Bitcoin Daily Brief — September 10, 2026: The Squeeze Breaks Downward as a PPI Overshoot Takes BTC to $76,537

Bitcoin price on September 10, 2026: BTC closed at $76,537, down 2.23%, in a $2,114 range between a $78,554 high and a $76,440 low — the largest daily decline since August 28, a fourth consecutive lower close, and the session in which the eleven-day Bollinger squeeze this brief has tracked all week finally expanded, taking %B to 0.07 and leaving the close $317 above the lower band after an August producer-price reading pushed thirty-year Treasury yields to a nineteen-year high. The snapshot and charts are first, the short version follows, and the full explanations are at the end for anyone who wants them.

New to the indicators? Every name below links to the plain-English explanations in Bitcoin Indicators Explained. Yesterday’s edition: September 9, 2026. About this series: Bitcoin Daily Brief.

Bitcoin price on September 10, 2026, at a glance

The charts

The 30-second read

Bitcoin fell 2.23% on Thursday to close at $76,537 — a $1,747 decline, the largest since August 28 and a fourth consecutive lower close — after an August producer-price report put wholesale inflation at 5.4% year over year, sent the thirty-year Treasury yield to 5.367% and the ten-year to 4.954%, both fifty-two-week highs, pushed Brent crude to $105.37 and dragged the S&P 500 down 0.58% to 7,591.79 in a session that produced $562 million of crypto liquidations. Underneath the price, the squeeze resolved: Bollinger band width rose for the first time in twelve sessions, from 5.57% to 6.07%, while %B collapsed from 0.41 to 0.07 — its lowest since June 30 — leaving the close $317 above a lower band at $76,220 that sits within a dollar of the September 2 monthly low, and the trend layer intact but narrowing, with a twenty-third straight close above the 200-day average and a golden-cross gap that widened again to $437. The calendar decides the next move rather than the chart: August consumer prices land Friday morning, and the FOMC meets on September 15–16 with CME futures pricing close to a 70% chance of a rate increase.

What happened in the last 24 hours

• Price. Opened $78,284, reached $78,554, fell to $76,440 and closed $76,537 (−2.23%, −$1,747) — a $2,114 range on volume of 6,391 BTC, 0.99 times the twenty-day average; the candle body took up 82.7% of the range, the eleventh-largest share of 2026’s 253 sessions, and the close finished less than 5% of the way up from the low.
• Flows. The two-day reporting gap cleared and both missing sessions were redemptions: −$46.6 million on September 8 and −$120.2 million on September 9, a two-day total of −$166.8 million on SoSoValue and Farside data, with ARKB at −$78.0 million and GBTC at −$27.2 million of Wednesday’s figure. September’s six reported sessions now stand at +$603 million; Thursday’s own total was not yet reported when this brief was written.
• Positioning. Bitcoin.com News reported $562 million of crypto liquidations across the twenty-four hours, the large majority of them long positions, with about $74 million of long liquidations in the four hours after the PPI release and $112 million of Bitcoin long liquidations over the full day. CoinGlass displayed open interest of $53.32 billion; its funding-rate and long/short fields did not load when the page was read.
• On-chain. CryptoSlate, citing Glassnode’s September 9 report, put the sell-side risk ratio at 7 basis points per day on a seven-day basis against an August peak of 16 — a 56% decline — with long-term holders now contributing 47% of realized profit against 88% at the August peak, while about 1.07 million BTC acquired between $83,000 and $86,000 sat largely unmoved for thirty days and spot cumulative volume delta stayed negative through September 8.
• Macro and policy. August producer prices rose 0.4% on the month, in line with expectations, but 5.4% on the year against 4.8% in July; the S&P 500 closed at 7,591.79 (−0.58%), the Nasdaq at 26,081.73 (−0.65%), the Dow at 52,064.46 (−0.60%) and the Russell 2000 at 2,891.49 (−1.02%), with the VIX up 8.69% to 17.89, WTI up 4.2% to $100.10, Brent up 3.6% to $105.37 and gold at $4,427.40; the CFTC separately authorized US institutional access to Singapore Exchange Bitcoin perpetual futures.

Section summaries

Price and trend

Thursday was the first decisive candle in two weeks: a $1,747.43 body inside a $2,114.18 range, with only a $270 upper wick and a $97 lower one, closing 4.6% above the session low. That is the mirror image of Wednesday’s near-doji, and it took price $2,066.96 below its 20-day average for a second straight close under that line, while the average itself fell for the first time since August 16, ending twenty-four consecutive rises. The long-frame structure was untouched: a twenty-third consecutive close above the 200-day average, 9.39% above it, and a golden-cross gap that widened from $272.36 to $436.77. Details ↓

Momentum and volatility

RSI fell 6.68 points to 52.77, a fourth consecutive decline and the lowest reading since August 16, while the MACD histogram widened a seventh day to −678.69 — the most negative it has been since June 10 — as the MACD line fell $348 to 2,266.87. ADX eased a third session to 45.66 with the directional spread narrowing a fourth time, from 18.31 to 12.97. The volatility story is the one that turned: band width rose for the first time in twelve sessions, from 5.574% to 6.065%, and %B collapsed from 0.4068 to 0.0664, its lowest since June 30. Details ↓

Flows and positioning

Five editions without a new ETF figure ended with two, and both were negative: −$46.6 million on September 8 and −$120.2 million on September 9, a two-day −$166.8 million that Crypto Briefing and Crypto Economy put at roughly 4.4% of the $3.8 billion gathered over the preceding three weeks. ARKB led Wednesday’s redemptions at −$78.0 million, GBTC followed at −$27.2 million and IBIT at −$19.5 million, with Morgan Stanley’s MSBT the only fund taking money in. On the derivatives side, Thursday’s move produced $562 million of crypto liquidations, heavily long — an order of magnitude above the $33 million reported for Tuesday’s session. Details ↓

On-chain and network

The day’s on-chain reading was sell-side risk — a measure of how much realized profit and loss is being locked in relative to the size of the network’s stored wealth — which Glassnode put at 7 basis points per day against an August peak of 16, with long-term holders’ share of realized profit down from 88% to 47%. The same analysis names the offsetting fact: roughly 1.07 million BTC bought between $83,000 and $86,000 that has not moved in thirty days. On the network, difficulty is unchanged at 127.45T after a +1.31% retarget at block 965,664 on September 6, with the next boundary at 967,680. Details ↓

Sentiment

alternative.me’s dated API could not be reached for a second consecutive run, so CoinGabbar’s September 10 mark of 69 (“Greed”) is used and is not compared with yesterday’s differently constructed 70. The Alpha Vantage news feed carried 15 Bitcoin-tagged articles, down sharply from 28 on Wednesday, scored none bullish, two somewhat bullish, eleven neutral, two somewhat bearish and none bearish. Published views split on horizon rather than on the day: Coinbase’s Brian Armstrong said he still regards $400,000 by 2030 as reasonable and that the bottom is in, while Ted Pillows named the 50-week EMA near $77,000 as the level he was watching into the inflation prints. Details ↓

