Recommendations
0 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 |
|---|---|---|---|---|---|---|---|---|---|
| 8AZ | Aztech Global | 0.57 | WATCH | 0.57 | 0.55 | n/a | 0.65 (+14.0%) | 0.33 (-42.1%) | Aztech is cheap on our valuation work and now deeply oversold, but for the fourth session running the model has cut its own score on exactly the same bounce setup - 80 on the sixth of August, 68, 63, and 58 tonight - while the shares fell another 0.9% to 0.575. We decline it a fourth time on unchanged grounds: it fails our financial-strength screen, which bars a new position outright, and our own twelve-week view sits forty-three percent below the current price. The engine has now spent a fortnight converging on the position we took two weeks ago. OPERATIONAL: FIRST NIGHT IN FOUR THAT adj SITS GENUINELY BELOW THE 65 FLOOR - adj 60, so the three-night 'adj sits exactly AT the floor, do not report as a conviction near-miss' caveat is RETIRED. Three blocks in precedence order: (1) WATCH is not a bullish actionable Action, so this row never enters the buy pool; (2) Quality_Pass False = hard gate on NEW entries, which also blocked adj 82 on 08-06 and cannot be out-scored; (3) T12 0.33 vs 0.575 close = -42.6%. adj 60 = cv58 + val_adj +2 (UNDERVALUED MoS +0.220 x20 = +4, HALVED for failed quality) + bias 0. Conviction decay now monotonic and unbroken: ->, a 22-pt collapse over a -7.3% price move. RSI 17.76 = new low for the move; Entry_Distance_ATR -0.31 puts price PAST the model entry, a no-chase condition in its own right. Liquidity gate PASSED cleanly - 2.609% range on SGD 733,930, comfortably inside the now-source-confirmed SG spec (turnover >= 100,000 AND range <= 3.000%). Valuation drift caveat: the FV surface is 6 days old and the basket has fallen 6.37% against it, so a re-mark would lift val_adj to roughly +3 and adj to 61 - still below floor AND still hard-blocked by the quality gate, so NON-BINDING. REFUSAL LEDGER: 08-06 BUY cv80/adj82 leg, card 0.62, L8 exec 08-07 open 0.6200 vs 0.575 tonight = -7.26% avoided (improved from -6.45%). RE-ARM unchanged: leg one (breakdown must fail first) SATISFIED; leg two binding and unmet - FY2026 revenue stabilisation repairing Quality_Pass, THEN a settled favoured-cell BUY >= 65 with a forward view not forty percent below price. Do NOT re-derive as a live BUY on leg one alone (SGM-4). · news |
| 5DD | Micro-Mechanics (Holdings) | 2.81 | WATCH | 2.81 | 2.77 | n/a | 2.96 (+5.3%) | 2.59 (-7.8%) | Micro-Mechanics was the single best-behaved name on our board tonight - the only one whose score rose and one of just three that gained, up 1.8% to 2.81 on a broadly negative session. We are still not buying it. The setup is a bounce-off-a-failed-breakdown pattern that our own trading record marks as one of the least reliable we track, the shares already look fully valued against our fundamental work, and our twelve-week projection sits below tonight's price. A name that is the best of a weak board is not thereby a buy. OPERATIONAL: adj 56 = cv63 (up from 58) + val_adj -7 (GROWTH_PRICED_IN MoS -1.426 x5) + bias 0 - nine points short of the 65 floor. Two independent blocks: (1) WATCH is not a bullish actionable Action, so the row never enters the buy pool; (2) L1 / MSCAP-1 explicitly discount FailedBreakdown_Up_20D chases (~0.25 hit rate, ~60% stop-first in the reference cell) - this is precisely the family the book was told not to starter cash into. Forward view also negative: T12 2.59 vs 2.810 close = -7.8%. Liquidity gate PASSED but MARGINALLY - 2.491% range on SGD 208,221, barely twice the now-source-confirmed SGD 100,000 turnover floor and the second-thinnest tape on the board. Entry_Price 2.81 = the close exactly; stop 2.77; 4wk target 2.96. MSCAP-2 caveat applies (x5 curve over-penalises a mild MoS) but is IMMATERIAL - even at val_adj 0 this prints 63 and still misses the floor, and WATCH bars it from the pool regardless. CONVICTION INSTABILITY NOTED: five-session series NT -> -> on closes 2.780/2.780/2.760/2.760/2.810 - a 20-pt conviction range over a 1.8% price range including a 15-pt round trip at an unchanged close (backlog item 3). RE-ARM: a settled PullbackBuy_VWAP or MomentumContinuation BUY >= 65 with a positive forward view; do not re-derive on a FailedBreakdown print at any conviction. · news |
