1An October follow-up crossed even money on both venues within a day of the hike
Two days ago the Federal Reserve raised its target range by 25 basis points, taking the upper bound to 4.00%. On the evening of the decision both prediction venues priced an October follow-up a little under even. Twenty-four hours later both had moved past it.
Polymarket's October contract now prices a 25 basis point increase at 50.5%, up 6.0 points on the day, against 49.5% for no change, down 5.0 points. That market traded $1.64m in 24 hours against $2.46m of standing liquidity, which puts it among the deepest macro contracts on the venue. Kalshi's October series moved the same way and by a similar amount: the hike leg last traded at $0.50, up from $0.46, and the hold leg at $0.50, down from $0.57. Open interest on Kalshi's hike leg rose from 190,948 to 227,744 over the same period, so the move came with new money rather than thin repricing.
Further out, the two venues agree on direction and differ on the December meeting's share of it. Polymarket puts any further hike before 9 December at 83.5%, up 2.5 points. Kalshi's December contract prices a 25 basis point increase specifically at $0.68, against $0.31 for a hold.
David Rosenberg, of Rosenberg Research, made the opposite case on Excess Returns — and made it on the morning of the decision, before the announcement, which he flagged himself. His argument was that the market had already gone too far: the bond market had priced roughly two hikes by year-end against a Fed that had signalled fewer, and had a 4.5% funds rate priced for the end of next year against a Fed projection nearer 3.6–3.8%. "I don't see the Fed ratifying what the markets are suggesting right now," he said, and drew from it a bullish call on Treasuries.
The market spent the following day moving further in the direction he called overdone. That does not settle the argument — his thesis runs to next year and rests on events in November, not on one session — but it is the cleanest test available of how much of this repricing was already in the price. What would settle it is the pair of dates he named: 3 November, the midterm elections, and 4 November, the Treasury's refunding announcement.
2The path repriced upward and inflation expectations repriced down, and the split shows in which assets recovered
The most informative print of the day was not on any screen the sources were watching. The five-year breakeven inflation rate closed at 2.32% on 17 September, down from 2.35% the day before, 2.41% on the 15th and 2.46% a week earlier. So in the same 24 hours in which the market added six points to the odds of another hike, it took three basis points off what it expects inflation to average for five years. That is what a market treating a central bank as credible looks like.
Rosenberg had the longer-horizon version of the same measurement. Since the 10-year yield's late-February low below 4%, he said, "90% of the increase in the 10-year Treasury note yield has come from the real interest rate" and only 10% from market-based inflation expectations, which he put at 13 basis points. Without the real-rate and term-premium move, on his arithmetic, "the 10-year note yield now would be 4.1%, not five." The 10-year inflation-protected yield backs the shape of that: it printed 2.68% on 16 September against 2.60% two sessions earlier, while the nominal 10-year went 4.97% to 5.01%.
The near-term picture points the other way, and the contradiction is real rather than an artefact. Polymarket's ladder on September annual inflation shifted up for a second consecutive day: 3.6% now prices at 45.0%, up 5.5 points; 3.7% at 27.0%, up 6.5 points; 3.5% fell to 18.5%, down 5.5 points. Roughly 82% of that ladder now sits at 3.6% or higher. It is a thin market — $2.4k of 24-hour volume against $17.6k of liquidity — so treat it as a weak reading. Taken together with the falling breakeven, the coherent interpretation is that the September print is expected hot because the oil pass-through is already in the data, and that the market expects the Fed to take it out afterwards.
Thursday's tape sorted itself along exactly that seam. The long end and the duration-sensitive assets recovered: closed +1.12%, +1.71%, +1.73%, +2.18%, with +1.14%. Semiconductors led it — +2.70%, +5.45%, +7.65% — against software at +0.66%. On The Compound's Animal Spirits, recorded on the Tuesday, Michael Batnick described semiconductors having "their worst day relative to software of all time" with semis down five and software up five. That relationship reversed inside two sessions.
