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 |
|---|---|---|---|---|---|---|---|---|---|
| BN2 | Valuetronics Holdings | 1.04 | WATCH | 1.04 | 1.04 | n/a | n/a | 0.99 (-4.8%) | Valuetronics is now the only name in the book still producing a trading signal at all, and it is a bearish one pointing about 5% lower over twelve weeks. Its indicators had briefly come into agreement yesterday - the first genuine agreement this universe had produced in weeks - and that agreement has already broken down, on volume two-thirds lighter, with the shares unchanged. We are long-only and hold none of it, so there is nothing to do. OPERATIONAL: conviction collapsed (-20 pts) in one session; consensus Mostly_Aligned -> Consolidation_Bullish; volume (-68%) on an unchanged 1.040 close. RSI 56.45, QPass False, T12 0.99 = -4.8%. Long-only + = no action either way (SGM-4b). Its clean gate PASS at a 1.92% range is tonight's only fresh cross-sectional lower bound (S71's 3.06% fail is a repeated bar - see the S71 card). |
| 5DD | Micro-Mechanics (Holdings) | 2.78 | WATCH | — | — | n/a | n/a | 2.55 | Micro-Mechanics was the only name in this book carrying a non-bearish signal for the last four sessions, and tonight our own liquidity filter removed it - not because anything changed about how much of it trades, but because the day's price swing was wider. It leaves the buy pool completely empty. Nothing about the company changed. OPERATIONAL: WATCH cv58 -> NO TRADE cv0, NoTrade_Reason 'Liquidity gate failed' on a 3.96% range; closed 2.780 (-1.42%). THIS NAME IS NOW SYSTEMIC FINDING 2'S CONTROLLED EXPERIMENT: the gate PASSED it on Sof turnover (08-13, 2.83% range) and FAILED it tonight on S - more turnover, opposite verdict - with realised range the only variable crossing the boundary (2.48% -> 3.96%). Turnover was flat within noise across all six sessions (S-0.39m). This removes the last explanation that the gate proxies liquidity via range. Shadow-test a turnover/spread specification; do NOT hand-tune the threshold. RE-ARM unchanged: settled PullbackBuy_VWAP/MomentumContinuation BUY >= 65 that clears the gate, with a stop wider than ~1 ATR (its 08-17 stop 2.80 vs entry 2.82 was 0.26 ATR). |
| AWX | AEM Holdings | 9.96 | WATCH | — | — | n/a | n/a | 12.22 | AEM fell 5% to 9.96, its second sharp drop in four sessions, and is now 9% below where it closed last Thursday. The business news is the best in the book and got better: analysts raised their price target 77% to S, see 2026 revenue of S against management's own S-510m guidance, and the shares are up 90% this year on record cash generation and a reinstated dividend. We still cannot buy it. It fails our quality screen, prices in more growth than any other name we track, trades 9% above its volume-weighted average price, and produces no signal. A good story at a stretched price is still a stretched price - which is precisely the case the floor exists for. OPERATIONAL: clean gate fail at 7.33% range on SGD 46.91m turnover (unmasked tonight, unlike 08-17's consensus-conflict masking - systemic finding 4). Our T12 12.22 sits 7.8% BELOW the street's 13.25; our projection is the conservative one. VALUATION LABEL RE-RATED ON TONIGHT'S REGEN: FAIR_VALUE -> GROWTH_PRICED_IN with MoS essentially unchanged (-10.780 -> -10.844) - the inconsistent pairing cited for weeks as MSCAP-2's headline exhibit was partly a staleness artifact (systemic finding 1). Ex-dividend SGD 0.024 on 2026-08-19: 0.24% of close, immaterial, not a catalyst. Refusal ledger: the 08-12 floor refusal FLIPPED from +4.80% foregone to -0.40% avoided; the 08-13 chase refusal is -9.45% avoided. RE-ARM (all three): Quality_Pass repair on the next fundamentals roll + a settled FAVOURED-cell BUY (PullbackBuy_VWAP / MomentumContinuation, NOT FailedBreakdown) >= 65 + a stop wider than ~1 ATR. · news |
