Friday's issue left one number open. BlackRock's IBIT had not reported its flow for Thursday 8 October, and we wrote down in advance what it would mean either way. A big outflow would make the largest day of the week also the latest, and that would be a direction. A small one would leave Wednesday standing on its own.

IBIT reported minus $5.5 million. Thursday finished at minus $244.1 million, about half of Wednesday's minus $484.9 million, and Friday 9 October came in at plus $21.1 million. On our sum of Farside's printed daily totals, the week of 5 to 9 October was minus $678.9 million. Wednesday stands on its own.

The total is not the useful part of the final table. The useful part is where the money came out.

Fund by fund, Monday 5 to Friday 9 October:

Two funds, Fidelity's and ARK's, account for $587.5 million of the $678.9 million, or 86.5%. The biggest bitcoin ETF there is finished the week one million dollars up. IBIT did have a bad day, minus $207.7 million on Wednesday, but it took in $69.9 million on Monday, $122.0 million on Tuesday and $22.4 million on Friday, and those netted it back to flat. FBTC had an outflow on four of the five days and an inflow on none.

Here is what that tells you, and where it stops. Fund flows show which product the money left. They do not show who sold, and Friday's point still holds: the filings that name holders are 13Fs, which report quarter-end positions, so October's sellers will not be visible until the middle of February. What the table does allow is narrower. If ordinary ETF holders had been heading for the exit in numbers, you would expect to see it in the fund that has the most of them, and IBIT was flat. A week in which two products did 86% of the selling looks more like a small number of large decisions than a crowd leaving. That is an inference from the shape of the table, not a fact about any holder, and the STACK below sets a test that can prove it wrong by Friday's close.

Then the bounce. Bitcoin closed Sunday at $83,678.29 on Coinbase, on UTC days, up 2.4% from Thursday's $81,692.79, a little under half of that on Friday and the rest over the weekend, when the ETF channel is shut. That recovers about 41% of the $4,814 it lost from the 4 October close. It was $83,513 at midnight New York time. We had no cause to offer for the fall on Friday and we have none for the bounce. The FILTER takes up the claim that it means the selling is over.

Three metrics.

Metric 1, Fear and Greed: 69. The reading stamped 12 October, which the index posts the evening before. Friday printed 59. Since then: 64, 61, 69. Ten points in three days on a move that got back less than half of what was lost. The gauge recovered faster than the price did, and the FILTER explains why that is less informative than it looks.

Metric 2, ETF flows: minus $678.9 million, 5 to 9 October, final. Daily: minus 89.8, plus 118.8, minus 484.9, minus 244.1, plus 21.1. Friday printed minus $694.5 million for the first four days with IBIT's Thursday missing. With it, those four days were minus $700.0 million. October to date, 1 to 9 October: minus $386.3 million, after two days of inflows at the start of the month.

Metric 3, Strategy's cushion: 10.7%. 848,000 BTC at an average cost of $75,440.70, per the 5 October filing, which is still the latest we have. Against $83,513 the stack sits 10.7% over cost, up from 8.6% on Friday. The company's weekly filing usually lands on a Monday, after this email has gone.

No tests grade this morning. The STACK sets one.

No new exchange or protocol hack that we can find in the window, no SEC or CFTC action, no Fed speeches between 9 and 11 October, and no movement on CLARITY since 15 September.

📣 THIS SPOT IS OPEN

Baseline Crypto reaches nearly 7,000 Bitcoin-focused readers, with an open rate of about 48% over the last four weeks. Direct sponsor placements are open, premium position, top of the issue, one per send.

Reply to this email for rates.

"Bitcoin is back above $83,000, Fear and Greed is back near 70, and the ETFs turned positive on Friday. The selling is done."

Maybe. But most of the bounce came on the two days the ETFs could not trade, it got back less than half, and Fidelity's fund was still negative on the good day.

Start with what is right in it. The price did recover, 2.4% from Thursday's close to Sunday's. The sentiment gauge did move, ten points. Friday's ETF figure was positive, plus $21.1 million, the first inflow since Tuesday. None of that is invented.

Now take each piece on its own terms.

The bounce. The selling we can see last week went through funds. Funds create and redeem shares on trading days, and Saturday and Sunday are not trading days. The channel that did the damage was shut for most of the recovery. A price that rises while the main seller cannot sell tells you very little about whether that seller has finished. The first real test is the flow figure for Monday, which is published after the US close.

Friday's inflow. Plus $21.1 million is the sum of IBIT at plus $22.4 million, HODL at plus $2.3 million and FBTC at minus $3.6 million. The fund that did most of the selling all week was still selling on the day the total turned positive. A small number, but still an outflow.

The gauge. The Fear and Greed index is not an independent reading of the market. On its own methodology page, half of the weight sits on volatility and on price momentum and volume, so a price bounce lifts the index almost by construction. Reading a rising gauge as confirmation of a rising price counts the same thing twice.

The size. Bitcoin is still nearly $3,000 below where it closed on 4 October, and October's ETF flows are still $386.3 million in the red for the month.

So what would actually show the selling is done? The price will not tell you, and the gauge copies the price. Watch the funds that did the selling. If FBTC and ARKB print zero or better for most of this week, the SIGNAL's reading survives: a few large decisions, apparently finished. If they keep printing outflows, it was not a few decisions and the weekend was a pause. That is checkable by Friday, and the STACK turns it into a test, so it can be graded from public numbers rather than taken on our word.

Three numbers that define the week.

