Talk Is Cheap. Numbers Matter.
Everybody had an opinion on the Fed this week. None of it mattered as much as one Friday morning number. Fed Chair Warsh’s hawkish tone from last week’s Jackson Hole speech split economists into two predictable camps, some said it raised the bar for him to actually deliver a hike, others said the case for one still isn’t there. A separate Fed official offered something closer to a truce, he could support holding rates steady if the next couple weeks of inflation data confirm real cooling. Japan wasn’t nearly as chatty about it, its central bank just came out and said it is hiking, and that alone hit the dollar hard, lifted the yen, and gave crypto and metals their momentum back. Then Friday’s jobs report showed up and made most of the week’s arguing look premature, coming in far stronger than anyone expected. Next week’s inflation data is what actually decides who was right.
The dollar broke, gold and crypto bounced, then Friday undid a chunk of it.
Bitcoin is up 1.6% since Sunday’s opener despite a rough middle of the week, while gold and silver both gave back a little ground even as funding on both actually built up real conviction underneath. WTI crude is the real outlier, up 7.1% on genuine supply fear out of the Middle East, not a rate story at all, even as its own funding went deeply negative. Stocks barely moved either direction. Section 03 has the actual story behind the whipsaw.
Grading every call from Sunday’s calendar against what actually happened.
Warsh split the room. Waller floated patience. The Bank of Japan did not hedge at all.
Chair Warsh’s hawkish comments at Jackson Hole, the same ones that woke the bear flattener out of hibernation, split the Fed and split the banks covering it. Two banks published opposite reads: Citigroup called it only marginally more hawkish than usual and does not see a hike consensus forming by September, while Bank of America called it a more credible Fed and said the onus is now on Warsh to actually deliver a hike. The Fed’s own governors split the same way. Governor Christopher Waller struck a patient tone Thursday, saying he could support holding rates steady if the next couple weeks of inflation data confirm real cooling. Governor Michelle Barr pulled the other way: she flagged that oil prices, driven by a fresh round of US strikes on Iran, could force the Fed’s hand if inflation does not cool, and crude trading above $90 a barrel this week gives her point real weight.
Japan’s central bank left no room for interpretation. Two officials, including Governor Ueda, all but confirmed a hike is coming at the September 18 meeting. A central bank about to raise rates makes its currency more attractive, and that alone is what pulled the yen up and pushed the dollar down this week, no rumor, no intervention, just investors pricing in a real hike. That dollar weakness is what gave gold and Bitcoin their bounce.
Friday’s strong report and positive revision settled the argument, for now. September hike odds landed at 58.4%, almost exactly where they sat the day after Warsh’s speech (57.0%), after dipping to a coin flip midweek on Waller’s comments. A whole week of arguing round-tripped back to almost the same number it started at.
A firm BOJ hike call beat a Fed still arguing with itself. Then the jobs number argued back.
The three scenarios from Sunday’s opener were all built around the US Treasury curve alone, none of them modeled a foreign central bank as the actual driver, worth being honest about that. Chain 4’s predicted outcome, a clean reversal of the prior week’s selloff, is roughly what happened to gold and crypto midweek, but the opener’s own mechanism for getting there was incoming US data softening. That is not what happened, it was a Bank of Japan versus Fed currency story instead. The one number the opener actually flagged as the real test, the 30-year Treasury yield, still has not moved, up just 1 basis point to 5.252% even after today’s blowout jobs number. The flattening question from Sunday is genuinely still open heading into next week’s inflation data.
A cascade-heavy stretch, with a late bounce.
BTC: $157M liquidated since the last Opener, majority long-side, before a late short squeeze pushed price higher.
Since we last checked in, BTC finished up about 1% overall, but it was a rough ride to get there. $157 million got liquidated across the market this stretch, and more of that was long positions getting forced out than short ones, a real cascade, not a squeeze. Almost all of the recovery happened in just the last day, when open interest fell close to 2% and price jumped nearly 3% as a fresh wave of short covering took over. Funding stayed calm and the profitable traders barely shifted their positioning. Net result, a choppy stretch that ended on a bullish note, with most of the actual bounce packed into that final day.
This comes straight from Hyperliquid, a crypto exchange where every trade and every wallet is public. We pull it to see who’s actually winning right now and which way they’re leaning, not just what price did.
ETH: $25M liquidated since the last Opener, three-quarters long-side, with only a partial bounce so far.
