Recommendations
0 to acquire · 2 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 |
|---|---|---|---|---|---|---|---|---|---|
| J36 | Jardine Matheson | 58.08 | SELL 50% | 58.08 | 56.74 | n/a | n/a | 45.42 (-21.8%) | Jardine Matheson has broken down: the scan flipped it to STRONG SELL - REDUCE at conviction 71 and ranked it the book's #1 sell, with the price now below its 20-, 50-, 100- and 200-day averages and a three-month projection of S, roughly 22% below today's Sclose. The shares look cheap on intrinsic value (margin of safety +0.42, discounted-cash-flow anchor S), and news is constructive in the abstract - Hongkong Land is in talks for over US$1bn of prime Tokyo property as part of Jardine's pivot to a returns-focused investment holding company - but the group has yet to find the right asset or partner, and neither cheapness nor strategy reverses a confirmed trend break. We take half off and keep half, rather than following the model's full 80% exit, precisely because the value case is real. OPS: signal-originated trend exit per CLAUDE.md s9 (SELL-REDUCE + below EMA200 + negative 12wk) - G1/G2/G3 do not apply, nothing here is uptrending. The row's PROFIT_ARMED flag is the inverted-span artifact (T12_Progress 127.87 on a +0.02% position) and is deliberately NOT the basis. Valuation amplify-only: UNDERVALUED + Quality_Pass True argues one tier DOWN, 80.07% -> 50% = (100-lot). Scorecard cell FailedBreakout_Down_20D|sell is 4-for-4 at 1w and 4w but n=4, report-only, bias 0. NOT a full exit - L2's second leg unmet, close is 2.36% above the 56.74 stop (low 57.21). Stop held at the 56.74 watermark; the row prints 55.25 and a watermark is never lowered. Retain 2,400. Third re-issue of an unfilled card, upgraded from 30% and repriced 57.63 -> 58.08 (+0.78%, inside the gap gate). 1wk/4wk n/a - Target_Price blank on this row; model near-term band 56.79-59.24. No DCF annotation (45.42 far below the 132.13 anchor). No J36 catalyst anywhere in the 10-day horizon window. L3/L4/L8. · news |
| Z74 | Singtel | 4.52 | SELL 50% | 4.52 | 4.50 | n/a | n/a | 4.59 (+1.5%) | Singtel's signal collapsed in a single session - conviction fell from 73 to zero, the momentum reading flipped bearish and the indicators now openly disagree, on the heaviest volume in the book (13.9m shares). The three-month projection of Sis barely 1.5% above today's Sclose, so the forward reward has gone flat while the position sits exactly at cost. The stock still screens well on the long view - up about 54% over a year, roughly double over three, and at 12.1x against an Asian telecom average of 16.5x - but that is a valuation argument, and valuation never cancels a deteriorating signal. We bank half and keep half on a tight stop rather than defend a position the model has stopped underwriting. OPS: ADR-0012 profit-take, ARMED (PROFIT_ARMED, T12_Progress 98.47) + 2 TICKS - (b) conviction, far past the 20-pt bar, and (c) dead 12wk, 4.59 < 1.02 x 4.52 = 4.6104, a real 0.44% margin this time (vs 0.026% on 09-09). Tick (a) no longer fires (RSI_DIV dropped off); tick (d) needs a 2nd consecutive NO TRADE. 2 ticks -> profit_take_2 = 50% =. G1 unavailable (NoTrade cv0, not BreakoutUp/MomentumContinuation at cv>=80). SGO-4 CHECKED AND SET ASIDE: NoTrade_Reason is 'Consensus conflict + low conviction', NOT 'Liquidity gate failed' - this is genuine deterioration, so tick (b) counts. Valuation inert (MoS -0.61, neither <=-3 nor UNDERVALUED, no tier change). Scorecard NoTrade (PROFIT_ARMED)|sell is ACTIONABLE at -2 and argues AGAINST - does not cross a tier boundary, fire rule is mandatory, disagreement recorded. CAVEAT RECORDED: position is FLAT (cost 4.52, close 4.52, Unreal_PnL -0.0%, Gain_ATR -0.0) - calling this a 'profit-take' is a category error, the same arming artifact as J36; the card stands on the independent deterioration read. Third re-issue of an unfilled card, repriced 4.56 -> 4.52 (non-fill cost ~S). 1wk/4wk n/a - Target_Price blank; model band 4.4296-4.6104. 12wk carries the exceeds-DCF-anchor flag (4.59 vs 1.95; Z74 is not financial so SGO-3 suppression does not apply). FTSE ST review 09-21 is 7 sessions out and weight-only. L3/L4/L8. · news |
