NOT fire here even though the stop-break limb does - unlike U11, where the 12wk limb crossed. RSI 34.3 is oversold. UNDERVALUED + MoS +0.241 but Quality_Pass FALSE. Stop held at 4.62. |
| HMN | CAPITALAND and ASCOTT STAPLE U | 0.86 | HOLD | 0.86 | 0.88⚠ | 0.86 (+0.0%) | 0.86 (+0.0%) | 0.88 (+2.3%) | Ascott Trust printed its loudest sell signal yet tonight and has been below its risk line for nine sessions running, but at Sthe position is the smallest in the book and the exit is blocked behind the same unrecorded holdings file. OPERATIONAL: SG-8 blocking sync flag, cycle 7. STRONG SELL - REDUCE cv71 at Sell_Rank #1, the strongest print this name has produced, on FailedBreakout_Down_20D whose sell cell is ACTIONABLE at exactly 0 (n=73, hit_1w 0.507 - a coin flip after 73 samples). Note SG-2's second limb still FAILS independently: Est_12W 0.88 is +2.3% ABOVE close, so even unblocked this would not clear the confirmation test. Stop held at 0.88. |
| A17U | CapitaLand Ascendas REIT | 2.44 | HOLD | 2.44 | 2.42 | 2.44 (+0.0%) | 2.46 (+0.8%) | 2.45 (+0.4%) | Ascendas REIT carries the board's second-loudest sell label and screens as the most expensive holding in the book against what the assets earn - but it has not broken its risk line and every forward estimate still points sideways-to-up, so there is no confirmed exit case. At Sthe position is also too small to matter either way. Holding. OPERATIONAL: SG-2 confirmation test FAILS on both limbs - close 2.44 vs the rebased 2.42 stop is NO break, and Est_12W 2.45 is +0.4% POSITIVE (1wk flat, 4wk +0.8%). Adjusted conviction 66 = 69 raw - 3 (RallySell_BearTrend|sell ACTIONABLE, n=9, hit_1w 0.333) does clear the 65 floor on arithmetic, but the floor is a necessary not a sufficient condition and SG-2 governs SG sell labels, which fade ~54% of the time. Valuation MoS -7.685 is deeply negative and would amplify a signal-based sell one tier, but valuation NEVER originates and there is no confirmed signal to amplify (CLAUDE.md-9). Would have been 80.07% x 400 = = SRe-arm: a decisive close below 2.42 on a NEGATIVE Est_12W. |
| J69U | FRASERS CPT TR | 2.19 | HOLD | — | 2.18 | 2.19 | 2.21 | 2.25 | Frasers Centrepoint has flipped from a sell label to an oversold-bounce read, vindicating eleven consecutive refusals to dump it - the twelve-week view is +2.7%, the best forward shape among the REITs. Holding the full stake. OPERATIONAL: label flipped STRONG SELL - REDUCE cv61 -> HOLD cv41 on ReversalUp_Oversold; RSI 34.1 oversold. Close 2.19 vs the 2.18 stop - one cent above, still no break after eleven cycles of declining this sell. Model stop 2.14 is LOWER than the recorded 2.18, so the stop is HELD at 2.18, never lowered. No valuation row (NA), so no amplify path exists. This remains the largest sell the board has repeatedly offered (S) and declining it has been correct. |
| D05 | DBS Group Holdings | 76.88 | HOLD | 77.22 | 74.47 | 76.41 (-1.0%) | 75.99 (-1.6%) | 79.13 (+2.5%) | DBS led the book today, up 1.8% to a fresh 52-week high and fully recovered from last week's 81c dividend, and it now carries the strongest twelve-week view of the three banks at +2.9%. No add - the signal is only a weak hold and the book's rule for bank cores is to add on pullbacks, never on extension - but the protective stop is being raised to lock in more of the run. OPERATIONAL: HOLD cv46 on BreakoutUp_52W, faded from cv51; NOT in the buy pool. Trail RAISED 73.04 -> 74.47 (live model stop, -3.1% from close). Exit_Warning RSI_DIV at RSI 72.0 plus a SPURIOUS PROFIT_ARMED (SG-7, no Avg_Cost) - with no real no profit-take can fire, and G1 would downgrade an RSI_DIV on a 52-week-high breakout to a trail-raise anyway, which is exactly what this is. 28.2% of book weight, the largest single position. FOMC minutes 08-19 and the Warsh Jackson Hole keynote 08-28 are both +/-2.0-2.5% two-sided on the bank leg - hold through, the trail does the work. SG-4 governs: re-arm on a non-extended BUY - ADD reprint at cv>=65. |
| S63 | ST Engineering | 11.13 | HOLD | 11.00 | 10.32 | 11.07 (+0.6%) | 11.15 (+1.4%) | 10.94 (-0.5%) | ST Engineering woke up from three quiet sessions with a 1.6% gain and the second-highest conviction reading on the board, but the signal is a hold rather than a buy, the shares are already extended past the model entry, and the twelve-week view is slightly negative - so the response is a materially tighter stop, not an add. OPERATIONAL: HOLD cv74 on AnticipationUp_VCP, up from NO TRADE cv0 on the three prior settles. Trail RAISED 9.58 -> 10.32 (live model stop, -7.3% from close). Entry_Distance_ATR -0.38 (extended past the 11.00 model entry); Est_12W 10.94 is -1.7% BELOW close, which is why no BUY-side case is opened. Deeply GROWTH_PRICED_IN at MoS -9.067 (val_adj -10) - amplify-only, no signal-based sell to amplify. 4c ex-dividend 08-25 (-0.36%), the only certain move in the 10-day Event Horizon; rebase the stop then. PROFIT_ARMED spurious (SG-7). |
