Five endorsements and a nine point drop

The list is genuinely remarkable, and a year ago it would have been unthinkable. In late July the following firms went on record in favour of the CLARITY Act:

  • BlackRock, in a statement to Politico, called it a step toward a framework that "puts investors first".
  • Fidelity emphasised the need for "clear rules of the road".
  • Franklin Templeton said the bill would make clear how crypto is regulated and urged Congress to pass it.
  • Goldman Sachs chief executive David Solomon said the bill "is not perfect" but would create a level playing field.
  • SoFi's chief executive called durable rules critical for competitiveness.

Now put the price of the bill next to the praise. The Polymarket contract on the CLARITY Act being signed into law in 2026 sits near 34%. It was 38% a day before, and 43% a week before. The endorsements went up and the odds went down.

That is not a contradiction, and understanding why is the most useful thing on this page. The bill has already passed the House and already cleared the Senate Banking Committee. Its problem was never a shortage of support, a drafting flaw, or industry opposition. Its problem is calendar. There is no scheduled floor vote, the Senate departs for its August recess on the 8th, and August 10 begins the state work period. If it does not move this week it waits until mid September.

None of those five firms controls a minute of Senate floor time. The person who does is the majority leader, and as we noted on July 28, John Thune has said the chamber will prioritise a Russia sanctions bill and is unlikely to advance CLARITY before it. A week of increasingly prominent endorsements has not changed that, and the odds have been marking it down accordingly.

The weekend reclaim, and the hour it turned

Friday ended with Bitcoin down almost 3% and $62,500 broken, a level this site had been naming since Tuesday. The weekend undid it, and the turn is locatable to a single hour.

Saturday drifted lower through the European session and then the selling concentrated. The 18:00 UTC hour carried $58.2 million, by far the largest of the weekend, at a 41.1% buy share. Inside it, the 18:30 candle printed the low of $62,275 on $18.3 million at a 34.5% buy share. That is a genuine flush: heavy, one sided, and it took out Friday's $62,466 low.

And then it stopped, immediately. The very next candle at 18:45 held $62,280 on a 58.1% buy share. By 18:55 price closed back above $62,500 with a 57.0% buy share, and the 19:00 hour aggregated to 59.8%. Saturday finished at $62,824, down just 0.10% on the day, having been 0.9% lower intraday.

Sunday extended it. The 02:00 hour rose 0.75% on $67.7 million at 58.3%, the largest hour of the weekend, and the 07:00 hour printed a 76.2% buy share on $26.2 million. Bitcoin trades near $63,178, up about 0.15% over 24 hours, with a Sunday buy share of 55.1% and a 24 hour aggregate of 52.8%.

So $62,500 is back to being support rather than resistance, which is the friendlier of the two outcomes we set out on Friday.

Our conditions contradicted each other again, and the fix we proposed did not work

This needs writing before anything else is claimed from the weekend, because it is the second consecutive instance and that makes it a pattern rather than bad luck.

Friday's article named two things to watch. That a lower low on a sub 45% hourly reading would be the continuation case. And that reclaiming $62,500 on an hourly buy share above 55% would mean Friday was a hedge unwind and nothing more. Both fired. They fired one hour apart, in opposite directions.

Hour (UTC)Condition triggeredVolumeBuy share
Aug 1, 18:00Lower low at $62,275 on a sub 45% reading$58.2M41.1%
Aug 1, 19:00Reclaimed $62,500 above a 55% reading$19.6M59.8%

On Friday, when the same thing happened, we diagnosed it as a resolution problem. The bullish signal that failed had come from a five minute window, so we said we would use sustained hourly readings instead. This weekend both readings were hourly, on real volume, and they still contradicted each other inside sixty minutes. The proposed fix did not fix it, and saying so is more useful than quietly moving on.

The better diagnosis is that the timeframe was never the problem. The buy share is a description of the hour it measures, not a forecast of the next one. The 18:00 reading of 41.1% was completely accurate: sellers were pushing and they made a new low. The 19:00 reading of 59.8% was also completely accurate: buyers absorbed that low and took the level back. Neither reading was wrong. The error was ours, in dressing them up as directional confirmations, because a level break is exactly the moment when control changes hands, so any threshold set near it will be crossed in both directions.

What the metric is actually good for is telling you who is acting right now, which is genuinely valuable and is not the same as telling you what happens next. From here we will report it that way and stop attaching if-then predictions to individual level crossings.

Bitcoin is the only thing being bought

Friday's session had every asset on the board below a 50% buy share, gold included, which is what told us it was a de-risking rather than a rotation. The weekend has changed that in exactly one place.

PairTaker buy share, 24h24h change
Bitcoin52.8%0.15% higher
XRP47.6%1.78% higher
BNB47.5%0.57% higher
Ethereum45.2%0.03% lower
Paxos Gold44.1%0.35% higher
Solana43.4%0.52% higher
Tether Gold38.2%0.17% higher

Bitcoin is the only pair above 50%. Several of the others are higher on price while their aggressive flow is net selling, XRP most notably at 1.78% up on a 47.6% buy share, which is drift on thin books rather than demand.

And gold is the weakest thing on the list. Tether Gold at 38.2% is the lowest reading we have recorded for it in this stretch, down from 40.3% on Friday, 44.9% on Thursday and 67.9% last Tuesday. The hedge unwind that explained Friday has not reversed at all. It has continued, quietly, through a weekend when Bitcoin recovered.

