SG Off-the-Radar — End-of-Day Verdict
1 ACQUIRE queued (S58 GTC buy-stop @S, exp 08-11); 2 DISPOSE (J36 3,400 double-fill reconcile; G07 900 = 30% profit-take, tick d); executor double-filled J36 - vs 3,400 authorised
Recommendations
1 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 |
|---|---|---|---|---|---|---|---|---|---|
| S58 | SATS | 4.83 | BUY — QUEUED @4.85 | 4.85 | 4.50 | 5.06 (+4.3%) | 5.55 (+14.4%) | 6.19 (+27.6%) | SATS is the strongest setup this book has seen: the aviation-services group has broken into a clean uptrend, with conviction jumping from in a single session on the heaviest volume in weeks, and the shares now sit above their 20-, 50- and 200-day averages with the trend reading at 98 out of 100. Unusually, every horizon points the same way - roughly 4% higher in a week, 14% in a month and 28% over three months - and the discounted-cash-flow estimate of Ssits far above today's S, so the projection assumes nothing heroic. The order is placed as a buy-stop just above the market, so it only fills if the move continues; risk is capped at about 1% of the book by a stop 7% below. Bought at three-quarters of standard size because the conservative earnings-power valuation disagrees with the cash-flow one. || OPERATIONAL: SGO-5 queue style - GTC buy-stop AT the anchor's printed model Entry_Price 4.85, on-strength style because Entry_Distance_ATR +0.17 >= 0; no carried-forward or self-invented level. Reprice-suppression not violated (that clause guards marketable limits; RSI 67.3 < 75 anyway). L1: MomentumContinuation_Up, not a breakout/FailedBreakdown chase. val_adj -10 from FV_method=EPV (1.399, no-growth floor) on a +9.0% grower - the systemic #2 artifact; MoS_DCF is +0.42 and Val_Signal reads FAIR_VALUE. Applied as written, no exemption. Quality_Pass True (F_Score 8, Quality_Rank 3/17) so the hard gate passes. Size 96.51% x Sx 20% x 0.75 = S / 4.85 = = S = 14.5% of book; 0.68% of daily volume; risk S = 1.04% of book. scorecard 0 - MomentumContinuation_Up|buy is n=3 hit_1w 0.333, report-only, sample too small to bias (and the weakest evidence behind this card). f=0.30 for the 1wk per L5/SGO-2. Re-armed the 08-03 near-miss trigger ('raw cv >=75 while the EPV anchor stands'). Withdraw if adj conviction < 65, expires_on passes, RSI > 75, or the printed entry falls to/below the close. · news |
| J36 | Jardine Matheson | 63.44 | SELL 50% | 63.44 | 61.76 | n/a | n/a | 59.40 (-6.4%) | Half the Jardine Matheson holding is being sold because the book never authorised it: the order placed on 3 August was repriced on 4 August, and the automated executor filled both versions this morning as if they were separate trades - so were bought where 3,400 were decided. Selling the excess simply restores the intended position. The timing is fortunate, because the shares turned the same day: the buy signal that justified the trade flipped to a failed-breakout warning, the price closed below its 20-, 50- and 200-day averages, and the three-month view was cut to 6% below today's level. The remaining half stays, because the value case is the best in the book - both intrinsic estimates put fair value far above the Sprice - and the shares are still above their stop. || OPERATIONAL: ORIGINATION = execution reconciliation (unauthorised double-fill), NOT a signal or valuation or portfolio-shape sell. transactions.jsonl sgotr-20260805-0001 ('fill of 2026-08-03 J36 verdict') and -0002 ('fill of 2026-08-04 J36 verdict'), 3,400 @ 63.55 each = Sdeployed vs Sauthorised; the 08-04 memory says verbatim the card 'replaces (reprices) the 08-03 GTC'. Signal independently corroborates: BUY cv61 PullbackBuy_VWAP -> AVOID cv46 FailedBreakout_Down_20D in one session, Target_Price blank, T12 62.19 -> 