Macro and cross-asset

Thursday was an inflation-and-energy session. August producer prices came in at +0.4% on the month, as forecast, but 5.4% on the year against 4.8% in July; WTI rose 4.2% to $100.10 and Brent 3.6% to $105.37; the thirty-year Treasury yield added 8.1 basis points to 5.367% and the ten-year 11.4 to 4.954%, both fifty-two-week highs, with Cointelegraph describing the long bond’s level as a nineteen-year high. Equities fell for a fourth straight day and the VIX rose 8.69% to 17.89. Bitcoin’s 2.23% decline was, unusually for this run, larger than any of the equity indices’. Details ↓

Calendar

Friday September 11: August consumer prices at 8:30 a.m. ET, the report that both the equity and crypto coverage has spent the week pointing at, with the September 10 ETF flow print due the same morning. Sunday September 13: the weekly candle closes, with the 50-week average near $79,630. September 15–16: the FOMC, decision Wednesday at 2 p.m. ET, and the Senate cloture vote on the CLARITY Act. September 16: August retail sales. September 18: the September options expiry and the Bank of Japan decision. Around September 19: the next difficulty adjustment at block 967,680. Details ↓

Reading the combination

Yesterday’s edition said Thursday would be the first band-width reading in eleven sessions that carried information about the market rather than the calendar. It carried information: the band expanded, and it expanded because price moved, not because an outlier left the window. The direction the compression broke was down, and the four layers now disagree more sharply than at any point in this run — trend still improving (twenty-third close above the 200-day, golden-cross gap widening to $437), momentum deteriorating on every measure (RSI, MACD, ADX spread), flows turning negative for two reported sessions, and macro actively hostile with yields at fifty-two-week highs into a Federal Reserve meeting that futures price as more likely than not to be a hike. The bracket is unusually tight beneath: the lower band at $76,220 sits within a dollar of the September 2 low at $76,219, and Thursday’s own low at $76,440 stopped $220 above both. The indicators do not say which way it resolves. Details ↓

Learn the indicators: every reading above is explained, with charts and the combinations traders watch, in Bitcoin Indicators Explained.