| BUOU | FRASER L&C TRUST | 0.96 | HOLD | — | 0.96⚠ | 0.96 | 0.96 | 0.98 | Frasers Logistics is sitting exactly on its stop after a small decline, but it has not broken it and the twelve-week view is still 2.1% higher, so there is no exit case yet. This is the book's fourth largest position and the next one likely to trigger. OPERATIONAL: NO TRADE cv0 ('Consensus conflict + low conviction'); close 0.960 vs stop 0.960 = 0.00% breach, touching not breaking. L2 does NOT fire - the negative-12wk limb fails (Est_12W 0.98, +2.1%). NoTrade|sell is ACTIONABLE at -5. MoS -6.033 is amplify-eligible (<= -3, would bump a trim one tier) but there is NO signal-based sell to amplify, and SG-3 says read the magnitude not the FAIR_VALUE band on SG REITs. S8.7% weight. A decisive close below 0.96 on a NEGATIVE Est_12W makes this the immediate lead dispose - the same pre-registered trigger structure that fired on U11 tonight. FOMC minutes 08-19 and Jackson Hole 08-28 are +/-2.5-3.0% two-sided on the S-REIT leg: tighten-trail stance. |
| N2IU | Mapletree PanAsia Com Tr | 1.27 | HOLD | 1.27 | 1.27⚠ | 1.27 (+0.0%) | 1.28 (+0.8%) | 1.28 (+0.8%) | Mapletree PanAsia is back to touching its stop after a 0.8% slip but has not broken it, and the forward view is mildly positive - no action on a Sposition. OPERATIONAL: AVOID cv44 on FailedBreakout_Down_20D (cell ACTIONABLE at 0, n=73). Close 1.270 vs stop 1.270 = 0.00% breach; Est_12W 1.28 is +0.8% ABOVE close so L2's second limb fails. The printed model Stop_Price 1.32 sits ABOVE the close (a bearish-family short-side level) and is NOT adopted - the recorded 1.27 stands. S-REIT leg: tighten-trail into 08-19 and 08-28. |
| ME8U | Mapletree Industrial Trust | 1.91 | HOLD | — | 1.89 | 1.91 | 1.90 | 1.89 | Mapletree Industrial was unchanged today and sits just above its stop with a flat forward view - the book's third-largest holding, quiet, and left alone. OPERATIONAL: NO TRADE cv0; close 1.910 vs stop 1.89, +1.1% clear. Est_12W 1.89 is -1.0% below close but with no stop break the L2 combo needs both limbs and only one is present. MoS -17.428 is the book's second-most-negative valuation - SG-3 says read the magnitude, and it is amplify-eligible the moment a signal-based sell appears, but valuation never originates. S12.0% weight. Watch the 1.89 line into the 08-19 minutes. |
| C07 | Jardine C&C | 27.65 | HOLD | — | 26.38 | 27.76 | 28.39 | 28.54 | Jardine Cycle & Carriage remains the single cheapest quality name in the book - the only holding that is undervalued, passes the quality screen and ranks first on composite quality - with a cash-flow value more than double the share price and a +3.2% twelve-week view. It still produces no tradable signal, so it stays on the watch list rather than becoming a buy. OPERATIONAL: NO TRADE cv0 ('Consensus conflict + low conviction'), so adjusted conviction is only 10 despite the maximum +10 val_adj (MoS +0.628, Quality_Pass TRUE, Quality_Rank 1). RSI 47.8, neutral. Re-arm on ANY BUY-family reprint at cv>=55. Note J36's ex-dividend on 08-20 is a mildly correlated drag (low confidence). |
| U13 | UOI LTD | 8.75 | HOLD | — | 8.03 | 8.75 | 8.84 | 8.85 | UOI slipped 0.5% but remains well clear of its stop with a positive forward view and no signal - a quiet Shold. OPERATIONAL: NO TRADE cv0 ('Consensus conflict + low conviction'); close 8.75 vs stop 8.03, +9.0% clear. Est_12W 8.85, +1.1%. One of only two names (with B61) that the Event Horizon cannot cover - touched by whole-book structural events only. PROFIT_ARMED spurious (SG-7). |
| 9CI | CapitaLand Investment | 2.74 | HOLD | — | 2.62 | 2.73 | 2.75 | 2.84 | CapitaLand Investment was unchanged and holds the best twelve-week view among the book's quiet names at +3.6%, but prints no signal. Holding. OPERATIONAL: NO TRADE cv0; close 2.74 vs stop 2.62, +4.6% clear. RSI 67.5. The 08-13 intraday BUY - ADD cv70 has still not reprinted on a settle (L4/L6 - no gap-chase). Watch for a non-extended reprint toward 2.65. |
Outcome & track record
Accuracy at the time of this record.
116 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.