The exception is the one that carries the information. closed −0.12% and finished dead flat, after what Prof G Markets recorded as bank stocks' worst day since February on concerns that higher rates would slow lending growth. A long end that rallies while the front end reprices higher is good for anything with duration and cash flows in the future, and does nothing for a business that lends against the front end. Rosenberg's two designated hiding places went along quietly rather than leading: +0.60% and +0.16%.
3The tightening is landing on an economy whose main transmission channel is already closed
Housing starts printed on 17 September at a seasonally adjusted annual rate of 1,275,000, down from 1,309,000 in July and 1,439,000 in June — a fall of 2.6% in the month and 11.4% over two. Lennar posted a downbeat third quarter the same morning; still closed +1.66% and the homebuilder group at +0.58%, on a day when the market was buying duration generally.
Two of the sources reached the same conclusion about what that means for policy, independently and in different registers. Robert Armstrong, of the Financial Times' Unhedged newsletter, on Prof G Markets: "the standard mechanism by which that happens is the housing market… Well, we are in the situation now, and you can call this fortunately or unfortunately as you please, where the housing market already sucks. So it can't be damaged that much more by higher rates." Cullen Roche, of Discipline Funds, on The Compound: "the Fed's policies work mainly through the real estate market… and the real estate market has been obliterated."
The rest of the economy is not behaving like something being slowed. Advance retail sales for August, released on 16 September, came in at $773.9bn against $764.5bn in July, a rise of 1.24% in the month. The Atlanta Fed's GDPNow estimate for the third quarter, updated on 17 September, stands at 5.08%, against 1.54% for the second quarter. Warsh's own framing at the press conference, as Armstrong relayed it, was that the 10-year yield has risen "first and foremost because the economy has strengthened" — a reading Armstrong only partly accepted, putting growth behind inflation and central-bank expectations as a driver, and hyperscaler competition for capital behind both.
Both Armstrong and Roche named the same alternative channel, and neither thinks the current move reaches it. Armstrong: the people building data centres are "extremely price insensitive", so "another 100 basis points on your interest bill are not going to bug you" — 25 wouldn't change a decision and 75 probably wouldn't either, but it becomes a factor six months out. Roche put data-centre investment as the offset that has been holding up total investment while residential and corporate spending declined, and named a pullback in it as the clearest recession risk he could identify. gave the mechanism a data point on Thursday, reversing a pre-market gain to close −4.20% after announcing a $3bn debt raise.
The consequence is that the visible cost of this tightening falls on the part of the economy that already stopped, while the part actually driving growth is being financed by borrowers who have not yet noticed the price. That is an argument for the tightening running longer than the first hike implied, not shorter.
4The barrel is being priced off the diplomatic track while the shipping data still reads like a war
Oil fell for a third consecutive session, Reuters reported, on hopes of limited Saudi supply disruption. closed −0.54% on Thursday after −3.51% on Wednesday, a fall of about 4% across two sessions, with +0.64% and +0.47% both diverging from the barrel. The supporting evidence for the de-escalation read is concrete. China's August refined-fuel exports exceeded pre-war levels with jet fuel at a record. US officials met Houthi representatives in Oman over the weekend. China is pressing Iran to help rein in the Houthis after a Saudi appeal. The United States will allow Iranian leaders to attend the UN General Assembly. Poland's Orlen covered the Saudi disruption by buying 16 extra crude cargoes.
Set against that, the physical data has not improved at all. Three ships transited the Strait of Hormuz on Wednesday. Container shipping rates could test record highs as the fuel spike feeds through, analysts told Reuters, and the conflict is driving a wave of supertanker orders. Saudi and Houthi forces exchanged strikes as the war spread into Yemen. JPMorgan said it sees no clear endgame for the oil market while the conflict drags on.
The prediction markets sit closer to the freight data than to the barrel. Polymarket prices a US invasion of Iran before 2027 at 16.5%, unchanged, on $281k of 24-hour volume against $819k of liquidity — no escalation to a ground war, and no resolution either.
So the two are reading different instruments. The price is tracking a diplomatic track that has produced meetings and statements; the transit counts, freight rates and tanker orders are tracking a physical constraint that has not moved. Near-term the price side is winning, three sessions running, and it is winning on something substantive — Saudi supply proving more resilient than feared, and Chinese export volumes showing product still flowing. What would break it is a single day on which the Hormuz transit count fails to recover, because the barrel is currently priced as though it will.