The details, for readers who want more

Signal board

Indicator	Reading (Sep 10 close)	Zone	Change since Sep 9	What this zone has historically meant
Bollinger Bands (20, 2)	%B 0.07; band width 6.07%	Squeeze released — first expansion in twelve sessions	%B 0.41 → 0.07; width 5.57% → 6.07%	The eleven-session contraction that began on August 29 ended here, with band width rising 0.491 points. The mechanism is the one this brief set out yesterday: the twenty-day window had finished expelling the August breakout, so from this session a move in band width reflects the market. The move was a fall in price, not a rise, and it showed up in %B rather than in the width: %B dropped from 0.4068 to 0.0664, the lowest since June 30, the day of 2026’s lowest close. %B measures where the close sits between the bands, 0 being the lower band and 1 the upper; 0.07 means the close finished $316.54 above a lower band at $76,220.01 and $4,450.46 below an upper band at $80,987.01. Two base rates, both struck on this series and both carrying the same methodological note as always — the instance count is the number of earlier days matching the filter, while each percentage is struck on the subset with the full forward window available, so the two differ slightly. For the state: of 203 earlier days with %B below 0.10 and price above its 200-day average, the close was higher ten sessions later 58% of the time (median +0.99%) and twenty sessions later 53% (median +0.65%) — against an all-day baseline of 56% at both horizons. For the release: of 155 earlier days on which band width expanded from below 8% with price above the 200-day, the close was higher ten sessions later 65% of the time (median +1.69%) and twenty sessions later 67% (median +5.82%). Neither is a strong signal, and neither speaks to direction.
Price vs 20-day SMA	−2.63% below ($78,604)	Below, second consecutive close	−0.52% → −2.63%	The close finished $2,066.96 under the 20-day average, five times Wednesday’s $408.98 gap, and this time price did all the work: the average fell $89.45 while price fell $1,747.43. That fall is itself a first — the 20-day had risen on twenty-four consecutive sessions since August 17, and this is its first decline since August 16. A short average turns when the close entering its window is smaller than the one leaving it; Thursday’s $76,537 replaced August 21’s $78,326, so the turn was narrow and would have happened on a much smaller decline. Of 940 earlier days in this series with the close below a falling 20-day average while price held above its 200-day, the close was higher ten sessions later 52% of the time (median +0.21%) and twenty sessions later 52% (median +0.39%) — which is to say, nothing at all.
50-day vs 200-day SMA	$70,401 vs $69,964 (gap +$436.77)	Golden cross — 50-day above, third day	Gap +$272 → +$437	The crossing kept widening on a 2.23% down day, which is the clearest illustration available of how slowly these lines move. The 50-day rose $209.00 to $70,400.82 — a thirtieth consecutive daily rise — while the 200-day rose $44.59 to $69,964.06, adding $164.41 to the gap. The 50-day window is still shedding the low closes of late July, which are far below anything the 200-day window is shedding, so the shorter average gains ground even as price falls. The base rates are unchanged from the last two editions because they are struck on the cross date, not on subsequent days: of the thirteen earlier golden crosses in this series since the 200-day average first became computable in February 2011, price was higher thirty sessions later 50% of the time (median −2.62%), sixty sessions later 67% (median +12.60%) and ninety sessions later 67% (median +10.67%), on twelve measurable instances.
Price vs 200-day SMA	$76,537 vs $69,964 (+9.39%)	Above (23rd straight close, since Aug 19)	+11.96% → +9.39%	Price above the 200-day average is the most common definition of a long-term uptrend, and this is a twenty-third consecutive close above it. The distance narrowed by 2.57 percentage points, the largest single-day compression of the run, and this time price did essentially all of it: the 200-day rose $44.59 while price fell $1,747.43. The 200-day has now risen on each of the last twenty sessions, since August 22. At $6,572.49 the cushion is still substantial — it would take a further 8.59% decline from Thursday’s close to reach the line — but it is the narrowest it has been since August 20.
RSI (14)	52.77	Mid neutral zone (30–70)	59.45 → 52.77	Down 6.68 points, a fourth consecutive decline and by far the largest of the four, the sequence reading −5.17, −2.55, −0.66 and −6.68. This is the lowest reading since August 16 and the first below 55 since August 17. RSI compares the average size of up-days with down-days over fourteen sessions, so a 2.23% decline registers properly where Wednesday’s 0.21% did not; the reading has now given back the whole of the August breakout’s overbought excess, having peaked at 86 on August 21. Of 662 earlier days in this series with RSI between 50 and 56 and price above its 200-day average, the close was higher ten sessions later 49% of the time (median −0.31%) and twenty sessions later 48% (median −0.95%) — the only genuinely flat-to-negative base rate on this board, and still within a few points of a coin flip.
MACD (12, 26, 9)	2,267 vs signal 2,946; histogram −678.69	Below signal, still far above zero	Histogram −500.09 → −678.69 (widening)	A seventh day below the signal line and a seventh consecutive widening, the sequence now reading −49.50, −108.96, −139.48, −261.93, −396.55, −500.09, −678.69. The MACD line fell $348.27 while the signal line eased $169.67. At −678.69 the histogram is the most negative it has been since June 10, when price was near $61,449 — but the MACD line itself remains 2,267 above zero, so this still describes a loss of upward momentum rather than a change of trend direction. The line would have to fall a further 2,267 points to signal the latter. Of 638 earlier days in this series with a negative and widening histogram, the MACD line above zero and price above its 200-day average, the close was higher ten sessions later 55% of the time (median +0.84%) and twenty sessions later 53% (median +0.80%).
ADX (14)	45.66 (+DI 28.5, −DI 15.5)	Strongly trending (above 25), falling for a third day	47.22 → 46.90 → 45.66	A third consecutive fall, of 1.24 points — nearly four times the size of Wednesday’s 0.32 — with the internals narrowing faster: +DI eased from 30.54 to 28.48 while −DI rose from 12.23 to 15.51, taking the spread from 18.31 to 12.97 points. That is a fourth straight narrowing, and −DI has now risen on three of those four sessions while +DI has fallen on all four. Because ADX is a fourteen-period smoothing of that spread, it lags it; a falling ADX from a high level describes a strong trend losing intensity, not a trend reversing. Days in this series with ADX at or above 45, +DI above −DI and ADX falling (255 earlier instances) were followed by a higher close ten sessions later 71% of the time (median +4.70%) and twenty sessions later 70% (median +5.60%) — the strongest reading on this board, with the standing caveat that such instances cluster in a handful of strongly trending years and the sample is therefore not 255 independent observations.
ATR (14)	$2,240 (2.9% of price)	Falling, sixth straight day	$2,250 → $2,240	A sixth consecutive fall, of $9.70 — the smallest of the six — on a high-to-low range of $2,114.18 that was 0.940 times Wednesday’s ATR and 2.76% of the closing price. That range ranks 122nd narrowest of 2026’s 253 sessions by dollars and 112th by percentage of the close, so Thursday was a slightly-narrower-than-median day by range even though it was a decisively directional one by close. This is the distinction ATR cannot see: it measures how far price travelled, not where it ended. The $2,114 range and the $1,747 body say the same thing from two directions — the market moved less than usual but kept almost all of what it moved.
30-day realized volatility	47.7% annualized	Elevated; highest since April 2	46.75% → 47.67%	Up 0.92 points, and this is the measure that has been right about the squeeze all along. Over the eleven sessions in which band width fell from 40.03% to 5.57%, the actual dispersion of daily returns over thirty days went from 44.03% to 46.75% — it rose. It has now risen again, to 47.67%, the highest reading since April 2. Band width and realized volatility use different windows on the same data and they answer different questions: band width asks how far apart the last twenty closes are, realized volatility asks how much the daily returns have varied. A market that grinds steadily in one direction can compress the first while leaving the second untouched, which is exactly what late August and early September did. Thursday reconciled them in the ordinary way — one large move, both readings up.
Volume vs 20-day average	6,391 BTC vs 6,443 average (0.99×)	At the average	5,256 → 6,391 BTC	Up 21.6% on Wednesday and the heaviest session since September 4, though still the 100th lightest of 2026’s 253 sessions. The ratio to the twenty-day average reached 0.99, its highest since September 3, and it got there from both ends: volume rose while the average kept falling, from 7,136 to 6,443, as the heavy late-August days continue to roll out of the window. A 2.23% decline on average volume is a more ordinary event than the same decline on a volume spike — it describes a market repricing rather than a forced one. Of 649 earlier days in this series with volume between 0.9 and 1.1 times its twenty-day average and price above the 200-day, the close was higher ten sessions later just 52% of the time (median +0.23%) and twenty sessions later 54% (median +0.75%) — both below the all-day baseline of 56%.
US spot ETF net flows	Sep 9: −$120.2M; Sep 8: −$46.6M; September ≈ +$603M; August +$3,524M; Sep 10 not yet reported	Two-day outflow after a three-week inflow run	Backlog cleared: two prints arrived, both negative	The five-edition gap closed and the answer was two outflow days. SoSoValue and Farside data, as reported by Crypto Briefing and Crypto Economy, put Tuesday September 8 at −$46.6 million and Wednesday September 9 at −$120.2 million, a two-day total of −$166.8 million. Cointelegraph’s headline framed the same episode as a single $167 million figure; the daily breakdown is used here because it is the more granular of the two and the two reconcile to within rounding. Wednesday’s per-fund split had ARK 21Shares’ ARKB at −$78.0 million, Grayscale’s GBTC at −$27.2 million and BlackRock’s IBIT at −$19.5 million, with Morgan Stanley’s MSBT the only product taking money in at +$4.5 million across the two days. Across both days GBTC shed $92.7 million, ARKB $69.9 million and IBIT $8.8 million. Context matters for the size: the same reports put the preceding three-week inflow run at about $3.8 billion, so a $166.8 million pullback gave back roughly 4.4% of it, and total US spot fund assets stood at about $99.3 billion. Cumulative net inflows since launch are around $55 billion; 2026 year-to-date flows remain negative at about −$1.07 billion. Ether ETFs took in $34.7 million on Wednesday and Solana funds $11.2 million.
Perpetual funding rate	Not reported when this brief was written	Unavailable	No comparable figure	Funding is the periodic payment between holders of perpetual futures that keeps the contract tethered to spot; positive means longs pay shorts. The CoinStats page that supplied this reading for the last week could not be fetched on this run, and the CoinGlass futures page, which was reachable, returned its weighted funding rate as an unloaded field. No figure is therefore quoted and none is estimated. What the liquidation data implies indirectly is set out in the row below: a day in which forced closures ran roughly seventeen times the prior session’s and skewed heavily long is consistent with positive funding going into it, but that is an inference about positioning, not a measurement of the rate, and it is labelled as such.
Open interest / liquidations	24-hour liquidations $562M across crypto; OI $53.32B displayed on CoinGlass	Heavily long; largest liquidation day of the week	Liquidations $33.48M → $562M (different sources)	Liquidations are forced closures of leveraged positions by an exchange when margin runs out. Bitcoin.com News reported $562 million of crypto liquidations over the twenty-four hours around the PPI release, describing the large majority as long positions, with about $74 million of long liquidations in the four hours immediately after the data and $112 million of Bitcoin-specific long liquidations across the full day. The per-side split that report published did not sum to its own headline total, so no long/short percentage is quoted here. The comparison with Tuesday’s $33.48 million is across different sources and different window conventions, so it is offered as an order of magnitude rather than as a ratio: Thursday was a liquidation event and Tuesday was not. Open interest is recorded at the $53.32 billion CoinGlass displayed; the page’s 24-hour change field did not load, so no day-over-day move is claimed, for an eighth consecutive edition.
Fear & Greed	69 (“Greed”) on CoinGabbar’s September 10 mark	Greed	Not comparable — different source from yesterday’s	alternative.me’s dated API could not be fetched for a second consecutive run, so the 74, 73, 73, 71, 69 series this brief carried for September 4 through 8 still has no entry for September 9 or 10, and none is invented. CoinGabbar’s September 10 market summary showed 69 (“Greed”). Because yesterday’s edition used CoinStats’ separately constructed index at 70, and today’s uses CoinGabbar’s, the two are not a series and no one-point change is claimed. The structural caveat survives every source change: these indices blend volatility, momentum, dominance and survey inputs rather than tracking price, and the collapse in volatility that chart 3 shows through early September pushed them up mechanically. Thursday’s expansion in band width works the other way, which is one reason to expect the next reading to fall regardless of what price does.
Hashrate / difficulty	Difficulty 127.45T; last retarget +1.31% at block 965,664 on Sep 6; block height 966,373 on CoinWarz	Between retargets	Height 966,271 → 966,373	The network data this brief has been unable to pin down for a week finally has a fixed point. Bitcoin.com News reported the most recent difficulty adjustment as +1.31% on September 6 at block 965,664, taking difficulty to 127.45T — which matches the figure CoinWarz has been displaying and confirms the next 2,016-block boundary at height 967,680, now 1,307 blocks away with the cycle 709 blocks in. That implies a retarget around September 19. The hashrate readings remain unusable and this brief continues to decline to quote one: Bitcoin.com News put the network at 934 EH/s, CoinWarz’s own page showed 861 EH/s at block 966,373, and a search summary of the same site showed 944 EH/s at block 966,350 — three figures spanning 10% across essentially the same moment. Hashrate is inferred from block intervals rather than measured, which is why short-window estimates scatter like this; the September 19 retarget will produce a number derived from 2,016 actual blocks. Hashprice — miner revenue per unit of computing power — was reported at $39.63 per PH/s per day, up 22.24% over thirty days.

