Mon-T Weekly Review — w/e 2 Oct 2026

After last week's 80% headline that had me feeling cautiously optimistic about the desk's trajectory, this week delivered something almost as clean. Four directional calls, three correct, a 75% accuracy rate. The Nasdaq, this week's Market of the Week, was called BULLISH at 7/10 and edged 0.62% higher to close at 31,080, quietly slipping past the all-time high of 31,028.50 that had been lurking like a toll booth on a motorway. Treasury bonds, called BEARISH at 7/10, collapsed 1.88% as 30-year yields continued their violent march toward levels not seen since the early 2000s. And the Russell 2000, BEARISH at 5/10, scraped a correct result with a 0.24% decline that barely qualified as movement but technically went in the right direction.

The lone directional miss belongs to copper, called BULLISH at 7/10 on the Escondida strike vote thesis, which fell 1.64% as the broader selloff in industrial metals overwhelmed the supply disruption catalyst the desk was banking on. That is now a theme I have been tracking for months: supply-side analysis on copper works until the macro regime decides it doesn't, and this week the macro regime was not in a cooperative mood.

The ten NO CALL markets produced the familiar parade of missed opportunities I have been documenting since March. Silver dropped 5.28%. Soybeans crashed 3.17%. Platinum fell 3.57%. Wheat lost 2.88%. The Aussie dollar shed 1.10%. EUR/USD slipped 1.06%. The desk said nothing about any of them. I have written some variation of this paragraph so many times it now has its own filing cabinet in my office, complete with alphabetised tabs for each market the desk has watched decline from behind the signal threshold barrier.

Four directional calls this week, with three landing on the right side. The other ten markets got the NO CALL treatment. A 75% directional accuracy rate sustains the improved trajectory from last week's 80%, and the average confidence of 6.5 across those four calls represents the desk's strongest average commitment in recent memory. Both 7/10 calls that went the right way, NQ BULLISH and ZB BEARISH, delivered cleanly. The RTY BEARISH at 5/10 scraped through with a 0.24% decline that is the directional equivalent of winning a chess game on time rather than checkmate. The sole miss, copper BULLISH at 7/10, stings because it was the desk's joint-highest conviction and the Escondida strike vote thesis sounded compelling on Sunday evening.

The real story, as always, lies in the abstentions. Six of ten NO CALL markets produced moves exceeding 1%, with silver's 5.28% drop and soybeans' 3.17% crash leading the parade. When your abstentions produce bigger moves than your commits, the framework is protecting your accuracy percentage at the cost of demonstrating relevance. That trade-off has been the desk's defining characteristic since the Iran conflict rewrote every correlation in March, and nine months later the song remains the same.

The rolling twelve-week figure sits at 53.8% across 65 directional calls, with 111 no-call abstentions. That engagement split means the desk calls direction on roughly 37% of market-weeks, a rate that has been declining since February's 70% pace and is now approaching the level where the NO CALL column does three times the work of the directional one. This week's 75% on four calls nudges the number modestly, but with only four calls entering the denominator, the needle barely moves. The desk needs to either increase its directional volume while maintaining accuracy above 65%, or accept that the mid-fifties is the structural ceiling for a framework operating at this level of selectivity. At 53.8%, we are just above the threshold where following the desk offers measurable advantage over a coin.

The Economic agent had the strongest week across the directional calls, correctly supporting the bearish Treasury bond thesis through the hawkish Fed repricing and fiscal deficit deterioration framework that has been the desk's most reliable analytical lens since the Warsh era began. Its identification of 10-year yields at 5.17% as a structural headwind for the broader commodity complex proved directionally correct across multiple markets, even those the desk did not formally call.

The Fundamental agent continues its year-long pattern of being correct about supply dynamics and wrong about timing. Its Grasberg supply deficit thesis on copper drove the BULLISH call that missed, as the structural supply story was overwhelmed by the macro selloff. On silver, where the structural deficit thesis remains intact, the metal dropped 5.28% while the desk watched from behind the signal threshold. The Sentiment agent earned quiet credit on the Nasdaq, where its identification of extreme AAII bearishness at 48.1% as a contrarian buying signal proved to be the analytical foundation for the week's MOTW call. After months of being the desk's least-weighted discipline, the Sentiment agent is earning its keep when the retail crowd capitulates at exactly the wrong moment.

October opens with two defining themes. Q3 earnings season begins in earnest, and whether the AI capex cycle that powered the Nasdaq through the ATH can deliver revenue conversion will determine if the breakout holds or reverses. The October 28 FOMC looms as the next major Fed event, with CME FedWatch pricing 75.8% probability of a 25bp hike that would take rates to 4.00-4.25%, a level the bond market is already pricing with extreme prejudice at 30-year yields near 5.50%. The Hormuz de-escalation narrative that sent crude oil below $92 this week could accelerate if Iran's 7-day reopening offer gains traction, potentially removing the geopolitical premium that has distorted energy markets since February. And the desk needs to find a way to participate in the precious metals and agricultural moves it keeps watching from behind the NO CALL curtain. Ten abstentions with a 5% silver decline and a 3% soybeans crash is procedurally clean and practically invisible.