| 575 | ASTI Holdings | 0.11 | WATCH | — | — | n/a | n/a | 0.10 | ASTI lost the tentative signal it showed on Friday after a single session and fell another 3.4% to 11.3 cents. It looks expensive against our fundamental work and our twelve-week view sits below the current price. No action. OPERATIONAL: WATCH cv55 (08-21) -> NO TRADE cv0 tonight, carrier status lost after one session. THIS ROW IS A LIQUIDITY FAIL MISREPORTED AS CONSENSUS - 4.425% range on SGD 372,572 turnover breaches the source-confirmed SG cap of 3.000%, but Pillar_Consensus is Indeterminate with score 0, so the post-conviction gate at engine.py:710 overwrote the reason string to 'Consensus conflict + low conviction' (backlog item 2). Turnover leg passed comfortably; this is a volatility rejection, not an illiquidity one. RSI 38.6. T12 0.10 vs 0.113 close = -11.5%. GROWTH_PRICED_IN MoS -2.385 -> val_adj -10 (CLAMPED - MSCAP-2 saturation again: 575 at -2.385 and AWX at -10.844 both clamp to an identical -10 despite a 4.5x difference in stated mispricing). QPass True. The internally-inconsistent print flagged 08-21 (4-week Target_Price 0.13 ABOVE the 12-week projection 0.11) has resolved by the 4-week target disappearing entirely with the signal. · news |
| S71 | Sunright | 0.43 | WATCH | — | — | n/a | n/a | 0.25 | Sunright is the cheapest name in this mandate on our valuation work and it remains uninvestable. It traded roughly ten thousand Singapore dollars of stock all day and fell 6.5%, the worst decline on our board, and our own twelve-week projection sits forty-two percent below the current price. Cheapness alone has never started a position in this book and does not tonight. OPERATIONAL: NO TRADE cv0. THE ONLY NAME IN THIS BOOK THAT HAS EVER FAILED THE TURNOVER LEG - SGD 10,406 on is a tenth of the source-confirmed SGD 100,000 floor, and with a 6.977% range it is a clean FAIL_Both. Closed 0.430 (-6.52%), RSI 32.1. UNDERVALUED MoS +0.459 -> val_adj +9, the largest positive valuation adjustment in the book, and it buys nothing: no chart, no signal, T12 0.25 vs 0.430 close = -41.9%. SGM-5 IN ITS PUREST FORM - valuation never originates a position (CLAUDE.md section 9). RE-ARM: a settled favoured-cell BUY >= 65 with a POSITIVE forward view; the negative T12 is the binding objection, not the valuation. A name that trades SGD 10k a day is uninvestable at any size this book would - the practical re-arm is a sustained turnover regime change, not a chart event. · news |
| M14 | InnoTek | 0.44 | WATCH | — | — | n/a | n/a | 0.25 | InnoTek rose 1.1% for a second session, but the fall it is bouncing from was an earnings collapse - first-half profit down sixty-four percent - not a technical dip. Our twelve-week view sits forty-three percent below the current price and the stock carries no signal. We are not chasing a bounce off deteriorating fundamentals. OPERATIONAL: NO TRADE cv0, clean Liquidity gate failed (FAIL_Spread, 5.682% range on SGD 716,936; turnover leg passes comfortably). Closed 0.440 (+1.15%), second consecutive gain. RSI 33.2, still deeply oversold. T12 0.25 vs 0.440 = -43.2%, the WORST forward view in the book tonight. GROWTH_PRICED_IN MoS -0.330 -> val_adj -2, QPass False. L6 APPLIES IN ITS LITERAL FORM: do not chase the post-earnings gap - the news finding from 08-21 stands unrevised (1H FY2026 attributable profit fell 64% to SGfrom SGshares -10%+ on the release). REFUSAL LEDGER: 08-12 WATCH cv68 leg, card 0.51, L8 exec 08-13 open 0.510 vs 0.440 tonight = -13.73% avoided, WORSE than last night's -14.71% as the bounce continues. This is the honest entry in the ledger: two consecutive up sessions have shrunk the book's largest favourable leg by five points. Recorded, never permitted to re-litigate the 65 floor (MSCAP-12). · news |
