Every forward-looking call in our reports is logged before the data prints, then graded after — hits and misses both. Every wire headline carries a verification label, and label changes are recorded, never erased. This page is the whole record, unedited. No other desk in this trade shows you this.
Grading is strict, on the stated basis (closing prices unless the call said otherwise). A miss stays a miss — we explain it, we never delete it.
| Said | Our words, exactly | What happened | Grade |
|---|---|---|---|
| 2026-07-10 | “MPOB stocks ~2.55 Mt (poll 2.50)” official June print |
Actual 2,544,066 t — off by 6,000 t. Published before the print. | HIT |
| 2026-07-10 | “Output ~1.68 Mt (hotter than poll 1.65)” official June print |
Actual 1,638,777 t (+8.08%) — cooler than poll. Owned: private trackers ran hot. | MISS |
| 2026-07-10 | “Exports ~1.28 Mt (below poll 1.30)” official June print |
Actual 1,204,013 t (+6.19%) — right direction vs poll's +17.1%, magnitude still too generous. | PARTIAL |
| 2026-07-09 | “The record stockpile is mostly already in the price — history says the report day itself is rarely the shock.” 10-Jul MPOB same-day reaction, Sep-26 |
RE-GRADED at close (we marked HIT prematurely at midday — owned): Sep-26 settled 4,511 (-1.81%), the FIRST record-stocks day in our 5-episode study to fall >1%. The base rate broke because this record was built by DEMAND failure (exports +6.2% vs +17.1% poll) — none of the 5 precedents were demand-miss records. Lesson: condition the record-day base rate on the print's COMPOSITION. | MISS |
| 2026-07-09 | “Scenario matrix: in-line print (2.40-2.55 Mt) = flat to -1%, the 4,552-4,642 box holds; >2.60 Mt = -1% to -2.5%; <2.40 Mt = +1% to +2%.” 10-Jul MPOB print vs same-day Sep-26 move (crude holding $74-80 assumed) |
Box 4,529-4,628 did NOT hold: settle 4,511 (-1.81%) vs our in-line row 'flat to -1%'. Even our beat row (-0.5 to -1.5%) under-sized it. Composition (demand-failure) mattered more than magnitude vs poll. | MISS |
| 2026-07-09 | “BNM: 24 of 25 economists say hold at 2.75% — the rate is a non-event; the statement language matters.” 9-Jul 3pm BNM decision |
BNM held at 2.75% (~3pm, as we said 24/25 expected); statement called Middle-East oil pressure on inflation 'contained' — the language mattered, as flagged. | HIT |
| 2026-07-09 | “Judge tomorrow's 1-10 July export prints on absolute tonnage vs ~377k t, not the flattering percentage.” 10-Jul AmSpec/ITS prints — the % is mechanically inflated by June's weak 1-10 base |
AmSpec July 1-10: 396,261 t (+5.1%) — the absolute-vs-377k yardstick was the right lens; print above the bar = genuine mild demand improvement, correctly not flattered. | HIT |
| 2026-07-08 | “Rangebound and slightly heavy into the two big events.” 8 Jul session bias, Sep-26 |
Wrong. The session closed +62 (+1.36%), a two-week-high breakout. We treated the Gulf conflict as background; when the whole energy complex jumps 3-5%, palm does not stay rangebound. Lesson encoded as the crude pass-through arithmetic (loaned-points) so it is sized, not hand-waved, next time. | MISS |
| 2026-07-08 | “Viral claims checked: three ships were hit, not four; the strait is open, not closed.” fact verification vs circulating rumours |
Confirmed by every major wire since (UKMTO/CENTCOM/Reuters). The '85 US sites' figure we flagged as Iran's own claim remains uncorroborated. | HIT |
| 2026-07-08 | “The picture is shifting from short-covering toward position-building into the stocks report.” positioning read off 7-Jul OI +1,288 (broker daily) |
The breakout came the next session. Graded modestly: the read flagged the shift; it did not predict the size. | HIT |
| 2026-07-08 | “Landed-cost figure withdrawn same day after a senior importer flagged it; method rebuilt on the trade's own formula.” process accountability (maund basis error, ~7% overstated) |
Our error, owned in print. The number stays out until it matches the market to the rupee. | MISS |
| 2026-07-07 | “Steady while it holds above 4,531.” daily close, Sep-26 |
No close below 4,531 (closes 4,547 / 4,609). Intraday pierced to 4,528 on 8 Jul (3 points) before closing +62 — noted in the scorecard for honesty. | HIT |
| 2026-07-07 | “A close above 4,576 (with open interest rising) would point to 4,608.” daily close, Sep-26 |
8 Jul closed 4,609 — cleared 4,576 and reached the 4,608 objective in the same session. OI condition: Sep-26 OI +1,288 on 7 Jul (broker daily), building beforehand. | HIT |
| 2026-07-07 | “This week — likely a range (about 4,480–4,608).” week of 7 Jul, Sep-26 |
Close reached 4,609 — one point above our top — but only via a war escalation we did not predict. Range top was right to a point; the 'range' framing was overtaken by the external shock. | PARTIAL |
Every wire headline carries a label — CONFIRMED (2+ named sources), REPORTED (one named source), CLAIM (an interested party says so). When a REPORTED item is later confirmed, we upgrade it and record the change. A walk-back would mean we printed something too confidently and had to retract — that number is the one we work hardest to keep at zero.
This record updates as our desk publishes. If you want the calls before they grade — the daily outlook, the weekly journal, and the verified wire — start your desk.
Market data & analytics — not investment or trading advice. Grades reflect our published wording, verified against official prints.