Price and trend (chart 1)

Thursday was the opposite of Wednesday in every respect that matters to a candle. Bitcoin opened at $78,283.98 — Wednesday’s close, as the daily candles here are continuous — reached $78,554.18 within the first part of the session, fell to $76,440.00, and closed at $76,536.55, a loss of $1,747.43 or 2.232%. The range was $2,114.18, or 2.76% of the closing price and 0.940 times Wednesday’s ATR, which ranks 122nd narrowest of 2026’s 253 sessions by dollars and 112th by percentage of the close: a slightly-narrower-than-median day by travel. But the body — the distance between open and close — was $1,747.43, which is 82.65% of the range, the eleventh-largest share of any session this year, with only a $270.20 upper wick and a $96.55 lower one (a wick is the thin line above or below the candle body, marking where price traded but did not close). The close finished 4.6% of the way up from the low. Where Wednesday’s candle was the technical definition of indecision — an $165 body inside a $2,026 range — Thursday’s is the definition of a one-way session: price opened, went almost immediately in one direction, and closed on its lows. Volume was 6,390.76 BTC against a twenty-day average of 6,443.46, a ratio of 0.992, up 21.6% on Wednesday and the heaviest session since September 4. That is worth pausing on. A 2.23% decline on volume at the twenty-day average is not a panic; it is a repricing that the ordinary flow of the market was able to absorb.

The short average broke in both directions at once. The close of $76,536.55 finished $2,066.96 below the 20-day simple moving average of $78,603.51 — a second consecutive close beneath that line, and five times Wednesday’s $408.98 gap. Wednesday’s crossing was mostly the average rising into a flat price; Thursday’s widening was entirely price falling away from it. And the average itself turned: it fell $89.45, its first decline since August 16, ending a run of twenty-four consecutive rises that began on August 17 and spanned the whole breakout. The turn was narrow — the close entering the window, $76,536.55, replaced August 21’s $78,325.54, a difference of $1,789, so the average would have fallen on a far smaller decline than Thursday’s — but a 20-day average that has stopped rising is a different object from one that is still climbing, because it stops manufacturing crossings on its own. Elsewhere the September range came under pressure at one end. The low of $76,440.00 was the lowest since September 2’s $76,219.18, which remains the month’s floor by $221 and is now the nearest reference point beneath the market; the month’s high remains September 3’s $82,283.00. Against longer reference points the close sits 39.40% below the all-time high of $126,296, 12.53% below the 2025 year-end close of $87,497.94, and 30.78% above the June 30 low close of $58,523.93. September is 2.58% lower across its first ten sessions, a $2,026.19 decline from August 31’s $78,562.74. This record contains sixteen completed Septembers; nine finished negative, and the month’s average return of −3.3% is the weakest of the twelve, though the last three Septembers closed +4.0%, +7.3% and +5.4%.

The long averages did not notice. The 50-day rose $209.00 to $70,400.82 — a thirtieth consecutive daily rise — while the 200-day rose $44.59 to $69,964.06, taking the golden-cross gap from $272.36 to $436.77, a widening of $164.41 on a day price fell 2.23%. This is the third consecutive session in which the gap has widened, and it is the clearest available demonstration of the timescale difference between these lines and the market: the 50-day window is still shedding the low closes of late July, so a single 2% down-day cannot arrest it. Bitcoin closed 9.39% above the 200-day average, 8.72% above the 50-day, 14.68% above the 100-day and 2.63% below the 20-day — a twenty-third consecutive close above the 200-day in the run that began on August 19. The compression in that last figure is the one to watch rather than the crossing: the distance to the 200-day fell 2.57 percentage points in a single session, the largest one-day narrowing of the run, and it now stands at $6,572.49, the smallest cushion since August 20. That is still a long way — an 8.59% further decline would be required to reach the line — but it is the first session of this run in which the gap moved meaningfully.

One point of housekeeping on the golden cross, because two respectable sources now describe it differently. This brief’s data has the 50-day crossing above the 200-day on September 8, and the gap has widened on that session and on each of the two since. Decrypt, on Thursday, published a piece headlined that a golden cross “is coming,” describing the pattern as about to flip for the first time since November. Both can be true of different series: moving-average crossings depend on which exchange feed, which close convention and which averages are used, and a cross that has printed on 00:00 UTC candles from one feed may still be a day or two away on another. The figures in this brief are computed on the Alpha Vantage BTC/USD daily series throughout, and the crossing date and gap quoted here follow from that. Readers comparing with other coverage this week will find the discrepancy and should know its cause. What neither version settles is what follows: of the thirteen earlier golden crosses in this series since February 2011, price was higher thirty sessions later in six of twelve measurable cases, with a median of −2.62% — a coin flip at one month — improving to two-thirds and a double-digit median at two and three months, on a sample small enough that one instance moves it.