| D05 | DBS Group Holdings | 76.79 | HOLD | 76.92 | 77.93⚠ | n/a | 79.63 (+3.5%) | 88.75 (+15.4%) | DBS closed below our recorded stop for a second straight session, but we are holding. The bank just posted a record quarter - net profit S08bn, up 9%, on record total income of S09bn crossing S$6bn for the first time, with a combined Sper share dividend - and the model's three-month view is still +15.6%. The uptrend structure is intact, with the price above every major moving average, and US inflation data lands tomorrow with the Fed decision on 16 September; both hit the bank directly, and the discipline is to hold through a scheduled event rather than sell into it. We exit only if the price actually breaks trend, not because a two-day fade clipped a tight stop. OPS: L2 requires a decisive close below stop AND a negative 12wk - the 12wk is +15.57% (88.75), second leg fails, no exit. Pre-authorized in the 09-08 card. Close 76.79 is 1.46% below the 77.93 stop (widened from 0.55%) but only 0.20% above EMA20 76.6361 - HAIRLINE. Trend Bullish, Regime Trending, EMA50 73.54, EMA200 62.85, row stop 75.90. Catalyst-aware hold: CPI 09-11 (1 session), FOMC 09-16 (4 sessions). WATERMARK 77.93 DELIBERATELY NOT LOWERED to 75.90 - the monitor will keep flagging it and that is the honest state. OPERATIVE EXIT TRIGGER: a close below EMA20 76.6361, or below the model stop 75.90. Adj conviction 59 (60 - 3 val + 2 bias) is 6 under the floor, so no add either; SGO-3 makes the Quality_Pass=False and the broken bank DCF non-binding. Position S = 59.8% of NAV, the largest single risk in the book. The Paredes retirement is a 2027 succession item, not a 2026 event. · news |
| O39 | OCBC Bank | 31.39 | HOLD | 31.39 | 31.11 | n/a | 32.39 (+3.2%) | 38.74 (+23.4%) | OCBC is sitting on a solid gain with a three-month view still 23% higher, and nothing in today's data says to take it. Conviction is unchanged, the trend reading is intact and no new warning appeared, so we keep the position and keep the protective stop where it is. The position is small - about 0.5% of the book - so this is a discipline exercise rather than a meaningful exposure decision. OPS: ADR-0012 ARMED (PROFIT_ARMED, T12_Progress 81.03, Gain_ATR 5.51, +9.64%) with ZERO TICKS for a FOURTH consecutive session - conviction flat, no 2nd Exit_Warning code, 12wk alive at 38.74 (+23.4%), Action HOLD, no NO TRADE. Mandatory trail-raise path applies but NO RAISE IS AVAILABLE: the tightened chandelier (31.3899 - 2.0 x ATR 0.5013 = 30.39) and the row Stop_Price 30.39 are both BELOW the existing 31.11 watermark, and a watermark is never lowered. STOP HELD AT 31.11. Today's low 31.10 pierced it by ONE CENT intraday; settled close 31.39 governs per L4, no stop-out - but flagged, because the identical one-cent pattern on Z74 last session preceded tonight's collapse. CPI 09-11 and FOMC 09-16 both hit this name; catalyst-aware hold reinforces. SGO-3: Quality_Pass=False non-binding (IsFinancial). |
| — | Venture Corporation | — | NO ACTION | — | — | n/a | n/a | n/a | Venture is the book's top-ranked buy candidate tonight and we are still not buying it. The entry setup is clean - a pullback to the volume-weighted average with the price not extended - but the model's own three-month projection of Ssits below today's Sclose, so the engine is simultaneously flagging a buy and forecasting a lower price a quarter out. Conviction lands five points under our minimum bar. The recent softness is benign, explained by a Sdividend going ex on 31 August, but that is a reason not to read the dip as damage, not a reason to commit capital. We wait for conviction to clear the bar with a forward view that actually points up. OPS: floor is adj conviction >=65; 60 raw - 2 val_adj + 2 scorecard bias = 60, FIVE SHORT. Buy_Rank #1, Entry_Distance_ATR -0.07 so the no-chase rule would NOT have blocked it - the floor and the dead 12wk did. NOT QUEUED: SGO-5 forbids a self-invented deeper limit and the only legal price (the live model entry 16.72) fails the floor; the 09-08 withdrawal of the old 16.91 order stays withdrawn. RE-ARM TRIGGER: conviction >=65 WITH Target_12Week above the close. Quality_Pass True so the gate is not the blocker. 1wk 16.83 shown at f=0.124 (scorecard-measured, 19 graded buys) for reference only - no position taken. Ledgered direction=none/optional=true so the counterfactual grader prices the pass. L1 (PullbackBuy is the favoured family) noted and still insufficient. NODX 09-17 touches this name. |
Outcome & track record
Accuracy at the time of this record.
49 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 →
0.124 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.