There was no issue on Monday. Our publishing system failed and we did not catch it in time. No excuse offered, and we are telling you rather than hoping nobody noticed.

It was an expensive week to be quiet.

Bitcoin is at $79,474. When we last published, on Wednesday 19 August, it was $63,064. That is up roughly 26% in six days, and the move happened almost entirely while this newsletter was dark.

Here is the part that should sting, and it is the reason for the subject line.

On 12 August we set a test that watched whether bitcoin could reclaim $65,000. It could not, and we read that as confirmation of weakness. On 19 August we went further and wrote that the biggest buyer had been dormant for eight weeks, that the ETF complex was running negative, and that “the floor is not where people think it is. There simply has not been a test.”

Bitcoin then broke through $65,000 and kept going. That break triggered the largest cascade of short liquidations since November 2021: more than $3 billion in twenty four hours, roughly 92% of it bearish positions. The level we were watching as a marker of weakness turned out to be the trigger for the squeeze.

We had the level right and the direction wrong, which is the least useful combination available.

What actually moved it was not crypto news. On 19 August, the day we last published, the Treasury announced it would double its long-term bond buyback operations from $2 billion to $4 billion per operation, starting 9 September. Yields fell, the dollar softened, and risk assets caught a bid. Then the shorts got run over.

Three metrics.

Metric 1, Fear and Greed: 74. Greed, and the highest reading we have printed in this newsletter. It was 46 when we last wrote, and 27 a fortnight ago. A twenty eight point move in six days is not sentiment improving, it is sentiment capitulating in the other direction.

Metric 2, ETF flows: roughly $2 billion in a week. This one is uncomfortable. On 17 August we argued the ETF complex was not the real bid, that one company's preferred stock had funded ten times more buying this year. That framing looks a lot weaker after a week in which the wrapper everyone dismissed pulled in two billion dollars while Strategy bought nothing at all.

Metric 3, the Treasury account: down $26.0 billion in seven days. For three issues we tracked this rising and called it a drain on liquidity. It has reversed. Money is flowing back out of the government's account and into the system, which is the mirror image of the story we were telling, and it happened in the same window as the rally.

📣 THIS SPOT IS OPEN

Baseline Crypto reaches 7,000+ Bitcoin-focused readers every Monday and Wednesday morning. Direct sponsor placements are open — premium position, top of the issue, one per send.

Reply to this email for rates.

Three things we published in August, measured against what happened.

One. We treated $65,000 as a ceiling that proved weakness. The 12 August test asked whether bitcoin could reclaim it. It could not, and we scored that as the bearish branch. What we never considered was that a level acting as resistance with heavy short positioning stacked beneath it is not evidence of weakness, it is stored fuel. When it broke, $3 billion of forced buying came through the door in a day. The level was the right thing to watch and we read it exactly backwards.

Two. We said the ETF complex was not the real bid. On 17 August we led on STRC having funded roughly 77,000 bitcoin against about 8,000 in net ETF inflows across the year, and used it to argue the wrapper was a sideshow. Then the ETFs took in about $2 billion in a single week while Strategy bought nothing. The annual figure was accurate. The conclusion we hung on it, that ETF flows were the less important engine, does not survive a week like this one. A ratio measured over a year told us nothing about which engine fires in a given week.

Three. We were building a case on a liquidity drain that had already stopped. Three issues tracked the Treasury account rising and called it a withdrawal of liquidity. It is now down $26.0 billion in seven days. The condition reversed and the reversal, plus the buyback announcement on 19 August, is a large part of why risk assets moved.

Now the part that is not self-flagellation, because there is a real read here.

A squeeze is not the same as a bid. Ninety two percent of $3 billion in liquidations were shorts. That is not patient money accumulating, it is people closing positions because they were forced to. The honest question after a move like this is what remains once the forced buying is finished, and the honest answer is that we do not know yet. The thing that would answer it is Strategy's $1.59 billion. If a company with a cash pool explicitly flagged for bitcoin does not buy near $79,000, that tells you something about what the marginal buyer thinks price is worth.

What we are changing, and it is a process change rather than a market call. We were writing tests that graded direction against a level. From here they grade against a mechanism. “Bitcoin fails to reclaim $65,000” told us nothing about why. “Strategy deploys or does not deploy a disclosed cash balance” is checkable and tells us who is actually buying.

New dated test.

Branch A, the war chest is real. A Strategy weekly 8-K filed on or before Monday 21 September discloses a bitcoin purchase, of any size, funded from the USD Cash account. Read: the pause was liquidity management, the buyer is back, and the 19 August framing of a retired buyer was wrong.

Branch B, the war chest is a story. No purchase is disclosed in any weekly 8-K through Monday 21 September, and the USD Cash balance remains at or above $1.0 billion. Read: the company is holding dollars against a bitcoin position rather than adding to one, and “may be used to acquire bitcoin” was optionality dressed as intent.

Graded Wednesday 23 September. Note that these two branches now cover the outcomes between them, which is more than we managed last time.

Three numbers that define the week ahead.

$1.59 billion, the most important number in this issue. Strategy's new USD Cash account, built from a $2.01 billion MSTR share sale, which the company says may be used to acquire bitcoin. That is roughly 20,000 coins at current prices. Watch the Monday 8-K each week for whether it gets spent, shrinks, or simply sits there. A buyer who announces capacity and then does not deploy it near $79,000 is making a statement about price, and it is not a bullish one.