Levels being discussed (attributed)

Level	Why it is being watched	Who cites it
$400,000	Named as reachable by 2030, in an argument that the downturn’s low is already in and the next halving is roughly eighteen months away	Brian Armstrong, Coinbase chief executive, via Decrypt (Sep 10)
$90,000	Described as the higher objective if $80,000–$82,000 were reclaimed as a floor	Cryptonews (Sep 10)
$83,000–$86,000	The band in which roughly 1.07 million BTC was acquired and has sat unmoved for thirty days — described as potential overhead supply rather than active selling	Glassnode, via CryptoSlate (Sep 10)
$82,793	Retracement resistance in Reuters’ technical mapping of the move	Reuters, via Cryptonews (Sep 10)
$82,283	The September 3 high and the month’s high, the highest print since May 11	Alpha Vantage feed
$81,616	The 365-day moving average, computed here	This brief’s calculation
$80,987	Upper Bollinger Band, 5.8% above the close and $101 higher than yesterday	This brief’s calculation
$80,000–$82,000	The resistance band named in the day’s published analysis, described as having rejected price on multiple attempts	Cryptonews (Sep 10); usethebitcoin (Sep 10)
$79,630	50-week simple moving average, computed here on weekly bars; the close is $3,093 below it	This brief’s calculation
$78,604	20-day simple moving average, $2,067 above the close and falling for the first time since August 16	This brief’s calculation
$78,554	Thursday’s high, reached before the producer-price release	Alpha Vantage feed
$77,600–$77,900	The nearer support zone in the day’s published analysis	usethebitcoin (Sep 10)
$77,000	The 50-week exponential moving average, described as “the line in the sand”; a hot inflation print was named as what could push price beneath it	Ted Pillows, via Cryptonews (Sep 10)
$76,900–$77,300	The lower short-term support zone in the day’s published analysis	usethebitcoin (Sep 10)
$76,651	The intraday low quoted in exchange-feed coverage, $211 above the 00:00 UTC candle low used throughout this brief — a reminder that “the low” depends on which feed is read	Bitstamp data, via Bitcoin.com News (Sep 10)
$76,537	Thursday’s close — also the lowest close inside the current 20-day Bollinger window, and one of the two numbers now setting band width	Alpha Vantage feed
$76,440	Thursday’s low on the 00:00 UTC candle used throughout this brief	Alpha Vantage feed
$76,350	The average cost basis of active holders as reported earlier in the week; the close finished $187 above it	Bitfinex, via CoinTurk (Sep 9)
$76,220	Lower Bollinger Band, $317 below the close and $280 lower than yesterday — within a dollar of the September 2 low	This brief’s calculation
$76,219	The September 2 low — the floor of the month’s range, $317 below Thursday’s close and $221 below Thursday’s low	Alpha Vantage feed
$76,000–$77,000	Named as the cluster of leveraged positions that would come under pressure on a break beneath $77,200	Cryptonews (Sep 10)
$73,000–$75,000	Described as the broader support shelf beneath the immediate zone	Cryptonews (Sep 10)
$72,854	200-day exponential moving average	This brief’s calculation
$72,000–$74,000	The zone named earlier in the week as where a deeper decline would be watched, now 3.3% to 5.9% below the close	Michaël van de Poppe, via CoinCentral (Sep 9)
$70,401 / $69,964	50-day and 200-day simple moving averages — the two lines that crossed on September 8, now $437 apart	This brief’s calculation

One note on the table. Eleven of these twenty-five entries carry a September 10 attribution, seven are this brief’s own calculations, five come straight off the price feed and two are carried from September 9. The most striking feature is how tightly the numbers now cluster beneath the close. Thursday’s low at $76,440, the lower Bollinger Band at $76,220, the September 2 monthly low at $76,219 and Bitfinex’s $76,350 active-holder cost basis occupy a span of $221 — narrower than a tenth of a single day’s ATR, and all of it within 0.5% of the close. Above, the nearest references are much further away: the 20-day average at $78,604 is 2.70% up, the 50-week average at $79,630 is 4.04% up, and the $80,000 round number is 4.53% up. That asymmetry is new. For most of this run the close sat inside a lattice of levels in both directions; Thursday’s session moved price to the bottom edge of the structure, where several independent reference points happen to coincide. Stripping out the long-horizon projections, the working band analysts named on Thursday runs from $73,000 to $82,793 — Cryptonews’ lower shelf and Reuters’ retracement level — which is $9,793 wide, or about 4.37 times Thursday’s ATR. The Bollinger envelope is much tighter at $76,220 to $80,987, a $4,767 span against $4,386 a day ago; that envelope widened for the first time in twelve sessions, but it widened by moving its floor down, not its ceiling up.

Momentum and volatility (charts 2 and 3)

Momentum stopped leaking and started draining. The MACD line fell to 2,266.87 from 2,615.14 while the signal line eased to 2,945.56 from 3,115.23, widening the histogram to −678.69 from −500.09 — a seventh session below signal and a seventh consecutive widening, the sequence now running −49.50, −108.96, −139.48, −261.93, −396.55, −500.09, −678.69. The $679 gap between line and signal is the widest since June 10, and the day’s $178.60 increase was the largest single-session widening of the run, against $122.45 on September 7 and $134.62 on September 8. The MACD line remains 2,267 above zero, so the description is unchanged in kind if not in degree: a loss of upward momentum, not a change of trend direction, with the twelve-day exponential average shedding the memory of September 3’s $81,264 close faster than the twenty-six-day average is. RSI fell 6.68 points to 52.77, a fourth consecutive decline and the largest of the four; it is the lowest reading since August 16 and the first reading below 55 since August 17, which means the whole of the August breakout’s overbought excess — RSI peaked at 86 on August 21 — has now been given back. ADX fell a third straight session, to 45.66 from 46.90, with +DI easing from 30.54 to 28.48 and −DI rising from 12.23 to 15.51, narrowing the directional spread from 18.31 to 12.97. That is a fourth consecutive narrowing, driven from both ends: +DI has fallen on each of the four sessions and −DI has risen on three of them, Thursday’s 3.28-point jump being the largest.

The volatility reading did what yesterday’s brief said it would be free to do. Bollinger band width rose for the first time in twelve sessions, to 6.065% from 5.574% — an increase of 0.491 points, ending the eleven-day contraction that began at 40.03% on August 29. The distinction this brief has drawn for four editions now pays off, so it is worth stating the arithmetic once more. Through September 9 the contraction was mechanical: the twenty-day window was expelling the enormous closes of the August breakout, so band width fell whether or not the market did anything. By Wednesday the window had nothing left to expel — its extremes were $81,263.99 and $77,054.44, a spread of $4,210 — and this brief calculated that a flat Thursday would leave band width at about 5.58%. Thursday was not flat, and the reading moved to 6.07% because a $1,747 decline widened the dispersion of the twenty closes: the standard deviation rose from $1,096.59 to $1,191.75. In other words, the first informative reading in eleven sessions said volatility is increasing, and the mechanism was price, exactly as designed. Two caveats keep this from being a tidy story. First, the expansion so far is small — 0.491 points against contractions that ran as large as 6.74 points a day. Second, the window still has one large object in it: August 22’s $77,054.44 close leaves on Friday, and rolling forward on an unchanged close would put the standard deviation at $1,230.13 and band width near 6.26%, so part of the next reading is again calendar rather than market. The other volatility measures moved as expected. ATR fell a sixth day, by $9.70 to $2,240.33, or 2.93% of price. %B fell to 0.0664 from 0.4068, its lowest since June 30, in a band now running $76,220.01 to $80,987.01. Thirty-day realized volatility rose 0.92 points to 47.67%, its highest since April 2 — which means the two volatility measures moved in the same direction for a second consecutive session, and rose together for the first time since August 28.