59.40, pillars 30/24/29, consensus Mixed, close 63.44 below EMA20 64.40 / EMA50 64.80 / EMA200 65.21; under the ADR-0012 queue-expiry rule this order would have been auto-withdrawn today (adj 53 < 65). Held=N on today's row is the SGO-4 one-session post-fill lag (scan 18:00:27, executor 18:00:29), not an unheld view. NOT a full exit: close 63.44 > the 61.76 stop so L2 is not met; note-B's AVOID case requires degradation FROM a sell label, and this degraded from BUY; and section 9's amplify says an UNDERVALUED name's trim argues DOWN a tier (MoS_FV +0.363, MoS_DCF +0.507, Quality_Pass True, Quality_Rank 7) - which is why the sale stops exactly at the 3,400 excess. Proceeds Srealised -S (-0.17%). 1wk/4wk n/a - AVOID prints no Target_Price, never synthesised. CURRENCY (systemic #3): Price_Currency=USD in fairvalue while the executor deducted 63.55 from SGD cash; if USD is right the outlay and market value are ~28% understated. Arithmetic here uses the book's own SGD units for consistency; the error is proportional and does not change the decision. · news |
| G07 | Great Eastern | 21.98 | SELL 30% | 21.98 | 20.64 | n/a | n/a | 31.32 (+42.5%) | A second slice of the Great Eastern winner is banked. The position is up 20% and more than four times its normal volatility above cost, and the signal engine has now failed to produce a tradeable read on it for two consecutive sessions - the pillars are in genuine conflict, with momentum collapsed and price stalling after the July results run. That combination is the model's own definition of a position whose move has stopped working, so 30% comes off and the trailing stop on the rest is raised to S. The remaining 70% stays invested: the three-month view is still 42% higher and the insurer pays a Sdividend on 19 August. || OPERATIONAL: ADR-0012 profit-take, armed + fired, 1 tick -> profit_take_1 = 30%. Armed = PROFIT_ARMED, Gain_ATR 4.46 (>=3.0), +20.31%, T12_Progress 70.18. FIRED on tick (d): 08-04 and 08-05 both NoTrade_Reason 'Consensus conflict + low conviction' = two consecutive GENUINE NO TRADEs; the 07-31/08-03 'Liquidity gate failed' prints stay suppressed per SGO-4. Tick (a) NO - RSI_DIV cleared 08-04 and has not returned. Tick (b) uncomputable - Conviction_Score 0 both sessions (systemic #5); momentum pillar flat at 14, consensus Conflicted, no substitution invented. Tick (c) NO - T12 31.32 > 1.02x close 22.42, which is why 70% stays. The 08-04 re-arm note anticipated profit_take_2 (50%) but that presumed two ticks; only (d) fires, so the rule as written governs at 30%. Valuation adds nothing - MoS -0.671 is not <=-3 and SGO-3 makes the insurer ResInc band non-informative; no tier bump. G1 does not protect (needs fresh BreakoutUp_*/MomentumContinuation_* at cv>=80; this is NoTrade cv0, momentum 14) and ADR-0012 fires ticks (b)/(c)/(d) regardless of G1. G2/G3 not engaged. Catalyst-aware hold cleared - dividend ex-date 08-19 is 9 sessions out (08-10 National Day intervening). Trail 20.48 -> 20.64 = 2.0xATR chandelier on the unchanged peak close 22.30 (ATR tightened 0.9115 -> 0.8312); a trail never falls. Action price = anchor close 21.98, derived by hand per SGO-4 (row blanked). 1wk/4wk n/a - no Target_Price to interpolate f against. Proceeds Srealised +S (+20.3%); the forgo Sof dividend. L3 - written down on 08-04 as the exact re-arm trigger; execute it. · news |
