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 |
|---|---|---|---|---|---|---|---|---|---|
| PCT | PC Partner Group | 2.86 | WATCH | — | — | n/a | n/a | n/a | Two days ago PC Partner gave us the best setup this book has produced in five weeks and we passed on it, three conviction points short of our bar. Tonight the market answered the question for us. The shares rose 2.9% to 2.86 - and traded up to exactly 2.95 intraday, the precise level our buy-stop would have sat at - then gave the move back and closed 3.1% below that touch. Had we bridged that three-point gap by hand and queued the order, we would be underwater tonight on the very trade the near-miss made tempting. Separately, the signal itself has gone: the stock's daily trading range widened to 3.50% against our 3.00% liquidity ceiling, so the model can no longer see the name at all and prints no signal, no entry and no stop. The investment case is untouched and remains the best in the book - cheap at a 7.7% discount to our estimate of fair value, second-highest quality score of the eleven names, and a twelve-week view of +16.8%. We are waiting on tradeability, not on the thesis. OPERATIONAL: BUY cv60 (08-26) -> NO TRADE cv0. Gate FAIL_Range: 3.497% range vs the 3.000% cap (0.497pp over) on SGD 4,217,069 turnover - the turnover leg passes comfortably, the range leg does not. NoTrade_Reason correctly reads 'Liquidity gate failed' (Pillar_Consensus Consolidation_Bullish, so this row is NOT affected by the engine attribution defect). REFUSAL LEDGER, NEW LEG: the 08-26 near-miss was a buy-STOP at 2.95; tonight's High was 2.9500 exactly, so the order fills, and the 2.86 close marks it -3.05%. MSCAP-21's 'a near-miss is evidence for the champion/challenger shadow test, not a gap to bridge' is vindicated on its first observable test. Val_adj +2 (UNDERVALUED MoS +0.0766 x20 = +2, QPass True, unhalved); bias 0 (no actionable cells). No estimates recorded - the row prints no Entry_Price, Stop_Price or Target_Price, and Step 7 forbids synthesising a level the live row does not print. RE-ARM UNCHANGED (MSCAP-21): a settled PullbackBuy_VWAP / MomentumContinuation at raw cv >= 63 that ALSO clears the gate. A news pass becomes mandatory before any card. L6 continues to bar a post-catalyst gap-chase into the Mon 08-31 MSCI review auction. · news |
| 8AZ | Aztech Global | 0.57 | WATCH | — | — | n/a | n/a | n/a | Aztech is the only name in the book carrying any conviction tonight, and it is pointing the wrong way. The model rates it AVOID on a failed-breakout-down pattern, with momentum at an extreme low (RSI 20) and a twelve-week view 42.6% BELOW today's price. The shares do screen cheap - roughly a 26% discount to our fair-value estimate - but the company fails our fundamental quality screen outright, and we do not buy a falling business on price alone. We run this book long-only, so an AVOID generates no trade in either direction; it simply confirms that there is nothing to buy. OPERATIONAL: AVOID cv56, adj 58 (val_adj +2 = MoS +0.2591 x20 = +5, HALVED on Quality_Pass False). SGM-4 governs: a lone AVOID print CONFIRMS an empty buy pool, it never populates it. MSCAP-31 governs the reporting order: the FIRST binding block is the HARD Quality_Pass gate on new entries, not the 65 floor and not the direction. Gate PASS (2.609% range, SGD 789,934) - one of only rows to clear tonight. T12 0.33 vs 0.575 close = -42.6%. Trail 08-18 -> 08-28: W68 -> W58 -> NT0 -> W58 -> AVOID56. RE-ARM (SGM-4, unchanged, LEG TWO STILL BINDING): FY2026 revenue stabilisation repairing Quality_Pass FIRST, then a settled favoured-cell BUY >= 65 with a forward view that is not forty percent below price. |