Since the squeeze has now released, the claim worth testing is the popular one: that a squeeze release produces a large move. It does not, in this record. Take every occasion since 2016 on which band width was below 8% and expanded after at least five consecutive contractions — Thursday is the thirty-seventh such release, and the twelve most recent are dated August 2023, September 2023, October 2023, November 2023, February 2025, June 2025, July 2025, September 2025, January 2026, May 2026, July 2026 and today. Across the thirty-six of them with ten sessions of data behind them, the median absolute move over the following ten sessions was 4.88%, against a 5.53% all-day baseline for the same period; over five sessions it was 2.52% against 3.84%; over twenty sessions, 6.34% against 8.93%. At every horizon the typical move after a squeeze release has been smaller than the typical move on a randomly chosen day. This is the mirror of the finding in yesterday’s edition, which looked at squeeze entries and found the same thing — and taken together the two say something fairly simple about compression in this series: it clusters. Quiet periods are followed by quiet periods more often than by violent ones, in both the approach and the aftermath, and the tail events that make the popular claim memorable are rare enough not to shift the median. The direction question is separate and equally unhelpful: of the thirty-six with forward data, the close was higher ten sessions later 64% of the time with a median of +2.09%, against a 56% baseline — a mild positive tilt that sits well inside the noise for a sample of that size, and one that is in no way a forecast for a particular release.

Flows and positioning (chart 4)

Chart 4 has two new bars for the first time in five editions, and both point down. The reporting backlog this brief flagged on Tuesday and Wednesday cleared on Thursday morning, and the missing sessions were outflows: Tuesday September 8 at −$46.6 million and Wednesday September 9 at −$120.2 million, a two-day total of −$166.8 million on SoSoValue and Farside data as reported by Crypto Briefing and Crypto Economy. Cointelegraph covered the same episode under a headline framing $167 million as a single figure; the daily breakdown is used here because it is more granular and because the two reconcile to within rounding. Wednesday’s per-fund split is the informative part. ARK 21Shares’ ARKB shed $78.0 million, the largest single-fund redemption; Grayscale’s GBTC shed $27.2 million and BlackRock’s IBIT $19.5 million, while Morgan Stanley’s MSBT was the only product taking money in, at $4.5 million across the two days. Across the pair of sessions, GBTC lost $92.7 million, ARKB $69.9 million and IBIT $8.8 million — a different ordering from the single day, which tells you the two sessions were not the same event repeated. September’s six reported sessions now stand at about +$603 million against August’s completed +$3,524 million, and the month remains net positive despite two consecutive negative days. Proportion matters here: the three-week run into September 8 gathered roughly $3.8 billion, so $166.8 million of redemptions gave back about 4.4% of it. Total US spot Bitcoin fund assets stood near $99.3 billion; cumulative net inflows since launch are around $55 billion, while 2026 year-to-date flows remain negative at roughly −$1.07 billion. Ether ETFs took in $34.7 million on Wednesday, recovering from a $24.3 million outflow the day before, and Solana funds $11.2 million. Thursday’s own figure — the one that would show how the ETF channel responded to the PPI release — had not been published when this brief was written.

The derivatives picture is the day’s largest number and its least precise one. Bitcoin.com News reported $562 million of crypto liquidations across the twenty-four hours spanning the producer-price release, with about $74 million of long liquidations concentrated in the four hours immediately afterwards and $112 million of Bitcoin-specific long liquidations over the full day. The report described the great majority of the total as long positions but published a per-side split that does not sum to its own headline figure, so no long/short percentage is carried here. The comparison with the $33.48 million this brief reported on Wednesday is across different sources and different window conventions and is therefore offered only as an order of magnitude: Thursday was a liquidation event and Tuesday’s session was not. The shape of it is consistent with what the price chart shows — a market positioned for the range to hold, met by a data release that broke it, with the forced selling concentrated in the hours immediately after the print rather than spread across the session. Funding, which would say whether that positioning was crowded, is missing: the CoinStats page that has supplied it all week could not be fetched, and CoinGlass’s weighted funding field did not load. Open interest is recorded at the $53.32 billion CoinGlass displayed, with no day-over-day change claimed for an eighth consecutive edition because the page’s change field was likewise unavailable. On the wider market, CoinGabbar put Bitcoin’s share of total crypto value at 56.9% on September 10 with Ether at 10.9% and total market capitalization at $2.76 trillion, down 0.9% on the day; that is a different construction from the tv-hub.org series carried earlier in the week, which showed 58.6% for September 9, and from CoinStats’ 59.0%, so it is quoted as its own reading rather than as a continuation.

On-chain and network

Thursday’s on-chain reading was sell-side risk, and it is worth explaining because the name is opaque. The sell-side risk ratio divides the total realized profit and loss locked in on-chain over a window by the realized capitalization — roughly, the aggregate price at which all coins last moved. A high ratio means holders are taking a lot off the table relative to the size of the stored wealth; a low one means they are mostly sitting still. CryptoSlate, citing Glassnode’s September 9 report, put the ratio at 7 basis points per day on a seven-day basis, down 56% from an August peak of 16 and back into the range the firm describes as rare lows. Long-term holders — coins that have not moved in at least 155 days — now account for 47% of realized profit, against 88% at the August peak, and Glassnode noted that “the realized-profit spike on Sept. 3 was less than half the size of August’s spike.” Cointelegraph made the same point from a different angle, reporting that Bitcoin investors showed little sign of panic selling as BTC held most of its August gains. The constructive reading of that is straightforward: the people who bought low are no longer distributing into strength at the rate they were three weeks ago. Two things sit against it in the same analysis. The first is supply: roughly 1.07 million BTC acquired between $83,000 and $86,000 has sat largely unchanged for thirty days, which is a body of coins bought above the market that has neither capitulated nor been resolved. The second is demand: spot cumulative volume delta — the running difference between aggressive buying and aggressive selling on exchanges — remained negative through September 8, meaning that even through the inflow run, aggressive exchange selling outweighed aggressive buying. Bitfinex’s contribution, reported earlier in the week, is the arithmetic one: more than 71% of circulating supply sits at an unrealized profit, against a historical mean of 74.7%. None of these is a timing signal, and Thursday’s decline is not evidence for or against any of them; they describe the composition of the market that took the decline.

