Home › Market › Daily brief

Daily Brief — October 2, 2026: Price made the high, the crowd went to a coin flip

By TradingPrimer · Published 2 Oct 2026

A new high is normally the easy part of a brief to explain: price went up, so somebody bought. Today the records of who did the buying all point somewhere else. The spot ETF channel was a net seller in each of the last two reported sessions. The share of leveraged accounts holding a long fell to an even split, most of it inside four hours on the way up. Funding has not reached its resting rate since 22 September, and the perpetual is still trading below the index it tracks. The book itself grew — so positions were opened — but the crowd that opened them stopped leaning long while doing it. None of this forecasts the next move. It does change what this high is evidence of, which is less than it looks.

Where the market actually is

Read at 03:01 UTC on 2 October, from the Binance spot and perpetual APIs:

All three were reading within 0.09% of their own 24-hour highs when the snapshot was taken — Bitcoin sat on its high to the cent. Turnover over the same 24 hours was $1.478 billion on BTC/USDT, $718.4 million on ETH/USDT and $263.1 million on SOL/USDT.

The brief of 28 September flagged a market that had compressed to a 0.81% daily range on $678 million of turnover, and warned that a position sized for the week before it would be far too large. That compression has now fully unwound in the other direction. The four completed sessions since run 2.95%, 2.16%, 3.25% and 2.51% — an average of 2.72%, against 1.02% for the weekend it replaced. Turnover on 1 October was $1.409 billion, roughly double that quiet Saturday. The range is back, and it came back in four days.

Why it moved

The macro half of the answer is clean and dates from Wednesday. Core PCE — the inflation measure the Federal Reserve actually targets — rose 3.0% year on year in August, against the 3.3% the market expected. Odds of a further increase at the 28 October meeting fell through a coin flip on the release: CME FedWatch was reported at 47.1% on 30 September, down from about 70% two days earlier, and at 38.2% in reporting that followed the full figure. That is the first genuine easing of pressure since the Committee raised its target range to 3.75%–4.00% on 16 September. The 10-year Treasury yield, around 5.17%, is still the competing offer that has sat on this market all quarter.

The other half does not cooperate. If softer inflation pulled money into Bitcoin, it did not come through the exchange-traded funds. Per Farside Investors data read on 2 October, the twelve US spot Bitcoin funds took $31.0 million on 28 September and $66.2 million on 29 September, then gave back $148.7 million on 30 September and $92.9 million on 1 October — a net outflow of $144.4 million across the four sessions. The week before that ran $2.39 billion of inflow. One caveat worth stating plainly: BlackRock's IBIT, much the largest of the twelve, had not reported its 1 October line when the table was read, so that day's total is provisional and could move either way.

So the two channels that usually explain a high are pointing opposite ways, and the honest reading is that whatever carried this one is not legible in either record. That is not a mystery to be solved with a guess. It is a limit on what today's chart can be used to claim.

The number most people skipped

Binance publishes the share of accounts on each side of its Bitcoin perpetual. On 1 October at 08:00 UTC, 58.76% of accounts held longs — in line with the 56.3% to 58.8% band of the previous two days. At the 03:00 reading on 2 October it was 50.38%. Half. A fall of 8.38 percentage points in nineteen hours, and most of it inside one stretch: 55.78% at 16:00 on 1 October, 51.11% by 20:00. Four hours, while price was rising.

Half is not an extreme, and it would be dishonest to sell it as one. Across the last thirty days — 180 four-hourly readings — 36 of them sat at or below today's 50.38%, and the month's low was 43.87% on 4 September. The level is ordinary. What is not ordinary is the company it is keeping: this reading arrived with a new high rather than after a fall, which is the opposite of how the crowd usually gets to an even split.

Over exactly the same window the book got bigger. Binance Bitcoin perpetual open interest bottomed at 92,474 BTC at 00:00 UTC on 30 September and read 97,029 BTC — 4,555 contracts added, up 4.9%; in dollars, $7.73 billion to $8.22 billion. Positions were opened, in size, by people who were on balance not choosing the long side.

The price of leverage says the same thing from a third angle. On Binance, perpetual funding has a resting rate of 0.0100% per eight hours — what you pay when the perpetual trades level with spot. Bitcoin has not reached it once in the last sixteen settlements charted below; the highest was 0.0080% and two printed negative. A $100,000 long held through all sixteen paid $61.50, against $160.00 at the resting rate. That gap is worth stating carefully, because the run is longer than the chart and far less remarkable than it sounds: funding has been below its resting rate for twenty-nine consecutive settlements, since 22 September — and across the last two hundred settlements the longest such stretch ran to fifty-three. On Binance the perpetual sits below its resting rate as the ordinary state of affairs. So the fact of a discount is not the signal; what it does while price makes a high is. When read, the mark price was $85,444.32 against an index of $85,477.39 — the perpetual $33 cheaper than the thing it tracks, at a 24-hour high. Aggressive flow agrees: of the last eight four-hour buckets, five had more market-order selling than buying, and the two covering the push itself read 1.0063 and 0.9983 — as near balanced as the measure gets.

Solana is the useful contrast, because it is the one major paying full price. Its funding was accruing at exactly 0.0100% — the resting rate, on the nose — and it is also the only one of the three whose book shrank, from 8,440,129 contracts at 20:00 on 30 September to 8,141,726, down 3.5%. Ether did neither: funding 0.00702%, open interest up 0.9% across the same window, a market that mostly sat still.

For the mechanics of each term from zero: open interest, funding rate and liquidity. Live readings sit on Market Pulse.

So what

The structural lesson is that a rising price and a growing book do not add up to "longs piling in" — you have to check, because the two readings are published separately and today they disagree. The reflex that open interest up plus price up means crowded longs is the single most common misreading of this data, and it would have told you the exact opposite of what the account ratio, the funding rate and the mark-to-index spread all said in the same hour. Three independent measures lined up against the reflex. That is the arrangement worth noticing, not the direction.

The practical consequence is the mirror image of the one in the 28 September brief, and it catches people the same way. Then, volatility had collapsed and the danger was a position sized for a wider market. Now the daily range has nearly tripled in four sessions, and the danger runs the other way: a stop set to last weekend's 0.81% day is inside the noise of a 2.72% one and will be taken out by ordinary movement. Stop distance follows volatility, and position size follows stop distance — so when the range changes, the size has to change with it to keep the money at risk constant. Position sizing covers that arithmetic, and leverage and margin covers what a multiplier does to it. Our live 4-hour BTC/USDT chart carries the current reading.

One last note on the shape of the book. It is 4.9% larger than it was on Tuesday, and that is fuel: a bigger book is more positions that can be forced out at once if a move goes against them, which is the mechanism described in liquidation cascades. That it is leaning less long than usual says which direction would hurt more, not which direction is coming. The event calendar has what is next; 28 October is the date that matters.

← Market analysis