| 558 | UMS Integration | 2.67 | WATCH | — | — | n/a | n/a | n/a | UMS carries the book's second-best forward view - our twelve-week estimate sits 14.2% above tonight's 2.67 close - but it is doubly blocked. The stock is priced well beyond what we think the business is worth on current earnings, and it fails our quality screen, so even a strong chart would take a maximum valuation penalty. Tonight the point is moot: its daily range blew out to 4.87% against our 3.00% liquidity ceiling and the model sees no signal at all. It surfaced on 08-26 only because its range briefly squeezed under the cap by five hundredths of a percentage point - cadence, not a change in the case. OPERATIONAL: WATCH cv52 (08-26) -> NO TRADE cv0. Gate FAIL_Range 4.869% vs the 3.000% cap on SGD 20,300,544. SGM-4a double block intact: GROWTH_PRICED_IN MoS -4.3424 -> val_adj -10 (clamped) AND Quality_Pass False (a HARD gate on new entries). T12 3.05 vs 2.67 = +14.2%, second-best in the book, and one of the NVIDIA supply-chain read-through names. Close -1.48% over the two sessions since 08-26. |
| AWX | AEM Holdings | 8.85 | WATCH | — | — | n/a | n/a | n/a | AEM is the deepest-traded name in this book and the one we came closest to buying earlier in the month - and staying out has kept getting more right. It fell another 2.3% to 8.85, so the entry we declined on 12 August is now 11.5% above the market and the chased version from 13 August is 19.6% above it. Tonight brings a genuine deterioration rather than just a lower price: our twelve-week estimate has crossed below the share price for the first time in this window, at 8.68 against an 8.85 close. The stock also remains expensive against what the current business earns. No signal tonight - the daily range widened to 5.31%, well outside our liquidity ceiling. OPERATIONAL: WATCH cv45 (08-26) -> NO TRADE cv0. Gate FAIL_Range 5.311% on SGD 45,221,731 (the book's largest tape). ENGINE DEFECT ROW: NoTrade_Reason misreports 'Consensus conflict + low conviction' over a provable range failure - Pillar_Consensus is Indeterminate, and see this session's systemic finding #1 (the label is a pure function of Pillar_Consensus, 8 sessions / 88 rows, 100% separation). T12 FLIPPED NEGATIVE: 9.40 -> 8.68 vs 8.85 = -1.9% (was +3.8% on 08-26), taking the book's positive-T12 count. val_adj -10 (clamped) and Quality_Pass False. Two open refusal legs, both WIDENED tonight. |
| E28 | Frencken Group | 2.35 | WATCH | — | — | n/a | n/a | n/a | Frencken was the worst performer in the book over the last two sessions, down 7.5% to 2.35, and it now trades below every trend line we track. Our twelve-week estimate sits 26% BELOW the current price - the weakest forward view of any name that is not already down more than forty percent. There is nothing to buy here and, with no position, nothing to sell. No signal in any case: a 5.96% daily range is double our liquidity ceiling. OPERATIONAL: NO TRADE cv0, fourth consecutive session. Gate FAIL_Range 5.957% (the widest in the book tonight) on SGD 22,704,289. Close 2.35 vs EMA20 2.588 / EMA50 2.672 / EMA200 2.335 - sitting on the 200-day. RSI 36.4. T12 1.74 vs 2.35 = -26.0%. val_adj -6 (FAIR_VALUE MoS -1.2266 x5); MSCAP-2 note stands - the FAIR_VALUE band paired with a -6 haircut remains the miscalibration this book keeps flagging. |
| 5DD | Micro-Mechanics (Holdings) | 2.80 | WATCH | — | — | n/a | n/a | n/a | Micro-Mechanics was flat over the two sessions and lost its signal entirely, missing our liquidity ceiling by two tenths of a percentage point - the narrowest miss in the book. That is a reminder of how marginal this universe is, not a reason to act: the name has swung between conviction 43 and 63 and back to nothing over a price range of under two percent, which is an argument against it rather than for it. The twelve-week view is 6.1% below the current price. OPERATIONAL: WATCH cv48 (08-26) -> NO TRADE cv0. Gate FAIL_Range 3.214% vs the 3.000% cap - a 0.21pp miss, the tightest of the nine failures - on SGD 393,399. MSCAP-34 extends to an EIGHTH data point: NT -> -> -> -> NT0 across a ~1.8% price range. T12 2.63 vs 2.80 = -6.1%. val_adj -7. L1/MSCAP-1 continue to discount its FailedBreakdown family. |