| D05 | DBS Group Holdings | 73.55 | HOLD | 73.12 | 70.95 | n/a | 76.58 (+4.7%) | 90.64 (+24.0%) | DBS is held unchanged into second-quarter results tomorrow morning. The position is 1.6% ahead, the three-month view is still 23% higher, and there is no warning on the row - the signal has gone quiet rather than turned. Bank results are to be held through rather than traded around, and the mid-August dividend takes about 1% off the price mechanically, which is cash to shareholders and not a reason to sell. || OPERATIONAL: Not armed - Gain_ATR 1.08 < 3.0, T12_Progress 81.15 < 90, no Exit_Warning. Conviction = -12, short of the 20-point collapse tick, and no label change. Model Stop_Price prints 71.82 and is NOT adopted; trail held at 70.95 and NOT lowered (a trail never falls). Known exposure, stated: the hold-through band floor is 73.55 x 0.96 = 70.61, so 70.95 now sits 0.5% INSIDE the in-band earnings-gap floor - an at-the-limit -4% print clips it, anything shallower does not. A gap-through on a print explicitly stanced hold-through is a next-session reassessment, not a forced exit; read tomorrow's monitor breach flag that way. Ex-date 13-17 Aug is a corporate action that must not originate or amplify a trim. ADR-0020: Peak Stop = live stop, 0% erosion. |
| O39 | OCBC Bank | 28.63 | HOLD | 28.70 | 27.73 | n/a | 30.23 (+5.3%) | 36.45 (+27.0%) | OCBC is held unchanged into second-quarter results on 7 August. The shares closed exactly at what the book paid, so the position is flat, but the setup has not broken - this is the fifth straight session printing the same pullback-buy pattern and the three-month view is 27% higher. Conviction at 61 is below the bar for adding more, and results are to be held through rather than pre-positioned around, so nothing changes. || OPERATIONAL: Not armed - Gain_ATR -0.00, Unreal_PnL 0.00%, no Exit_Warning, no sell case. cv61 < 65 so no add; Entry_Distance_ATR flipped to +0.15 (entry 28.70 now ABOVE the close), so a re-arm would need buy-stop style, not the withdrawn limit. Model Stop_Price prints 28.13 and is NOT adopted; trail held at 27.73. Known exposure, stated: hold-through band floor 28.63 x 0.96 = 27.48, so 27.73 sits 0.9% inside it - an at-the-limit -4% gap clips it. Same disposition as D05: hold the stop, treat an in-band gap-through as a next-session reassessment, not a forced exit. The 07-31 queue stays withdrawn; re-arm requires BUY - ADD at cv >= 65. ADR-0020: Peak Stop = live stop, 0% erosion. |
| 9CI | CapitaLand Investment | 2.76 | NO ACTION | 2.73 | 2.56 | 2.84 (+4.0%) | 3.09 (+13.2%) | 2.84 (+4.0%) | CapitaLand Investment is still breaking out and is still not being bought. The property group fails the quality screen outright - return on of 1%, revenue shrinking 24%, and the lowest quality rank but three in the book - and the shares trade at nearly three times the discounted-cash-flow estimate of S. On top of that the price is already well through the model's entry with the momentum gauge at 79, so buying here is chasing. Yesterday's stronger signal has faded from in a session, and the company reports its own half-year results on 13 August. Logged, not taken, so the cost of standing aside can be measured. || OPERATIONAL: adj = 62 - 9 = 53, eleven under the 65 floor. Quality_Pass False (F_Score 5, ROE 1.15%, Rev_Growth -24.2%, Quality_Rank 14/17) = HARD BLOCK on new entries; SGO-3's financials carve-out does not apply (IsFinancial=False, Real Estate, FV method DCF) - a genuine quality failure, not the structural artifact. Chase: RSI 79.40 > 75 (SGO-5 suppression threshold) and Entry_Distance_ATR -0.53, price already through the 2.73 buy-stop. L1 additionally discounts breakout chases. Ledgered direction=none/optional=true for the counterfactual grader - note the name closed UP again (2.73 -> 2.76), so yesterday's block is only partially vindicated so far. Size_mod 0.5 recorded as what would have applied (MoS_FV -1.81) had the gate passed. · news |
Outcome & track record
Accuracy at the time of this record.
8 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 →
0.141 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.