| I07 | ISDN Holdings | 0.68 | WATCH | — | — | n/a | n/a | 0.75 | ISDN fell another 5.6% and the product-launch spike that lifted it in mid-August is now completely unwound - the shares are back below where they started and thirteen percent off the high. We identified that move as a marketing announcement rather than an order-book or earnings event and declined to chase it; that judgement has now played out in full. No signal, no action. OPERATIONAL: NO TRADE cv0, clean Liquidity gate failed (FAIL_Spread at 7.407% - the WIDEST range on the board tonight - on SGD 4,877,415; turnover leg passes by 49x). Closed 0.675 (-5.59%), FOURTH consecutive decline. The 08-19 Automation Taipei launch spike (ApexTwin MFF dual-gantry motion system, 33mm vacuum low-profile precision module) is fully round-tripped: 0.775 -> 0.675 = -12.9% from the high, RSI 67.2 -> 47.8, VWAP distance collapsed from +9.68%. T12 0.75 vs 0.675 = +11.1%, one of only four positive forward views left in the book, but eroding (0.87 -> 0.90 -> 0.89 -> 0.84 -> 0.75). GROWTH_PRICED_IN MoS -0.742 -> val_adj -4, QPass False. Waiting was correct and we now know WHY it was correct rather than merely THAT it was. · news |
| AWX | AEM Holdings | 9.14 | WATCH | — | — | n/a | n/a | 9.98 | AEM fell for a fifth straight session, down 2.4% to 9.14, and is now eight percent below where we declined to chase it a fortnight ago. The business is fine - guidance was reaffirmed and the stock is still up sharply on the year - but the shares are unwinding an extended run, they look richly priced against our fundamental work, and our own twelve-week projection has been cut in every one of the last five sessions. This is the name most exposed to NVIDIA's results on Wednesday, which SGX prices on Thursday. We wait for that. OPERATIONAL: NO TRADE cv0, clean Liquidity gate failed (FAIL_Spread, 4.705% range on SGD 26,403,633 - the deepest tape in the book; the turnover leg passes by 264x, confirming this is a VOLATILITY rejection, not an illiquidity one). Closed 9.140 (-2.35%), fifth consecutive decline. RSI 47.0, clearly below 50 now. T12 9.98 = +9.2%, eroding monotonically and hard: 12.22 -> 11.90 -> 11.38 -> 10.65 -> 9.98, the projection has shed 18% in five sessions while the price fell 8.2%. GROWTH_PRICED_IN MoS -10.844 -> val_adj -10 (CLAMPED; MSCAP-2 saturation). QPass False. REFUSAL LEDGER, BOTH LEGS IMPROVED: 08-12 floor leg card 9.63, L8 exec 08-13 open 10.00 vs 9.140 = -8.60% avoided (from -6.40%); 08-13 chase leg card 10.98, L8 exec 08-14 open 11.00 vs 9.140 = -16.91% avoided (from -14.91%). These two legs alone account for 25.5 of the 49.3 percentage points of avoided drawdown across the book's five open refusals. EVENT: NVIDIA Q2 FY2027 on 08-26 after the US close, SGX reaction 08-27 - the key read-through, 2 sessions out. · news |
| PCT | PC Partner Group | 2.78 | WATCH | — | — | n/a | n/a | 3.37 | PC Partner remains the strongest fundamental case in this mandate - trading essentially at our estimate of fair value, passing our financial-strength screen, and carrying the best twelve-week upside on the board at plus twenty-one percent - and for the twelfth time in thirteen sessions our system will not let us buy it. It gained 0.7% tonight. The blocker is not the business and not the price; it is that the stock swings too widely intraday to pass our execution filter. We keep it first on the watchlist. OPERATIONAL: NO TRADE cv0, FAIL_Spread at 3.957% against the now-source-confirmed SG cap of 3.000%. CRITICAL RECLASSIFICATION: this is definitively a VOLATILITY rejection, NOT an illiquidity one - PCT turned over SGD 4,133,860 tonight and has NEVER failed the SGD 100,000 turnover leg in thirteen sessions. The 'liquidity gate' label has been misleading this book's reading of PCT for a fortnight. Closed 2.780 (+0.73%), RSI 47.5. T12 3.37 = +21.2%, BEST forward view in the book, though eroding (4.09 -> 4.05 -> 3.66 -> 3.45 -> 3.37). FAIR_VALUE MoS -0.020 -> val_adj 0, QPass True - the only name in the book with a clean valuation AND quality profile. PCT IS THE NAME MOST DIRECTLY PENALISED BY SG'S LEGACY FIXED-3% spread_pct SPEC (backlog item 1: US and HK were migrated to volatility-aware range_to_atr as the NBIS-class fix; SG was left on the legacy cap on the strength of a large-cap sweep). Proposed remedy is a shadow-tested small-cap segment overlay - NEVER a hand-tuned threshold (forward-test rule). REFUSAL LEDGER: 08-06 floor leg cv62, card 2.74, L8 exec 08-07 open 2.86 vs 2.780 = -2.80% avoided, worse than last night's -3.50%; this leg has now flipped sign or magnitude in six of seven sessions and is the least informative of the five. · news |