| I07 | ISDN Holdings | 0.77 | WATCH | — | — | n/a | n/a | 0.87 | ISDN was the only name in the book to rise tonight, up 7.7% to 0.770 on nearly twice the previous day's volume - the heaviest trading in the book by share count. We can find no announcement that explains it, and what coverage exists is stale and negative in tone. An unexplained spike on heavy volume in a thin small-cap is not an entry; it is a reason to wait for the reason. It has also just re-rated from fairly valued to pricing in growth, having run 18% in five sessions. OPERATIONAL: gate-blocked on a 9.09% range while transacting SGD 22.08m - tonight's starkest cross-sectional exhibit for systemic finding 2 (55x BN2's SGD 0.40m, resolved the wrong way). RSI 49.7 -> 67.2 over four sessions; VWAP distance +11.35%, extended. VALUATION RE-RATED ON TONIGHT'S REGEN: FAIR_VALUE MoS -0.447 -> GROWTH_PRICED_IN MoS -0.742, correctly tracking the +18.5% run since 08-11 (systemic finding 1, the surface behaving as designed). Quality_Pass False. T12 0.87 = +13.0%. NEWS: none found for August 2026. · news |
| PCT | PC Partner Group | 2.92 | WATCH | — | — | n/a | n/a | 4.09 | PC Partner fell another 4.6% to 2.92, extending Monday's reversal off its 52-week high into a two-day 10% retreat, with no announcement to explain either session. It remains the book's best forward case - a 12-week view 40% above today's price, passing our quality screen, priced almost exactly at fair value - and it remains excluded by our own liquidity filter for the eighth time in nine sessions. The exclusion has now paid two sessions running; that is worth stating plainly and it still does not repair the specification argument, because a filter can be wrong in construction and right on a given day. OPERATIONAL: gate-blocked on a 6.51% range. Regenerated valuation FAIR_VALUE MoS -0.036 -> -0.020, QPass True, still the cleanest valuation in the book. Regime 'Trending', Trend 77. RE-ARM unchanged and two-legged (AND): a settled PullbackBuy_VWAP/MomentumContinuation BUY that clears the gate AND prints raw cv >= 65. Refusal ledger: the 08-06 floor refusal has decayed from +12.9% foregone (08-16) to +7.00% (08-17) to +2.10% tonight on the L8 execution basis (08-07 open 2.86). NEWS: none found; search-index snippets again contradicted our settled close (2.70 for 02-Jul, 1.35 for 01-Apr) - fourth consecutive night, L4 stands, disregarded. · news |
| 8AZ | Aztech Global | 0.60 | WATCH | — | — | n/a | n/a | 0.34 | Aztech is the one genuinely new tell tonight: after four straight sessions where its indicators agreed the trend was down, they flipped to a bullish reading - the shape of a stock that has stopped falling. It screens as the cheapest name we track on a quality-adjusted basis. But its own 12-week outlook is 43% below today's price and our liquidity filter removed its signal entirely, so there is nothing to act on. A watch item, not a position. OPERATIONAL: consensus Consolidation_Bearish -> Consolidation_Bullish, breaking a four-session AVOID streak (cv67/68/62/66), but the row prints NO TRADE on a 3.33% range. Closed 0.600 (-1.64%); RSI 20.6, still pinned oversold; VWAP distance -9.96%. UNDERVALUED MoS +0.220 with Quality_Pass False (val_adj +2, halved). T12 0.34 = -43.3% would disqualify it even with a clean signal. RE-ARM: FY2026 revenue stabilisation repairing Quality_Pass FIRST, then a settled favoured-cell BUY >= 65 - and the forward view has to turn. |