33.5%, September hike odds, up nine points. They were 24.5% when we last published and no-change has fallen from 74.5% to 66.5%. This is the market noticing that risk assets ripped, the dollar softened and financial conditions loosened, all of which argues for a Fed that stays restrictive rather than one that eases. The ten year breakeven has moved too, from 2.27% to 2.32%. Small, but it is the first time in a month that inflation expectations have moved at all, and we have said repeatedly that a breakeven that moves matters more than a CPI print that does not.

2.69%, high yield credit spreads, still near record tights. Unchanged through all of this. We spent two issues arguing credit was asleep while liquidity drained. Credit was right and we were wrong, and the spread never budged in either direction. Keep watching it anyway, for the opposite reason now: spreads this tight into a squeeze-driven rally are the configuration where a reversal gets amplified.

Forget analyst predictions. Polymarket is a real-money prediction market, where traders put actual dollars on outcomes. Scorecard first, then the board. Not financial advice. Our read. Disclosure: we hold personal positions in Polymarket itself and may earn a commission from Polymarket referrals. These markets are thin and the odds below move, so read direction over ticks.

Scorecard: today's dated test cannot be graded, and that is our fault, not the market's. On 17 August we set a test to resolve today. Branch A required STRC below $100 on Friday's close and a fifth Strategy bitcoin sale. Branch B required STRC at or above $100 and no new sale. What actually happened: STRC closed Friday at $96.18, below par, and the 24 August 8-K disclosed no sale and no purchase. So one leg of Branch A is true and the other is false, and Branch B fails on price. Neither branch resolves.

We wrote branches that do not cover the outcomes between them. Two possibilities, below-par-with-no-sale and above-par-with-a-sale, simply fall through the gap. Last issue we were pleased with ourselves for fixing a test that passed on noise by requiring magnitude. We then immediately shipped a test that cannot resolve at all. Recorded as VOID, not as a pass, and not quietly dropped. The rule going forward is that branches must be exhaustive as well as specific, and the new test above is written that way.

The 23 September test stays open and just got interesting. It asks whether Strategy discloses any bitcoin purchase in a weekly 8-K through 21 September. Two more filings have landed with nothing, taking the drought to ten weeks. But the company now holds $1.59 billion in a cash account it says may buy bitcoin, so the question has gone from “will a dormant buyer wake up” to “will a buyer with loaded capacity pull the trigger.”

Still no position, sixth issue running. We watched a 26% move without a trade on. That is the cost of a rule that says we do not enter without a stated edge, and we are not going to pretend a squeeze we did not see coming was a missed opportunity we identified.

Fed decision in September, 25 bps increase | Market: 33.5% ⚪ NO TRADE, but this is the one to watch
Up from 24.5% when we last published. No-change has slipped from 74.5% to 66.5%. The rally itself is the argument for a hike: financial conditions loosened materially in a week, which is precisely what a restrictive Fed does not want to see. If this pushes north of 40% the September question genuinely reopens, and that is the level at which we would look at it seriously.

Zero Fed rate cuts in all of 2026 | Market: 86.2% ⚫ EXITED, and the exit now looks correct
Up from 85.2%, and drifting back toward the 88.75% where our stop took us out. We have spent four issues saying the exit was early and the loss was real. On this week's tape it no longer looks early. We are not claiming that as a win, because the rule fired for risk reasons rather than because we foresaw this, but the record should show the contract moved back our way.

CLARITY Act signed into law in 2026 | Market: 14.5% ⚫ EXITED, cleanest call on the board
Down from 19.5% and now roughly a third of the 41% where the position opened. Congress returns in September. This one has done nothing but validate the cloture rule we exited on, and it remains the single best decision in the ledger.

Track the whole board live at polymarket.com, free, no account required.

Last issue we asked what holds this market up if Strategy stops issuing. The tape answered a question we had not asked.

Strategy kept issuing, raised two billion dollars, bought no bitcoin, and the price went up 26% anyway. So the marginal bid was not Strategy and it was not steady ETF accumulation either. For one week it was people who had bet against it being forced to buy.

That is the least durable form of demand there is. It also happens to be the most powerful in the short run.

So this week's question, and it is the one we genuinely cannot answer from a spreadsheet.

Strategy is sitting on $1.59 billion it has told the market may buy bitcoin. If you were running that book at $79,000, would you deploy it? The company spent eight months buying at an average of $75,385. Adding here raises that average. Waiting risks the price running away. Doing nothing while holding a disclosed war chest invites the question of what it is for.

There is no obviously right answer, which is why we want yours. If you have run a treasury, managed a corporate buyback, or simply have a view on what you would do with a public balance sheet and a public promise, tell us. We especially want to hear from anyone who thinks the squeeze changes nothing.

Hit reply. We read every response, and the best answers run Friday.

See you Friday, with whether the squeeze held or gave back, what the USD Cash balance does, and September hike odds after a week of loosened financial conditions.

The Baseline Crypto Team

Three tools we built because we wanted them and could not find them. Free, no signup, no accounts, and nothing you type ever leaves your browser.

Is your hardware wallet screwed? Check your device and firmware against every logged advisory, including the ones where updating your firmware does not save you, because the seed itself is already weak.

How fragile is your self-custody? Nine questions, scored out of 27. Every question exists because of a documented loss.

Our track record, losses included. Every dated call we have made, graded in public, including the one this morning we had to void because we wrote it badly.

A 26% week pulls people back in who left after the last drawdown. If you know someone who is about to move coins onto an exchange for the first time in a year, send them the second link before they do it, not after.