| 558 | UMS Integration | 2.66 | WATCH | — | — | n/a | n/a | 3.15 | UMS fell 4.0% and the stretched condition we have flagged for two weeks is finally unwinding. It still carries the second-best twelve-week upside in the book, but it looks expensive against our fundamental work and it fails our financial-strength screen, so a chart improvement alone cannot make it buyable. Like AEM, it is a direct NVIDIA read-through name and Wednesday's results are the real test. No action. OPERATIONAL: NO TRADE cv0, clean Liquidity gate failed (FAIL_Spread, 5.639% range on SGD 16,873,179; turnover leg passes by 169x - volatility rejection, not illiquidity). Closed 2.660 (-3.97%). RSI cooled sharply 60.0 -> 53.3, the extension unwinding at last. T12 3.15 = +18.4%, second-best forward view here but eroding (3.48 -> 3.46 -> 3.38 -> 3.15). SGM-4a DOUBLE BLOCK PERSISTS: GROWTH_PRICED_IN MoS -4.559 -> val_adj -10 (clamped) AND Quality_Pass False. Needs a valuation AND quality re-rate, not a chart event - do not re-derive as a live BUY on a chart turn alone. EVENT: NVIDIA Q2 FY2027 08-26, SGX reaction 08-27, 2 sessions out. · news |
| E28 | Frencken Group | 2.50 | WATCH | — | — | n/a | n/a | 2.03 | Frencken fell 3.1% to 2.50 and our twelve-week projection has now dropped to nineteen percent BELOW the current price - a signal here would mean buying against our own forward view by nearly a fifth. The valuation is reasonable and the balance sheet passes our screen, but the trend and the projection are both pointing down. No action. OPERATIONAL: NO TRADE cv0. THIS ROW IS PROVABLY A LIQUIDITY FAIL MISREPORTED AS CONSENSUS - 4.400% range breaches the source-confirmed 3.000% SG cap, on SGD 10,003,500 of turnover (turnover leg passes by 100x), but Pillar_Consensus is Indeterminate with score 0, so the post-conviction gate at engine.py:710 overwrote the reason to 'Consensus conflict + low conviction' (backlog item 2 - this is the entire 'masking' anomaly logged as unknown across five prior entries, now located exactly). RSI 42.3 and falling. T12 2.03 vs 2.500 = -18.8%, third-worst forward view in the book and deteriorating monotonically: 2.33 -> 2.39 -> 2.25 -> 2.18 -> 2.03. FAIR_VALUE MoS -1.341 -> val_adj -7; MSCAP-2 flags this as an over-penalty on a fair-valued small-cap (the x5 curve), FLAGGED NOT OVERRIDDEN - and immaterial at cv0. QPass True. · news |
| BN2 | Valuetronics Holdings | 1.02 | WATCH | — | — | n/a | n/a | 0.97 | Valuetronics closed unchanged for a third consecutive session at 1.02 and has no signal in either direction. Our twelve-week view is modestly below the current price. Nothing to do. OPERATIONAL: NO TRADE cv0 for a third session, bearish carrier fully extinguished (FailedBreakout_Down_20D AVOID cv48/cv58 on 08-18/08-19 -> cv0 since). TONIGHT'S CLEAN CONTROL CASE FOR THE MASKING FINDING: BN2 PASSED the liquidity gate honestly - 1.961% range on SGD 490,314, both legs clear of the source-confirmed SG spec - so its 'Consensus conflict + low conviction' label is GENUINE, unlike E28 and 575 which were overwritten at engine.py:710. This is what a real consensus block looks like and it is what made the defect diagnosable. Closed 1.020 flat on. T12 0.97 = -4.9%. GROWTH_PRICED_IN MoS -0.136 -> val_adj -1, QPass False. Long-only + => no action either way (SGM-4b): a lone AVOID print confirms an empty buy pool, it never populates it. · news |
Outcome & track record
Accuracy at the time of this record.
0 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.
0.000 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.