Since we last checked in, ETH finished down about 1% overall. $25 million got liquidated across the market this stretch, and three out of every four dollars of that came from long positions getting forced out, a real cascade. The most recent day did turn calmer, open interest eased while price rose close to 3%, but that bounce hasn’t been enough to erase the damage from earlier in the period. One large long position is also still sitting close enough to forced liquidation that it’s worth watching. Net result, a rough stretch overall with only a partial recovery in that final day.
Same read as the BTC section above, pulled from Hyperliquid for ETH specifically. Same exchange, same method, different asset.
Reading The Pulse. Six things get pulled and scored every check. Open interest against price shows the crowd’s real position, not just its size: both rising means fresh longs are opening, both falling means longs are capitulating, OI rising while price falls means fresh shorts are opening, OI falling while price rises means shorts are covering. Funding’s 1-day trend against its 7-day baseline: running above the 7-day average means the cost of holding longs is rising, running below means that cost is easing. TWAPs shows both the largest order on the board and the fastest one, often different wallets, since the largest shows commitment and the fastest shows who’s actually pressuring price right now. Liquidations already realized are read by which side got forced out, since forced-out longs and forced-out shorts mean opposite things. Liquidations still ahead looks at the nearest-to-forced-out wallets on each side, since the single largest position is often not the closest to getting wiped out. And Tug of War compares the profitable wallets’ positioning to the underwater crowd’s: whichever side the profitable wallets are leaning is the side currently winning.
Metals funding built real conviction. QQQ flipped negative. Oil went deeply negative despite the rally.
The best single day of the stretch landed right in the middle of the dollar’s slide.
A Bitcoin ETF is a fund big investors use to get exposure to Bitcoin without holding it directly. When money flows into the fund, that usually means institutions are buying. When it flows out, that usually means they are selling. This week, the tracker showed a net inflow of $812.26 million over four trading days, pushing total assets in Bitcoin funds past $103 billion for the first time. Wednesday alone brought in $730.87 million, the single largest day of the stretch, landing right in the middle of the dollar’s slide.
Ethereum funds took in a smaller but still positive $191.94 million over the same window. Real, positive demand on both, running straight through a week that also produced the heaviest liquidation stretch this tracker has logged, buying pressure and forced-selling pressure running side by side rather than one canceling out the other.
The Fed never actually reversed course. This week just made that clearer.
This week’s bounce came from a genuine cross-currency divergence, not a Fed reversal, and Friday’s jobs report argues the underlying hawkish case never actually went away. That’s worth sitting with rather than reacting to the week’s price swings alone.
Gold and silver funding both built real conviction this week even as price gave ground, QQQ funding flipped negative for the first time in a while, and Bitcoin logged its heaviest liquidation stretch on record, with a narrow long-side tilt.
This week gave a real answer on where Fed leadership actually stands, split, not unified, and Friday’s jobs report just handed the hawks the stronger argument again.
Bitcoin’s funding rate eased to 6.40% annualized over the past seven days, down modestly from 7.24% last issue. The buy-spot-sell-perpetual trade from Lesson 3 is still paying a real, live yield at these levels.
Read the Lesson →The covered call from Lesson 5 is currently short the October $90,000 call, 55 days out, carrying roughly $1,600 in total premium, an annualized yield of about 13.3% on the position. With BTC sitting at $79,580.20, that strike sits about 12.7% above spot. For comparison, a tighter October $88,000 call, roughly 10% out of the money, is pricing at $1,975 in premium right now, a 16.3% annualized yield, a real reminder that pulling the strike closer to spot buys a meaningfully higher yield at the cost of less room before assignment.
Ethereum shows the same tradeoff, richer across the board. The October $2,800 call, roughly 14% out of the money, is paying $84 per coin in premium, a 22.5% annualized yield. Pull that strike in to the 10% out-of-the-money level instead and premium jumps to $108 per coin, a 28.9% annualized yield, the same strike-versus-yield tradeoff, just a lot more expensive to buy room on right now.
Read the Lesson →The Bitcoin against the Nasdaq 100 spread sits at roughly 110.9 now, up from 108.7 at last issue as Bitcoin outpaced the Nasdaq 100 this week. Still well above the 87.08 entry.
A second rotation worth watching from the same lesson: buying earners, companies actually profiting from the AI buildout, and selling spenders, the ones pouring money into it without the profits to show yet. Intel against Tesla is this week’s example. That ratio ranged from about 0.236 to 0.270 this week, a real dip Thursday that popped decently back up by Friday. Funding sits positive on both legs right now, a small cost on the long Intel side, a small credit on the short Tesla side. Watch for the next dip toward the low end of that range to add to the rotation.
Read the Lesson →This content is produced by Harmonic for educational purposes. It is strategy education, not investment advice.