On Wednesday we told you the rally was a short squeeze. Forced buying, not conviction. We asked what would remain once the people who had to buy were finished.

The answer arrived in two days, and it is not the one we set you up for.

Spot bitcoin ETFs have now taken inflows for eight consecutive sessions. $314 million on Tuesday alone. The week's net was $1.92 billion, the largest since early October. And August is now above $3 billion, the strongest month of 2026 and roughly double what April managed.

That is not a squeeze. That is a bid, and it kept buying after the squeeze finished.

Bitcoin is at $80,776, above $80,000 for the first time since May.

Now the uncomfortable arithmetic, and we would rather do it ourselves than wait for a reader to do it for us.

This is the third time in eleven days we have been wrong about the same thing. On 17 August we led an issue arguing the ETF complex was not the real bid, because STRC had funded roughly 77,000 bitcoin across the year against about 8,000 in net ETF inflows. On Wednesday we conceded that framing looked weak, then immediately replaced it with a new reason the buying was not real: it was a squeeze. Two days later the ETFs printed their best month of the year.

Three swings, one target, three misses. The pattern is not bad luck. We kept looking for a reason the rally was not what it appeared to be, and each time we found one that was true in isolation and wrong about where the money was actually coming from.

Three metrics.

Metric 1, ETF flows: eight straight days, $3 billion in August. This is the number the whole issue turns on. Thirteen US-listed funds, net positive every session for over a week, into a month that is now the biggest of the year. Whatever else is true, somebody with a mandate is buying bitcoin steadily and has not stopped.

Metric 2, Fear and Greed: 71. Down slightly from the 74 we printed Wednesday. Still Greed. What matters is that it did not spike and collapse, which is what a pure squeeze normally leaves behind. Sentiment is holding a level rather than round-tripping.

Metric 3, real yields: 2.32%, down from 2.40%. The quiet one, and possibly the most important. Real yields have fallen eight basis points while the ten year breakeven has risen eleven basis points over thirty days, to 2.32%. Falling real rates with rising inflation expectations is the configuration that historically favours assets with no yield. We have said for a month that a breakeven that moves matters more than a CPI print that does not. It has moved.

📣 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.

Wednesday's issue took our August calls apart one by one. Today we have to do it again, because the correction we published on Wednesday was itself wrong within forty eight hours. That is worth stopping on rather than rushing past.

What we actually got wrong is a method, not three separate calls.

Look at the sequence. On 17 August we compared a year of STRC-funded buying against a year of ETF flows and concluded the ETFs were the minor engine. On 26 August we conceded that and said the rally was a squeeze, forced buying rather than demand. Today the ETFs are eight sessions into their best month of the year.

Each of those claims was supported by a real number. The 77,000-against-8,000 ratio was accurate. The $3 billion of liquidations, 92% of them short, was accurate. What was wrong each time was the inference: that because we could identify one mechanism, we had identified the mechanism. A short squeeze and steady ETF accumulation are not competing explanations. They ran at the same time, and we kept treating whichever one we had measured most recently as the whole story.

So here is the correction, stated as plainly as we can.

There is a real bid under bitcoin right now and it is coming through the ETF wrapper. Eight consecutive sessions, $3 billion in a month, continuing after the liquidation cascade cleared. We spent two issues arguing otherwise and we were wrong both times.

Now the part that is not capitulation, because there is one thing we said that the tape has not touched.

The largest single buyer of 2026 is still not buying. Strategy holds 840,447 bitcoin, unchanged for ten weeks, and $1.59 billion in a cash account it has publicly flagged as available for bitcoin. It has not spent a dollar of it into a 27% rally. That is not us looking for a bearish angle, it is the most conspicuous fact on the board. If the price is right at $80,000, the buyer with a loaded account and a public mandate is behaving very strangely.

What we are changing, and this one is structural. We have been writing the FILTER as a single-mechanism argument, and markets are not single-mechanism. From here, when we name a driver we will name what else is running alongside it and say which we can actually measure. "The ETFs are not the main bid" should have been "STRC outbought the ETFs over the year, and here is what would tell us that had changed." The second version would have survived this month. The first did not survive eleven days.

New dated test, and it is deliberately about the thing we keep getting wrong.

Branch A, the ETF bid is structural. US spot bitcoin ETFs record net positive flows for the month of September as a whole, on Farside's monthly total. Read: August was not a one-off squeeze artefact, the wrapper is now the marginal buyer, and our year-ratio framing is dead.

Branch B, August was the anomaly. September closes net negative on the same measure. Read: the eight-day streak was the tail of a liquidation event rather than a new regime, and the year ratio was telling us something after all.

