Three sessions, in order

SessionRangeVolumeCloseBuy share
Fri Aug 285.97%$1,560.9 million3.00% lower47.1%
Sat Aug 291.22%$545.9 million+0.49%51.2%
Sun Aug 30, partial0.44%$128.3 million0.22% lower52.5%

The range fell by a factor of about five from Friday to Saturday, and volume by about two thirds. Bitcoin currently trades near $78,060, roughly $3,400 below Friday's high and almost exactly where Friday closed.

Where those figures sit historically is the part worth knowing.

Percentile of 999 sessionsDaily rangeDaily volume
1st0.65%$408 million
5th1.15%$579 million
10th1.55%$758 million
25th2.13%$1,152 million
Median3.31%$1,663 million
90th6.59%

Saturday's 1.22% range lands in the 5.8th percentile: only 58 of 999 completed sessions were quieter. Its $545.9 million of volume lands in the 3.7th percentile. Friday's 5.97% range, for contrast, was in the 86.9th.

So the week ended with a session near the top decile for activity followed immediately by one near the bottom twentieth.

The arithmetic of a quiet weekend

On August 22 we published a measurement that surprised us and, judging by the response, surprised readers: weekend depth is not thinner. Measuring dollars of hourly volume per 1% of price movement, weekend readings came out at 87% to 109% of weekday depending on the asset. We used that to reject the standard explanation for a violent Saturday.

That was correct and it was incomplete, and a reader could reasonably have taken it to mean weekends are not quieter at all. They are. Here is the missing half.

Measure, median of 999 sessionsWeekdayWeekendWeekend as % of weekday
Daily range3.72%2.04%55%
Daily volume$1,928 million$953 million49%
Volume per 1% of moveBaselineMeasured Aug 2287% to 109%

Read the three rows together and the picture resolves. Weekend volume is roughly half. Weekend range is roughly half. Divide one by the other and you get approximately the same number as a weekday, which is exactly the flat depth reading we reported eight days ago.

So all three measurements are true simultaneously and each answers a different question. How much trades? Half. How far does price travel? Half. How many dollars does it take to move price one percent? About the same. The last of those is the one that governs how violent an individual move can be, which is why an 8% hour in XRP on a Saturday was not explained by thin books.

This is the same lesson as the taker buy share guide we published on Thursday, arriving from a different direction. A ratio and its two components carry different information, and confusing them produces confident wrong answers.

A rare sequence, and what we will not do with it

A session with a range of 5% or more followed immediately by one with a range of 1.5% or less has occurred seven times in 999 sessions, this weekend included.

Loud sessionIts rangeNext session's range
June 7, 20245.23%0.60%
March 14, 20255.56%1.27%
October 17, 20255.52%1.11%
December 19, 20255.04%0.89%
March 13, 20265.01%1.41%
August 28, 20265.97%1.22%

Now the part that matters more than the table. We are not going to tell you what happened after those six prior instances, and we are not going to look for a pattern in them.

Six observations cannot establish what typically follows anything. This site published five falling gold buy share readings on August 3 and called them the clearest available read on an ongoing unwind. The pattern reversed the next day. We have cited that error in four separate articles this month because it is the exact temptation this table presents.

What the table does support is a narrow claim: this particular sequence is uncommon, occurring in well under 1% of sessions. That is a fact about how unusual the last two days were, not a hint about tomorrow.

What Friday actually closed at

We published Friday's article at 16:20 UTC with the session still open, and labelled every figure from it partial. Here is what the completed session turned out to be, because a label is only worth something if you come back and settle it.

FigurePublished mid sessionFinal
Session change2.58% lower3.00% lower
Session low$77,944$76,888
Range4.53%5.97%
Taker buy share47.2%47.1%

The price kept going after we published, another $1,056 lower, and the range ended nearly a point and a half wider than the figure we printed. Anyone treating our partial numbers as final would have understated the day.

The last row is the interesting one. While the price moved that much further, the taker buy share barely shifted at all, from 47.2% to 47.1%. That is the bounded ratio behaving exactly as Thursday's guide described: the day got materially worse and the metric moved a tenth of a point.

Positioning into the quiet

Three readings, and they do not all point the same way.

MeasureReadingNote
Long/short accounts1.1259The most long skewed of this run
Perp open interest, 30h2.22% higher106,103 to 108,454 BTC
Funding, last four settlements0.0100%, 0.0100%, 0.0083%, 0.0100%Back at the venue ceiling

The long/short ratio has moved from below parity during Friday's decline to 1.1259 now, which is the most long skewed reading we have recorded across this entire run. Open interest has climbed 2.22% over thirty hours. Funding has returned to the 0.0100% ceiling after coming off it mid week.

So positions have been added, and they have been added on the long side, during the two quietest sessions of the month. We are reporting that arrangement and not forecasting from it, which is the same discipline we applied on Monday when the ratio was below parity and no squeeze followed.

The ETF streak ended on the same day

US spot Bitcoin ETFs recorded roughly $201.9 million of net outflows on August 28, ending a run of nine consecutive inflow sessions that had brought about $3.04 billion into the funds since August 17.

Two things are worth holding at once. The streak broke on the same session as the third rejection above $80,000 and the deepest give back of the week, which is not a coincidence worth much on its own but is worth noting. And the outflow amounts to roughly 6.6% of what the streak brought in, so a nine day accumulation was not undone by one day of selling.

There are no flow figures for Saturday or Sunday. ETFs do not trade at weekends, which is one of the genuine mechanical reasons weekend volume runs at half of weekday levels.

Where things stand

AssetPrice24h24h range
Bitcoin$78,060+0.66%$77,507 to $78,330
Ether$2,455.74+0.91%$2,431.26 to $2,467.93
XRP$1.3906+0.66%$1.3793 to $1.4039
Solana$104.89+1.49%$103.03 to $105.88
BNB$692.78+0.67%$687.39 to $695.56
Dogecoin$0.08458+0.19%$0.08418 to $0.08571
Tokenised gold$4,463+0.10%$4,455 to $4,467

Every major is up between 0.19% and 1.49% and every 24 hour range is narrow. Total market capitalisation is $2.631 trillion, down 1.89% over 24 hours, with Bitcoin dominance at 59.48%, the highest of this run. The price sits 12.57% above the 200-day moving average of $69,347.

Fear and Greed reads 69, Greed, easing from 73 on Friday. The ETH to BTC ratio is 0.03146, up 0.22%, the eighth day we have reported it. Up, up, down, up, down, up, up, up. We have declined to call it a trend eight times and we decline again.

What we are not saying

We are not saying quiet precedes anything. That is the whole point of the section above: six prior instances cannot tell you what follows a seventh.

We are not saying the long positioning is right or wrong. It is a fact about how accounts are arranged this morning.

We are not saying the ETF outflow marks a turn. It is one session and 6.6% of what nine sessions brought in.

We are not saying weekends are safe because depth holds up. Half the volume and half the range is a real reduction in activity, and our August 22 measurement was about a ratio, not about calm.

What we are saying is that the last two sessions were unusually quiet by a measurable amount, that the arithmetic behind a quiet weekend resolves an apparent contradiction in our own coverage, and that Friday closed considerably lower than the partial figures we published while it was still trading. All of it is reproducible from public endpoints.