| I07 | ISDN Holdings | 0.69 | WATCH | — | — | n/a | n/a | n/a | ISDN edged up 0.7% to 0.69 and holds one of only three positive forward views left in the book, with our twelve-week estimate 5.8% above the current price. It is also the mildest valuation penalty of any name here. But it has not produced a tradeable signal in this window, it fails our quality screen, and tonight its 3.62% daily range again puts it outside what we can trade. A watch item, not a buy. OPERATIONAL: NO TRADE cv0, unbroken across all four snapshots. Gate FAIL_Range 3.623% on SGD 3,067,739. T12 0.73 vs 0.69 = +5.8%. val_adj -3 (FAIR_VALUE MoS -0.5424 x5); Quality_Pass False is a hard block on new entries regardless. Note its 08-19 Automation Taipei launch fully round-tripped -12.9%. |
| BN2 | Valuetronics Holdings | 1.01 | WATCH | — | — | n/a | n/a | n/a | Valuetronics closed unchanged at 1.01 on the quietest tape in the book and produced no signal. It is one of only two names we could actually have traded tonight, and there was nothing to trade - the model's internal read is genuinely conflicted here, not blocked by liquidity. Our twelve-week view is 5.9% below the current price. Nothing to do. OPERATIONAL: NO TRADE cv0. Gate PASS on BOTH legs (0.990% range - the tightest in the book - on SGD 216,745), so its 'Consensus conflict + low conviction' label is GENUINE. This row is the CONTROL for systemic finding #1: BN2 is Pillar_Consensus Indeterminate AND gate-passing, which is why its consensus label is correct while AWX's and 575's identical labels are misattributions. T12 0.95 vs 1.01 = -5.9%. val_adj -1. MSCAP-34 volume-artifact concern retained (a byte-identical OHLC bar previously moved conviction with volume the only differing input). |
| M14 | InnoTek | 0.44 | WATCH | — | — | n/a | n/a | n/a | InnoTek fell a further 2.3% to 0.435 and remains one of the weakest names in the book, with a twelve-week view 42.5% below the current price and a first-half result that saw profit collapse 64%. We declined it at 0.51 on 12 August and that refusal is now worth 14.7%. No signal tonight. OPERATIONAL: NO TRADE cv0 across all four snapshots. Gate FAIL_Range 4.598%. CORRECTION TO THE 08-26 ENTRY: M14's turnover leg has RECOVERED - SGD 201,492 tonight vs SGD 43,788 on 08-26, so it now clears the SGD 100,000 floor and fails on range alone. S71 is once again the only name failing on turnover. T12 0.25 vs 0.435 = -42.5%. val_adj -4, the book's only valuation deterioration at the last regen (-0.330 -> -0.721); DCF_per_share unavailable. Prior finding retained: 1H FY2026 profit -64% to SG. |
| 575 | ASTI Holdings | 0.12 | WATCH | — | — | n/a | n/a | n/a | ASTI rose 2.6% to 0.117 on heavy share volume but the sums involved are small and the twelve-week view remains 14.5% below the current price. No signal - its daily range exceeded our liquidity ceiling again. Nothing here. OPERATIONAL: NO TRADE cv0. Gate FAIL_Range 3.419% on SGD 1,576,551. ENGINE DEFECT ROW: NoTrade_Reason misreports 'Consensus conflict + low conviction' over a provable range failure (Pillar_Consensus Indeterminate - see systemic finding #1). T12 0.10 vs 0.117 = -14.5%, continuing to deteriorate (-13.8% -> -21.1% -> -14.5%). val_adj -10 CLAMPED at MoS -1.9014 - MSCAP-2 saturation, identical to AWX's -10 at MoS -9.7737 despite a 5.1x difference in the underlying reading. |
| S71 | Sunright | 0.44 | WATCH | — | — | n/a | n/a | n/a | Sunright is the cheapest name in the book on paper - roughly a 53% discount to our estimate of fair value, which earns it the maximum valuation credit our framework allows - and it is completely untradeable. It changed hands just once tonight, SGD 13,006 in total, about a tenth of the minimum daily turnover we require. A discount you cannot buy or sell out of is not a discount. The twelve-week view is also 42.5% below the current price, so the cheapness is not corroborated by anything forward-looking. OPERATIONAL: NO TRADE cv0. Gate FAIL_Turn - SGD 13,006 vs the SGD 100,000 floor (recovered from SGD 2,838 on 08-26 but still ~13%). Range 0.000% (High = Low = Close = 0.435 - a single print), so it trivially clears the range leg and fails turnover alone; with M14's turnover recovering, S71 is again the ONLY name in this book failing that leg. val_adj +10 clamped - MSCAP-2's positive-end saturation: the maximum valuation credit the framework can award is being spent on a name with no tradeable market. |
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.