Sentiment is inching back. The Fear and Greed index reads 27, with Friday's crash having taken it to 25, the only Extreme Fear print of the stretch, from 28, 29 and 29 before that. Total crypto market capitalisation is $2.251 trillion, up 0.20%, with Bitcoin dominance at 56.3%.

The caveat that outweighs the recovery

Everything above should be read against one number, and it is not flattering. Participation has drained away.

SessionVolumeBuy shareChange
Fri Jul 31$1.30B47.2%2.92% lower
Sat Aug 1$0.48B49.9%0.10% lower
Sun Aug 2, so far$0.28B55.1%0.56% higher

Friday was the heaviest session of the week. Saturday came in 63% lighter. Sunday so far is $0.28 billion, which is not just thin for a weekend, it is about 45% below the previous Sunday's $0.51 billion. This is the emptiest tape in the sample.

A buy share is a ratio, so it stays perfectly readable on low volume. But it takes far less money to move it, which means a 55.1% Sunday on $0.28 billion carries much weaker evidence of conviction than the same number would on a Wednesday. The reclaim of $62,500 is real and it happened on decent hourly readings. It has simply not yet been tested by anyone who trades size.

Saturday deserves a note of its own for the opposite reason. Its buy share was 49.9%, which is as close to a perfect coin flip as this metric produces, and the session moved 0.10% lower. Balanced flow in, nothing out. It is a small and rather elegant demonstration that the number is measuring something real: when neither side has the initiative, price does not go anywhere.

July's ETF book closed worse than it read mid week

On Friday we flagged the final July ETF number as something the session would settle. It settled badly.

Spot Bitcoin ETFs finished July with roughly $172.4 million of net inflows, after about $265.4 million left on the final trading day. That final day is precisely the session in which Bitcoin fell almost 3% and lost $62,500, so the two things are one thing. Mid week the running total was near $205 million, and Friday alone took a third of it back out.

The figure needs its context in both directions, as it did before. July snapped a two month outflow streak, following roughly $2.43 billion out in May and $4.52 billion out in June, so a positive month is a real improvement. It is also the smallest positive month on record. Both remain true.

The more interesting line is the comparison. Ether ETFs took in about $365.2 million across the same month, extending their run to four consecutive weeks of inflows, and that is more than double the Bitcoin total. The rotation inside crypto that we first flagged on Thursday now has a full month of finalised data behind it rather than a partial read.

What to watch this week

  • Any CLARITY floor vote being scheduled, at all. This is the binary. The Senate leaves August 8 and August 10 begins the state work period. A calendar announcement is worth more to price than another endorsement, and the absence of one through midweek effectively settles it until mid September.
  • The Polymarket contract itself. It has moved 43 to 38 to 34 in a week, which makes it the fastest reacting instrument on this story. Watch the direction rather than the level.
  • Whether $62,500 holds on weekday volume. The reclaim happened on the thinnest tape of the month. Monday is the first real test. Losing it again on above average volume would say the weekend was an absence of sellers rather than a presence of buyers.
  • $62,275 below, $63,634 above. Saturday's low and Sunday's high are the weekend's boundaries, and the range is only about $1,360 wide.
  • Tether Gold at 38.2%. The hedge unwind is still running. If gold's buy share keeps falling while Bitcoin's holds above 50%, that argues the money leaving the inflation trade is partly landing in Bitcoin rather than exiting entirely.
  • August ETF flows from day one. July closed at $172.4 million for Bitcoin against $365.2 million for Ether. Whether that gap narrows in the first week of August tells you if the rotation was a July artefact or a trend.

How to read this as a trader

Two things from this weekend are worth carrying into the week, and one of them is uncomfortable.

The comfortable one is about the CLARITY story, and it generalises well beyond this bill. Endorsement is not the binding constraint, scheduling is. When BlackRock and Goldman Sachs publicly back a bill and its odds fall nine points in a week, the market is telling you it has already priced industry support and is now pricing only floor time. The tradeable variable is the calendar, which is controlled by one person, and it has been pointing away from crypto since July 28. If you are watching headlines for encouragement you will keep being encouraged while the probability keeps falling.

The uncomfortable one is about us. For two consecutive publications, conditions we set at a level fired in both directions within an hour, and the fix we proposed after the first instance did not prevent the second. The honest conclusion is that the buy share describes the hour it measures and does not forecast the next one, and level crossings are the worst possible place to ask it for a prediction, because that is definitionally where control is changing hands. That does not weaken the metric for its actual purpose, which is knowing who is acting in size while it is happening. It does mean we were using it for the wrong job, and we have stopped.

Practically, that leaves this week with a clean setup and no forecast attached: Bitcoin reclaimed a level on the thinnest volume of the month, the only asset on the board attracting aggressive buying, into a week where the single biggest scheduled catalyst is whether one senator puts a bill on the calendar. Our live dashboard separates aggressive buying from aggressive selling minute by minute across the most liquid pairs, which is how the 18:30 flush and the 18:55 reclaim were both visible inside half an hour. For the arc behind this, see Friday's break and the 49.5% July, Thursday's reclaim and the pre-release bid, the Fed's 9 to 3 hold, and the Tuesday that first named Thune and the CLARITY delay. If the buy and sell split is new to you, the volume guide covers the mechanics.