On the network itself, a week of contradictory readings finally has an anchor. Bitcoin.com News reported that the most recent difficulty adjustment was +1.31%, at block 965,664 on September 6, taking difficulty to 127.45T — which matches what CoinWarz has been displaying and fixes the next 2,016-block boundary at height 967,680. With the chain at 966,373 when the page was read, that is 1,307 blocks away and the cycle 709 blocks in, implying a retarget around September 19. Difficulty is the protocol’s self-correcting mechanism: every 2,016 blocks it adjusts so that blocks keep arriving about every ten minutes, so a positive adjustment means the preceding fortnight’s blocks came faster than target. The hashrate figures remain unusable and this brief continues to decline to quote one as a reading: Bitcoin.com News put the network at 934 EH/s, CoinWarz’s own page showed 861 EH/s at block 966,373, and a search summary of the same site showed 944 EH/s at block 966,350 — a 10% spread across essentially the same moment, which is an artefact of inferring hashrate from block intervals over short windows rather than a real change. What can be said with more confidence is the economics: hashprice, the revenue a miner earns per unit of computing power per day, was reported at $39.63 per PH/s, up 22.24% over thirty days, and the same coverage noted that miners have not expanded hashrate much despite that improvement, which is consistent with hardware rather than economics being the binding constraint. Elsewhere, the CFTC authorized US institutional access to Singapore Exchange Bitcoin perpetual futures, a venue that has done about $5.8 billion of cumulative volume since November 2025 and averaged $19 million a day in August — small in absolute terms, but the first regulated onshore route for US institutions into a perpetual contract. The European Securities and Markets Authority separately flagged tokenized equities, DeFi exploits and prediction markets as channels through which crypto could transmit risk to traditional finance. And the Liquid Network resumed block production after a $320 million exploit, with transactions and peg operations still suspended while recovery continues — a reminder that sidechain risk and base-layer risk are different things, and that this was the former.

Sentiment

The index this brief normally leads with is missing for a second consecutive run. alternative.me’s dated API — the source of the 74, 73, 73, 71, 69 series carried for September 4 through 8 — could not be fetched, so that series still has no entry for September 9 or September 10, and none has been estimated. CoinGabbar’s September 10 market summary showed a Fear & Greed reading of 69, in “Greed” territory. Yesterday’s edition used CoinStats’ index at 70; today’s uses CoinGabbar’s at 69; these are not the same measure and the one-point difference between them is not a change, so none is claimed. The structural caveat carries across all three sources: these indices blend volatility, momentum, market dominance and survey components rather than tracking price, and the collapse in volatility through early September pushed them mechanically higher — which is why a “Greed” reading persisted through a week in which price went nowhere. Thursday’s expansion in band width and the 0.92-point rise in realized volatility work the other way, so the direction of the next reading is more likely to be down than up whatever price does, and that will be a property of the construction rather than of the market.

Of the 15 Bitcoin-tagged articles in the Alpha Vantage news feed published between 00:00 UTC on September 10 and 04:00 UTC on September 11 — down sharply from 28 on Wednesday — the feed’s sentiment model scored none as bullish, two as somewhat bullish, eleven as neutral, two as somewhat bearish and none as bearish. Sourcing was eight from Cointelegraph, four from Decrypt and three from Motley Fool. The collapse in article count is itself the most interesting feature of the tally and probably says more about the news cycle than the market: Wednesday’s 28 included a heavy policy and corporate docket, while Thursday’s coverage concentrated on a single macro event. The two somewhat-bearish scores attached to Cointelegraph’s account of the PPI-driven decline and to its report on the Trezor and BitBox phishing emails — one about market conditions, one not. The two somewhat-bullish scores went to Decrypt’s report of Brian Armstrong’s long-horizon comments and to a Motley Fool piece on Cathie Wood’s crypto-equity purchases. As always, a tally like this measures the tone of coverage as a machine reads it across everything tagged to the ticker, and is reported for continuity rather than as a market indicator.

Published views, in both directions. The constructive case on Thursday was almost entirely long-horizon, which is what tends to happen on a down day. Brian Armstrong, Coinbase’s chief executive, told Decrypt he still regards $400,000 by 2030 as a reasonable figure, arguing that the bottom of the downturn is already in a year into it and that the next halving is roughly eighteen months away. Glassnode’s sell-side risk work, via CryptoSlate, is the nearer-term version of the same argument: holders are locking in far less profit than they were in August, with long-term holders’ share of realized profit down from 88% to 47%. Cointelegraph’s read of the same data was that investors “showed little sign of panic selling” as Bitcoin held most of its August gains. Against that, the cautious voices were about the next forty-eight hours rather than the next four years. Ted Pillows, writing before the release, named the 50-week exponential moving average near $77,000 as “the line in the sand” and said explicitly that “any sign of inflation heating up could push Bitcoin below the 50W EMA” — a conditional statement whose condition arrived the same afternoon, and whose level the 00:00 UTC close then finished $463 below. Cryptonews’ own framing set out three paths without endorsing one: reclaiming $80,000–$82,000 as a floor, consolidation between $77,000 and $80,000, or a break beneath $77,200 running into a cluster of leveraged positions between $76,000 and $77,000. Glassnode’s 1.07 million BTC of unmoved supply bought between $83,000 and $86,000 is the structural version of the bearish case, and the negative spot cumulative volume delta through September 8 is the flow version. Beneath all of it, the macro voices remain the most consistently bearish, on the argument that a Federal Reserve now more likely than not to raise rates next week lifts bond yields — which set fresh fifty-two-week highs on Thursday at both the ten- and thirty-year points — and redirects capital away from an asset that pays no yield.

Macro and cross-asset

Thursday’s session had a single cause and it arrived at 8:30 in the morning in New York. August producer prices rose 0.4% on the month, matching expectations, but the annual rate came in at 5.4% against 4.8% in July — an acceleration driven, according to the day’s coverage, by a 4.2% surge in energy commodities. The bond market took it badly at the long end: the thirty-year Treasury yield added 8.1 basis points to 5.367% and the ten-year 11.4 basis points to 4.954%, both fifty-two-week highs, with Cointelegraph describing the thirty-year’s level as a nineteen-year high. Energy compounded it: WTI crude rose 4.2% to $100.10 a barrel and Brent 3.6% to $105.37, the highest since July, which matters because it feeds directly into the next month’s inflation prints. Equities fell for a fourth consecutive session — the S&P 500 down 0.58% to 7,591.79, the Nasdaq Composite down 0.65% to 26,081.73, the Dow down 0.60% to 52,064.46 and the Russell 2000 down 1.02% to 2,891.49 — and the Cboe Volatility Index rose 8.69% to 17.89, its largest one-day gain of the week. Gold closed at $4,427.40, quoted as down 0.75% on the day; that level is $14.90 below the $4,442.30 mark this brief carried for Wednesday, which would imply a 0.34% decline instead, so the direction is agreed between the two marks but the size is not, and the level is reported as published. On the rate expectations that all of this feeds, the readings cluster but do not agree: TheStreet reported CME FedWatch odds of a hike at “nearly 70%”, Bitcoin.com News put the CME figure at 67% and Polymarket and Kalshi bettors at 63%, a nine-point jump on the day. What is not in dispute is the direction of travel — a week ago the same coverage was quoting roughly 60%.