What the sources recommended
Three single-name or single-fund calls came out of six videos, and each one carries a disclosed conflict or a caveat from the person making it — which is most of what there is to say about the day's recommendations.
Ross Givens made the clearest buy call: , Helis Pharma, a clinical-stage biotech whose lead candidate is a psilocin-based oral pill for major depressive disorder, taken as two doses three weeks apart alongside an existing antidepressant. His evidence is insider buying — the founders bought roughly $7m of stock on the open market between 31 August and 8 September, about seventy times their previous purchase — ahead of Phase 3 topline data expected in the fourth quarter, with a TD Cowen price target lifted from $8 to $62 on 20 August. The caveats are his own and they are substantial: the striking Phase 2 numbers he quotes, 100% response and 71% remission at a year, come from seven patients, and the lower-dose group was 60% and 50%. He also discloses that he sent the same name to paying subscribers on 15 September, two days before publishing the video, and roughly a third of the runtime is a pitch for a $5 membership. closed at $14.02 on Thursday.
Andrew Mitchell, of Ophir Asset Management, on Equity Mates, gave the most worked-through thesis of the day for , Codan, which lists on the ASX. Half the business sells gold detectors, where he argues a doubled price on a newly released model is not reflected in the margin the market is forecasting; the other half sells the communications systems inside drones sold into Ukraine, where he places the company second on the battlefield behind Motorola and taking share. He attaches three risks unprompted: the stock trades near 40 times earnings and cannot afford a slip; a falling gold price would hurt detector demand in Africa; and an end to the Russia-Ukraine conflict, which he calls good for the world and bad for the company. The hosts disclosed they are unit holders in his global fund.
Alex Morris named , his own firm's ultra-short-duration TIPS fund, on The Compound's Ask the Compound — answering a question about whether a 5.2% bond yield is real, and doing so in a segment that opened by noting his firm had just sold to T. Rowe Price. Take it as the product pitch it is presented as.
Rosenberg's calls were at the asset-class level rather than the single name, so they do not appear in the day's ticker tables: long Treasury duration on a 2s10s curve he expects to steepen, long gold on central-bank reserve demand, short the US dollar as a high-conviction theme, and an explicit refusal to own the S&P 500 as an index on concentration grounds, with the top ten names above 40% of it. He also disclosed raising 10% cash recently, leaving his model portfolio at 50% equities, 30% bonds, 10% cash and 10% commodities. One number in that segment does not match the public series: he said high-yield spreads had widened "more than 20 basis points" so far this month, where the ICE BofA US High Yield index option-adjusted spread went from 2.65% on 1 September to 2.70% on the 16th, a move of five.
The refunding announcement, not the next meeting, is where this gets settled
Nothing material prints on Friday, and the September quarterly expiry falls on it, so the session is likelier to be about positioning than about news. The day's live catalysts are all next week and beyond: Xi Jinping arrives in Washington, with Alice Han of Greenmantle telling Prof G Markets to expect symbolism rather than substance, and at most a modest tariff reduction or an extension of the existing pause alongside China's rare-earths pause; Scott Bessent meets He Lifeng in New York over the weekend; and the UN General Assembly convenes, which is where the Iran diplomacy either acquires substance or visibly does not. Earnings resume on 22 September with , , and — directly on the housing thread above — the homebuilder .
The two dates that actually test the claims here are further out and both belong to Rosenberg. He named 30 September, when a revised set of PCE deflator data lands that he expects to show inflation lower than currently measured, and 4 November, the Treasury's refunding announcement, where he expects a repeat of the autumn 2023 manoeuvre: flood the market with bills, cut issuance at the long end. On his reading that, rather than the October meeting, is what determines whether the long end follows the policy path higher or breaks away from it. It sits one day after the midterms, which Polymarket prices at 89.5% for a Democratic House — up one point, on $2.23m of liquidity — and at 59.5% for a Democratic Senate, the outcome behind his fiscal-gridlock thesis.
