Mon-T Weekly Review — w/e 17 Jul 2026

Two markets told the entire story of this week, and they told it in opposite directions. Wheat, the Market of the Week and the desk's first ever MOTW in the agricultural space, was called BULLISH at 7/10 conviction and exploded 8.35% from 632 to 684.75 as the worst US production since 1957 finally stopped being a thesis and started being a price. Trading Economics confirmed wheat climbed toward $7 per bushel in mid-July, its highest level since May 2024, with Black Sea strikes and relentless drought adding fresh accelerant to a fire the desk had been trying to light for months.

Then there was crude oil. Called BEARISH at 3/10, the lowest conviction the desk has ever whispered on a directional call, WTI surged 14.16% from $71.41 to $81.52. Forbes confirmed WTI opened the week at $78.08 on July 14, already well above Friday's $71.41, as renewed Strait of Hormuz tensions and depleted US petroleum reserves sent energy prices on one of their most violent weekly rallies since the March war peak. The geopolitical premium that the desk has spent the entire summer declaring dead apparently has a different opinion about its own mortality.

Four of eight directional calls correct, a 50% accuracy rate. Not the worst week on the books, but firmly at the threshold where following the desk offers precisely zero statistical advantage over flipping a coin. The saving grace, if we are feeling generous, is that the four correct calls included the week's largest move by a considerable margin, and the four misses included a crude oil call that was barely a call at all.

Eight directional calls this week, with four landing on the right side. The other seven markets got the NO CALL treatment. A 50% directional accuracy rate matches the late-March coin-flip week that I described at the time as the desk's lowest point. The average confidence of 5.4 tells you the desk was barely committing on most of these, which is reflected in the fact that the two highest-conviction calls (wheat at 7/10 and platinum and RTY both at 6/10) all delivered, while three of the four misses came from calls at 5/10 or below. When your highest conviction produces your best result and your lowest conviction produces your worst, the calibration system is doing something right even if the aggregate number does not show it.

The NO CALL markets included the Nasdaq shedding 4.44%, gold dropping 2.33%, and soybeans rallying 1.76%. I have written some variation of this paragraph approximately forty times in 2026. The desk sits out the Nasdaq. The Nasdaq moves dramatically. I write an annoyed paragraph about it. The Nasdaq does not care. At this point, the NQ NO CALL saga has its own narrative arc, character development, and what feels like a deliberate refusal to reach a satisfying conclusion.

The rolling twelve-week figure sits at 56.6% across 76 directional calls, with 100 no-call abstentions. That engagement split means the desk calls direction on roughly 43% of market-weeks, a rate that has continued its slow decline from February's 70% pace. This week's 50% does nothing to improve the rolling number, though the late-April catastrophes are now ageing out of the window. The desk has been lodged in the 53-58% rolling band since the Iran conflict rewrote the playbook in March, and at 56.6% we are at least in the upper half of that range. To crack meaningfully above 60%, the desk would need to increase directional volume while maintaining accuracy above 65%. At the current pace, that feels aspirational rather than imminent.

The Fundamental agent had a split week that perfectly illustrates its year-long identity crisis. On wheat, it was magnificent, driving the correct BULLISH call through its identification of the WASDE production catastrophe and the drought thesis that has been its pet project since March. On crude oil, its structural oversupply and demand destruction framework was the intellectual foundation for a bearish call that missed by 14%. Same agent, same week, opposite outcomes. The lesson remains the same one I have been writing since the Iran war began: the Fundamental agent reads medium-term supply-demand dynamics correctly but cannot price binary geopolitical catalysts. When Hormuz tensions flare, structural analysis stops working.

The Economic agent had a poor week across equities, correctly identifying the June CPI at 4.0% as a bearish bond catalyst (which then went the wrong way as bonds rallied marginally) while supporting the ES BULLISH call that missed by 1.74% as the Nasdaq's 4.44% decline dragged everything lower. The Sentiment agent continues to be the desk's most underweighted discipline, and this week its cautious reading on equity put/call complacency at 0.55 looks prescient in retrospect. Silver's bearish call was well-served by the Economic agent's real yield framework, delivering another 6.30% decline.

The calendar is heating up. The June CPI that landed on July 14 within the grading window appears to have rattled equities and supported crude oil's surge. The ECB meets July 23 with markets pricing 70% probability of a rate hike, which could finally break the desk's nineteen-week NO CALL streak on EUR/USD. The FOMC follows on July 28-29, where Warsh faces the question of whether the June CPI at 4.0% demands hawkish escalation or whether the June NFP miss at 57K argues for patience. Wheat's 8.35% breakout raises the question of whether the run toward 700+ continues or exhausts itself at the 52-week high. And crude oil at $81.52 is back in a zone where the desk's bearish thesis has historically been destroyed by geopolitical headline risk. The desk will have its Sunday views. Given what crude oil just did to the scorecard, I expect the conviction on energy to be approximately zero.