$380.3 million, Fidelity's FBTC, 5 to 9 October. The single largest source of last week's selling, with outflows on four days and an inflow on none. It is also half of this week's test. Farside publishes each fund's figure after the US close, so the first reading comes tonight. One quiet day settles nothing. Three in a row would start to.

3.15%, high yield spreads on 8 October, and a deadline a week out. On 5 October we named the two exits for the credit question: a break above 3.46%, or a return under 3.00% inside a fortnight, which runs to the 19th. Spreads came in to 3.03% on 6 October and then went back out, 3.09% on the 7th and 3.15% on the 8th. That is further from the under-3.00% exit, not nearer. If neither exit has come by the 19th, the question did not resolve, and the window does not stretch to make it.

Wednesday 14 October, 8:30 a.m., September consumer prices. The BLS calendar puts CPI on Wednesday, producer prices on Thursday and import prices on Friday. Going in, the bond market's ten-year inflation expectation is 2.33%, down from 2.36% a week earlier, and the real yield was 2.87% on 8 October. The rise in long rates this autumn has been a rise in real yields, not in expected inflation. A hot print that leaves breakevens where they are is the market saying it is noise. And on the evidence of the last two weeks, even a print that moves rates may not move bitcoin, which is the thing we would most like to find out.

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 this morning, and Friday's open question is closed.

Friday was not a dated test, but it was a stated condition, so it gets the same treatment. Rule: if Thursday 8 October finished below minus $484.9 million once IBIT reported, the selling had a direction; if IBIT reported something small, Wednesday stood on its own. The number: IBIT minus $5.5 million, Thursday minus $244.1 million. The verdict: Wednesday stands on its own. The consequence is the SIGNAL above. With the full week in, the selling turns out to have been concentrated in two funds, which is a better question than the one we were asking.

A new dated test, set today, for 12 to 16 October.

The rule: the combined net flow of Fidelity's FBTC and ARK's ARKB from Monday 12 to Friday 16 October, on our sum of Farside's printed daily figures for those two funds. Branch A is zero or above. Branch B is below zero, down to and including minus $100 million. Branch C is below minus $100 million.

What our argument implies. The SIGNAL reads last week as a small number of large decisions that have now been made. If that is right, these two funds should be quiet this week, which is A or B. Branch C would mean the reading was wrong and the selling was not a few decisions but a steady leak, and that is what we would report. For scale, last week $587.5 million went out of the pair, an average of $117.5 million a day.

The board has no live rule. Nothing has been added since the October entry was retired on 5 October. A flow test is not a position, and we are not taking one on it.

Fed decision in October, 25 bps increase | Market: 16.5% ⚪ NO TRADE, no rule
No change is 83.5%, as of midnight Sunday New York time. It was 15.5% on Thursday evening, so it is back where it was on 5 October. A point either way in a market this thin is noise. What would move it: a voter saying October out loud before the 28th. Nobody has, that we can find.

Fed decision in December, 25 bps increase | Market: 74.5% ⚪ NO TRADE
No change is 22.5%, as of midnight Sunday. Down a point from Thursday's 75.5%. The hike is still the clear base case for December, and nothing in the window changed that. What would move it: the 28 October statement, which either sets December up or does not.

Zero Fed rate cuts in 2026 | Market: 95.55% ⚫ EXITED at 88.75%, a loss
As of midnight Sunday, unchanged from Thursday. Short from 78%, trimmed at 85%, covered at 88.75% on a published stop on 3 August, and recorded as a loss in every issue since. What would move it: a dot below the current midpoint in the December projections.

CLARITY Act signed into law in 2026 | Market: 5.05% ⚪ NO TRADE, nothing to trade on
As of midnight Sunday, unchanged from Thursday. The re-entry rule here was retired in print on 16 September. What would move it: a markup of the substitute text, and a new rule written before the fact.

The book. Flat for the twentieth consecutive issue, recomputed from the archive: nineteen issues published after the 3 August exits, so this is the twentieth. The test above is the first one in a while aimed at bitcoin's own buyers and sellers rather than at the rates around it. After two weeks in which the rates were right and the asset ignored them, that is where the evidence is.

Track everything live at polymarket.com, free, no account required.

On Friday we asked whether bitcoin's break from the macro picture was noise that fades or a shift that lasts.

Nothing came back. Our file of reader replies has nothing new since 3 August. We are still asking, and we will say so when nobody answers.

This issue is about where the money sits. Last week, the selling we can see happened inside two funds, and most of the bounce came on days the funds were shut.

One question: where does your bitcoin live? Reply with one word: fund, exchange or keys.

Nobody is asking for amounts, and nobody will be named. We want to know which of last week's channels our own readers are actually in.

Hit reply. We read every response.

The Baseline Crypto Team

Two tools, both free, no account needed.

The hardware wallet checker takes your device and firmware version and tells you whether you sit inside a population with a known problem, and what to do if you do. The custody audit is nine questions about what happens to your coins when something goes wrong, and it is deliberately unkind about backups and inheritance.

The $679 million that left the ETFs last week left through machinery that runs every trading day. Holders sold shares, and the funds' custodians released the bitcoin as market makers redeemed them. Whatever you think of funds, that process gets exercised daily.

Holding your own keys replaces all of that machinery with you. Nobody exercises your backup unless you do. If you have never restored a wallet from the words you wrote down, the most important part of your setup has never been tried. The custody audit asks about that directly, along with what happens if you are not the one doing the restoring.

Send the custody audit to the friend who moved their coins off an exchange this year and has never tried recovering the wallet from the words they wrote down. They did the hard part. This is the part that tells them whether it worked.