Mon-T Weekly Review — w/e 20 Mar 2026

There are weeks where you grade a prediction desk and feel satisfied. There are weeks where you feel sympathetic. And then there are weeks like this one, where the FOMC delivered a hawkish hold on Wednesday, gold flash-crashed 11% in two sessions, silver dropped 17%, copper shed nearly 8%, and the Strait of Hormuz remained largely closed while Iraq declared force majeure on its oil exports. If you thought the Iran war week three weeks ago was violent, this one made it look like a warm-up act.

The desk's directional accuracy was solid on paper: five of seven directional calls correct, a 71.4% hit rate that maintains the recovery trajectory from the 46.7% horror show of w/e 6 Mar. But the real story lies in the eight NO CALL markets, five of which experienced moves so large they would make a directional forecaster weep. Gold's 11.26% collapse on a NO CALL. Silver's 16.63% freefall on a NO CALL. Copper losing 7.79% on a NO CALL. Soybeans shedding 5.27%. These are not background noise. These are the kind of moves that define quarters, and the desk sat on its hands for all of them.

To be fair, the system did what its rules demanded. The FOMC meeting created binary event risk that the agents correctly identified as unforecastable. But when FinancialContent runs a headline reading 'Precious Metals Flash Crash: Gold and Silver Plummet Amid Middle East Chaos and Hawkish Fed Pivot,' and your desk issued NO CALL on both metals, the intellectual honesty of the process offers cold comfort. The desk was right about uncertainty. It just failed to be right about direction.

Seven directional calls this week, five on the right side, two wrong. The other eight markets received the NO CALL treatment. On the surface, 71.4% directional accuracy is respectable, the joint-best since the 93% outlier of w/e 27 Feb. But the average confidence of 6.1 tells you the desk was not exactly pounding the table. The highest conviction call among the directional group went to the Aussie dollar at 7/10 BULLISH, which promptly fell 0.62%. Every other directional call sat at 5 or 6 out of 10.

The confidence calibration puzzle this week is not about the directional calls. It is about the NO CALLs. When eight of your fifteen markets are abstentions, and five of those eight produce moves exceeding 2%, you have to ask whether the desk's noise threshold framework is fit for purpose during wartime FOMC weeks. The system correctly anticipated that the March 18 Fed meeting would create binary risk. It did not anticipate that the resolution would be this violent. Gold's 11.26% drop, triggered by what CNBC described as 'gold and silver crushed as Fed's hawkish hold reshapes 2026 outlook,' is the kind of move that a NO CALL cannot capture by design. The question for subscribers is whether missing a crash is better or worse than being on the wrong side of one. I know which I'd choose.

The rolling twelve-week figure sits at 54.4% across 125 directional calls, with 53 no-call abstentions. That engagement rate tells you the desk calls direction on roughly 70% of market-weeks, sitting out the rest. The rolling number remains stubbornly in the mid-fifties, weighed down by the late-December catastrophe and the w/e 6 Mar war-week meltdown. This week's 71.4% directional accuracy helps, but with only seven directional calls adding to the denominator, the needle barely moves. The desk needs several more weeks above 65% on higher directional volume to crack through the 57% ceiling that has felt permanent since March began.

The Fundamental agent had its strongest week relative to the rest of the pack, correctly driving the bearish crude thesis through producer hedging analysis and supporting the bearish bond call via the $1.9 trillion fiscal deficit framework. Its commodity supply-demand work on CL, identifying the IEA demand downgrade and structural oversupply, proved to be the most useful lens. The Economic agent also performed well, with its focus on the FOMC binary risk and hawkish hold scenario proving prescient after the March 18 decision.

The weakest performer was the collective framework's inability to position for the precious metals crash. Gold fell 11.26%, silver 16.63%, and copper 7.79%, and the desk had NO CALL on all three. The agents correctly identified the FOMC as a binary risk event, but the system's response was to abstain rather than acknowledge that the hawkish scenario had meaningfully higher probability given oil at $100 and inflation fears mounting. I have been saying for weeks that the Fundamental agent needs a macro override switch for wartime conditions. The March 19 precious metals flash crash is what happens when that switch does not exist.

The wreckage from the Fed's hawkish hold and the precious metals crash will dominate early next week, with gold at $4,492 and silver at $67.81 representing generational buying opportunities or the beginning of something far worse depending on your time horizon. The BoE delivered its rate decision on March 19, the ECB met March 18-19, and the full implications of a 'no cuts in 2026' dot plot scenario are still being priced across asset classes. Oil at $112 Brent with the Strait of Hormuz effectively closed and Iraq on force majeure creates a genuine energy crisis scenario that Goldman Sachs warns could push prices to $150 in a bull case. The March 31 quarter-end is now ten days away, with window-dressing flows adding to an already volatile mix. The desk will have its Sunday views. Given the scale of this week's moves, I expect the NO CALL count to decrease as the post-FOMC clarity the agents were waiting for has now arrived. Whether that clarity produces better calls than this week's abstentions did remains to be seen.