Mon-T Weekly Review — w/e 28 Aug 2026

Two weeks ago I wrote that the precious metals complex was the desk's profit engine and everything else was noise dressed up as analysis. This week, the engine seized. Gold, called BULLISH at 7/10, fell 2.45%. Silver, called BULLISH at the desk's maximum conviction of 9/10, dropped 3.41%. Platinum, BULLISH at 7/10, surrendered 2.89%. Three precious metals calls, all wrong, all at high conviction. The NFP regime change thesis that powered three weeks of metals glory has run headlong into a wall of US data releases and dollar strength that the agents apparently forgot to price.

And yet, buried beneath the metals wreckage, the desk produced one of the best individual market calls of the entire year. Wheat, the Market of the Week, was called BULLISH at 6/10 conviction and exploded 11.93% from 700 to 783.5 as the Black Sea export crisis escalated from serious to generational. December wheat futures hit the daily limit on Wednesday, rising 45 cents in a single session as Russian grain export capacity remained offline. Trading Economics confirmed wheat at 758.18 on August 28, up 14.75% over the past month and 46.37% year-on-year. The desk identified the supply shock. It identified the speculative short squeeze setup. And the market delivered a move so violent it would make a Hollywood screenwriter blush.

Five directional calls this week, two correct. Forty percent accuracy. Ten NO CALL markets, six of which were graded correct. The scorecard looks poor, and it should, because calling three metals bullish at high conviction and watching all three go the wrong way is the kind of week that separates honest reviewers from cheerleaders. But the wheat call at 11.93% is not a consolation prize. It is the call of the month, perhaps the quarter, on a market where the desk has been building the thesis since July.

Five directional calls this week, with two landing on the right side. The other ten markets received the NO CALL treatment. A 40% directional accuracy rate is the joint-worst of the recent stretch, matching the w/e 5 Jun and w/e 7 Aug horror shows. The average confidence of 7.0 across those five calls is the highest in months, which transforms this from a bad week into a painful one. When you commit with your strongest conviction of the year (silver at 9/10) and it goes wrong, the confidence column stops being a selling point and starts being evidence.

The calibration story is brutally split. The two correct calls, wheat at 6/10 and soybeans at 6/10, were the lowest conviction on the board. The three misses, gold at 7/10, silver at 9/10, and platinum at 7/10, were the highest. When your strongest conviction produces your worst results and your most measured commitment produces your best, the system is telling you something about the difference between having an opinion and having an edge. The precious metals thesis was an opinion. The agricultural thesis was an edge.

The rolling twelve-week figure sits at 56.1% across 66 directional calls, with 107 no-call abstentions. That engagement split means the desk calls direction on roughly 38% of market-weeks, a rate that has been declining since February's 70% pace when the metals thesis was minting money. This week's 40% does nothing to help, and the persistent mountain of NO CALLs means the denominator barely moves from week to week. The mid-fifties range has become the desk's permanent postcode, and breaking free requires either more directional calls at higher accuracy or acceptance that this is the structural ceiling for the current level of selectivity.

The Fundamental agent had a violently split week that deserves to be studied in a textbook. On wheat, it was magnificent, identifying the Black Sea supply disruption and US production shortfall as a generational squeeze setup that delivered 11.93%. On precious metals, the same agent drove bullish calls on gold, silver, and platinum through structural deficit theses and central bank buying narratives, and all three missed. The supply-demand framework works brilliantly when geopolitical supply shocks are genuinely one-directional. It fails when the macro environment, specifically real yields and dollar dynamics driven by this week's data cluster, overwhelms the physical scarcity story. The Economic agent was bearish on wheat and wrong, bearish on soybeans (technically, via discipline) and the synthesis correctly overrode it. On metals, the Economic agent's persistent warnings about elevated real yields were the voices that should have been heard.

The Institutional agent earned genuine credit on wheat, where its identification of speculators adding shorts into a supply catastrophe was the week's most profitable single insight. On silver, the same agent called bullish on washed-out positioning at the 17.7th percentile, and while the positioning data was real, calling the timing of the re-investment cycle from a monthly floor required more than COT data. It required the macro catalyst to cooperate, and this week it did not.

Jackson Hole on August 27-29 was the week's macro centrepiece, and whatever Kevin Warsh said on Friday will shape September's rate expectations across every asset class. The September 11 WASDE is the next test for wheat's extraordinary breakout. Does the USDA incorporate the full scale of Black Sea damage, or does the global stocks-to-use ratio provide a ceiling? The September FOMC is now the defining event for precious metals, and whether the NFP-driven dovish pivot survives the Core PCE data that landed this week will determine if the desk's bullish metals thesis has legs or was a three-week mirage. NVIDIA earnings on August 26 provided the binary test for the AI capex narrative. Crude oil at $83.35 dropped 4.26% while the desk watched from behind its NO CALL barrier. The desk will have its Sunday views. Given what happened to silver at conviction 9, I expect some serious recalibration on the conviction column.