Graded Wednesday 7 October, once September's monthly total is final. Exhaustive, single measure, public source, and neither branch can pass on noise.

Three numbers that define the week ahead.

$3 billion, August ETF inflows, with two sessions still to run. Already the strongest month of 2026 and roughly double April. The number to watch is not the total but whether the streak survives the month boundary. Flows cluster around month-end rebalancing, so a first week of September that stays positive tells you far more than another big print on Monday does. That distinction is the whole of our new dated test.

2.32%, and 2.32%. The ten year real yield and the ten year breakeven, now identical, having moved toward each other from opposite directions over thirty days: real yields down twelve basis points, breakevens up eleven. Cash is getting less attractive and expected inflation is getting less quiet at the same time. If you only track one macro pair for bitcoin, this is a better one than CPI, because it updates daily and nobody writes headlines about it.

87.9%, the odds of zero Fed cuts in all of 2026. Up from 86.2% on Wednesday, and drifting higher through a week when risk assets ripped. Worth sitting with: the market is simultaneously pricing no easing at all this year and buying bitcoin at $80,000. Either the rally does not need the Fed, which would be a genuine regime change from the last three years, or one of those two prices is wrong. We do not know which and we are not going to pretend to.

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: nothing grades today, and the two open tests point in opposite directions. The 23 September test asks whether Strategy discloses any bitcoin purchase through 21 September. Ten weeks of nothing so far and $1.59 billion sitting unspent, so as of today it is tracking toward Branch B. The new 7 October test set above asks whether September ETF flows close net positive. On the current streak it is tracking toward Branch A. Those two resolving as they currently point would mean the ETFs replaced Strategy as the marginal buyer of bitcoin, which would be the most consequential thing this newsletter has covered all year. We will grade both on their dates whichever way they land.

Still no position, seventh issue running. We have now watched a 27% move without a trade on, and it would be easy to dress that up as discipline. It is partly discipline and partly that we read the move wrong three times, so we would not have taken the trade anyway. Both of those are true and we are not going to only print the flattering one.

Fed decision in September, no change | Market: 67.5% ⚪ NO TRADE, and the hike leg is the live one
No-change at 67.5%, roughly where Wednesday left it, with the hike leg near 31% and a cut priced at 1.1%. The setup we flagged on Wednesday has not fired: we said a push north of 40% on the hike would make this genuinely interesting. It has not happened, despite a week of visibly looser financial conditions. That is mildly surprising and worth watching into next Friday's data.

Zero Fed rate cuts in all of 2026 | Market: 87.9% ⚫ EXITED, and it keeps drifting our way
Up from 86.2% on Wednesday and 85.2% last week. We shorted at 78%, stopped out at 88.75%, and the contract is now within a point of where our stop fired. The loss was real and stays on the record. What the record should also show is that the thesis has been correct for four straight issues since the stop took us out, which is the most expensive kind of right.

CLARITY Act signed into law in 2026 | Market: 14.5% ⚫ EXITED, unchanged, still the cleanest call
Flat since Wednesday at roughly a third of the 41% where the position opened. Congress returns next week, which is the first real catalyst since July. If the merged text moves and this stays under 30%, the re-entry trigger we published fires.

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

Wednesday's question, answered by the flows before anyone could reply.

We asked whether you would deploy Strategy's $1.59 billion at $79,000. Two days later the price is $80,776 and the company still has not spent a dollar. So the question is no longer hypothetical, it is a live observation: the biggest buyer of 2026 is watching a 27% rally from the sidelines with a loaded account.

Meanwhile the ETFs bought for eight straight days.

That is the whole tension in this market right now, and it is a strange one. The buyer who publicly built a war chest is not buying. The buyers nobody credits are buying every session.

So this week's question is about which of those two you would trust as a signal.

If Strategy still has not bought by the end of September, does that tell you anything about price, or only about Strategy? One reading is that a company with the best information on its own cost basis is declining to add at $80,000, and that is a bearish tell. The other is that its balance sheet problems have nothing to do with what bitcoin is worth, and reading it as a market signal is a mistake we have been making all month.

We genuinely do not know. We have argued the first version twice and been wrong on adjacent claims both times, so we would like to hear the second version argued properly by someone who believes it.

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

See you Monday, with Strategy's next 8-K, whether the ETF streak survived the month boundary, and Congress back in session on CLARITY.

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 graded in public. This month it makes uncomfortable reading, which is rather the point of publishing it.

A 27% month is when people who swore they would never buy an exchange again quietly open an account. If you know someone who is about to buy their first bitcoin in a year, send them the first link before they pick a wallet, not after.