| M14 | InnoTek | 0.44 | WATCH | — | — | n/a | n/a | 0.25 | InnoTek fell another 5.4%, its seventh decline in eight sessions, which puts the watch-level signal we declined at 0.51 a week ago down nearly 14%. It stays the counter-example to the AEM miss and deserves equal weight: the same discipline that kept us out of a name that ran has kept us out of one that has not stopped breaking. The 12-week outlook is 43% below the current price. OPERATIONAL: gate-blocked on a 5.68% range; closed 0.440, -5.38%. RSI 27.8, VWAP distance -13.80%, the weakest in the book. The declined 08-12 card is held by the counterfactual grader and is this channel's clearest favourable datapoint: -13.73% avoided on the L8 execution basis (08-13 open 0.510), improved from -8.82% last night. Its valuation improved mechanically as price fell (MoS -1.358 -> -0.330); that is the surface tracking price, not a thesis. |
| 558 | UMS Integration | 2.82 | WATCH | — | — | n/a | n/a | 3.34 | UMS gave back 1.7% after leading the book higher yesterday, and we still could not buy it in either direction. Two things independently block it regardless of the chart: it fails our quality screen, and it prices in more growth than we are willing to pay for - and tonight's fresh valuation made the second one worse. A clean buy signal alone would not be enough; this needs the fundamentals or the valuation to re-rate first. OPERATIONAL: double-blocked (SGM-4a) - GROWTH_PRICED_IN MoS deteriorated -4.016 -> -4.559 on tonight's regen, holding val_adj at the -10 floor, AND Quality_Pass False, so it cannot reach the 65 floor even with a clean signal. Also gate-suppressed on a 4.26% range while transacting SGD 34.17m - the book's second-largest turnover, blocked. T12 3.34 = +18.4%, and it ROSE (+1.2%) on a -1.74% price bar, a fresh instance of systemic finding 5 running in the confirming direction. · news |
| E28 | Frencken Group | 2.67 | WATCH | — | — | n/a | n/a | 2.33 | Frencken fell 3.3% to 2.67, giving back most of yesterday's gain, and our indicators cannot agree on what it is doing - the reading is formally indeterminate. Our own 12-week projection sits 13% below today's price. The company itself is fine; the stock gives us nothing to act on. OPERATIONAL: consensus Indeterminate, so NoTrade_Reason prints 'Consensus conflict + low conviction' and the liquidity verdict is MASKED, not passed (systemic finding 4 - E28 has now alternated masked/gated across 08-12, 08-13, 08-14, 08-17, 08-18). Regenerated FAIR_VALUE MoS -1.159 -> -1.341, val_adj -7; MSCAP-2's sizing claim applies here - lean on the sign, not the x5 magnitude - and that claim is UNAFFECTED by tonight's label re-ratings elsewhere (systemic finding 1). QPass True. The 5.4x external DCF divergence logged 08-17 stands unresolved and remains advisory only. |
| S71 | Sunright | 0.49 | WATCH | — | — | n/a | n/a | 0.28 | Sunright is still the deepest discount on our valuation screen and still impossible to actually buy - and tonight it did not trade at all. Every price and volume figure in its row is identical to yesterday's, so there is no new information in it. A discount you cannot transact in is not an opportunity, and a repeated bar is not a session. OPERATIONAL: DATA INTEGRITY - the 08-18 row is byte-identical to 08-17 across Close/Open/High/Low/Volume(137,400)/RSI(42.0814489518346)/ATR/EMA5. Either the name did not trade or the feed carried the bar forward; the scan cannot distinguish these and emitted a live NoTrade_Reason regardless (systemic finding 3). CONSEQUENCE FOR TONIGHT'S ANALYSIS: S71's 'clean gate FAIL at 3.06%' is a REPEAT, not independent evidence, so it does not serve as tonight's cross-sectional upper bound - 5DD's fresh 3.96% fail does. Recorded against our own bracket claim. T12 0.28 = -42.9%; val_adj +9 is the largest in the book and irrelevant absent a chart and size (SGM-5). |
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.