Bitcoin’s 2.23% decline was larger than any of the four equity indices’ and larger than the move in gold, which breaks the pattern of the preceding week: on Tuesday and Wednesday Bitcoin fell less than small-cap equities, and this brief noted that relative resilience as the most defensible bullish observation available about those sessions. Thursday removed it. That is what a correlated risk asset does when the shock is a rate shock rather than an equity-specific one, and it is the transmission channel the coverage keeps describing: higher policy rates raise the yield available on dollar cash and reduce the relative appeal of an asset that pays none. At Thursday’s marks Bitcoin buys 17.29 ounces of gold, against 17.62 on Wednesday — a ratio that fell because Bitcoin fell considerably more than gold did. The policy question is now three working days away: the FOMC meets September 15–16 with the target range at 3.50%–3.75%, and the single report standing between here and that meeting is Friday morning’s August consumer price index, for which economists surveyed before the release were looking for an annual rate near 3.3% to 3.4%, down from 3.4% in July, on a 0.4% monthly increase. A reading in line with that would leave Thursday’s producer-price acceleration looking like an energy story confined to the pipeline; a reading above it would put a second inflation surprise in front of a committee that futures already price as more likely than not to tighten.

Calendar: what’s ahead

Date (ET)	Event	Why Bitcoin traders watch it
Fri, Sep 11, 8:30 a.m.	August consumer price index	The report the coverage has described both Bitcoin and the S&P 500 as waiting on all week; economists surveyed beforehand were looking for an annual rate near 3.3%–3.4% on a 0.4% monthly rise. The last report before the FOMC convenes
Fri, Sep 11, morning	US spot ETF flow print for Thursday	The first read on whether the ETF channel added to the PPI-day selling or absorbed it; the two prior sessions were both outflows
Fri, Sep 11, 8 p.m.	Next Bollinger reading	August 22’s $77,054 close leaves the 20-day window; on an unchanged close that alone would take band width to about 6.26%, so part of the move will again be calendar rather than market
Sun, Sep 13, 8 p.m.	Weekly candle closes (00:00 UTC Monday)	The frame on which the 50-week averages are defined — the simple average computed here is near $79,630 and the exponential average Ted Pillows cites near $77,000
Tue, Sep 15	Senate cloture vote on the CLARITY Act	Cloture is the vote to end debate and move to a final vote; it needs 60 senators. Cointelegraph’s earlier analysis set out that a failure could see the bill remade or scrapped under a different Congress in 2027
Tue–Wed, Sep 15–16	FOMC meeting; rate decision Wednesday 2 p.m.	CME futures were pricing a 25 basis-point increase at roughly 67% to 70% after Thursday’s data, against about 60% a week earlier; the target range is 3.50%–3.75%
Wed, Sep 16, 8:30 a.m.	August retail sales	Lands the morning of the FOMC decision; the consumption side of the same inflation question
Fri, Sep 18	September options expiry; Bank of Japan rate decision	The expiry’s notional size was still not reported when this brief was written. The BOJ meeting matters for the yen carry trade — borrowing in a low-rate currency to buy higher-yielding assets elsewhere — which has been a live risk to global risk assets since the yen reached 153 per dollar on Wednesday
~Sat, Sep 19	Next Bitcoin difficulty adjustment at block 967,680	1,307 blocks away when this brief was written. A measured retarget also settles the week’s contradictory hashrate readings, because difficulty adjusts on what actually happened over 2,016 blocks
Tue, Sep 29, 10:00 a.m.	August JOLTS job openings	The labour-market side of the case for or against further tightening, two weeks after the decision
Wed, Sep 30, 8:30 a.m.	August PCE inflation and the third estimate of Q2 GDP	The Fed’s preferred inflation measure, two weeks after the decision it will not have informed
Through September	Calendar-month seasonality	September has finished negative in nine of the sixteen completed years in this series and is its weakest month by average return, at −3.3% — though the last three Septembers closed positive. The month is 2.58% lower so far
Ongoing	Oil and the long end of the Treasury curve	Brent at $105.37 and a thirty-year yield at 5.367% are the two prices that drove Thursday’s session; both feed the next inflation print rather than the last one

Reading the combination

Thursday is the day the squeeze answered. For four editions this brief tracked a Bollinger band width that fell from 40.03% to 5.57% in eleven sessions and argued, with the arithmetic on the page each day, that the fall was mechanical — old outliers leaving a twenty-day window rather than a market going quiet — and that once the window had finished expelling the August breakout, the next reading would carry information. That reading arrived: band width rose to 6.07%, and it rose because price fell $1,747.43, widening the standard deviation of the twenty closes from $1,096.59 to $1,191.75. The compression broke downward. It is worth being precise about what that does and does not establish. It establishes that the quiet period ended, and which way the first move went. It does not establish that the move continues, and the base rate says the opposite of the popular intuition: across the thirty-six earlier squeeze releases in this series since 2016, the median absolute ten-day move afterwards was 4.88% against a 5.53% all-day baseline — quieter than average, not louder. The layers now disagree more sharply than at any point in this run, and the disagreement is legible rather than confusing. The trend layer is still improving: a twenty-third consecutive close above the 200-day average, a 50-day average rising for a thirtieth straight session, and a golden-cross gap that widened from $272.36 to $436.77 on a 2.23% down day. The momentum layer deteriorated on every measure: RSI down 6.68 points to 52.77, its lowest since August 16 and its fourth consecutive fall; the MACD histogram at −678.69, the most negative since June 10, on a seventh straight widening; ADX down a third day with the directional spread narrowing a fourth time and −DI up 3.28 points, its largest rise of that sequence; and a second close below a 20-day average that itself fell for the first time since August 16. The flow layer turned over: two reported ETF sessions, both outflows, totalling −$166.8 million after three weeks that gathered $3.8 billion. And the macro layer was actively hostile — a 5.4% annual producer-price rate, Brent at $105.37, a thirty-year yield at a nineteen-year high, and rate-hike odds that moved from about 60% to nearly 70% in a week. On the base rates the board splits the way it has all week and no more decisively: for, 155 earlier band-width expansions from below 8% above the 200-day were higher ten sessions later 65% of the time (median +1.69%), and 255 days with ADX at or above 45, +DI above −DI and ADX falling were higher 71% of the time (median +4.70%); against, 662 days with RSI between 50 and 56 above the 200-day were higher just 49% of the time (median −0.31%), and 940 days below a falling 20-day average above the 200-day, 52% (median +0.21%). Two things are worth holding on to. First, the cushion that has defined this run narrowed materially for the first time: the distance to the 200-day average fell 2.57 percentage points in a session, to 9.39%, its smallest since August 20 — still $6,572 of room, but no longer untouched. Second, the level structure beneath the close is the tightest coincidence this brief has recorded. The lower Bollinger Band at $76,220.01 sits $0.83 above the September 2 monthly low of $76,219.18; Thursday’s low at $76,440.00 stopped $219.99 above the band and $220.82 above the September low; and Bitfinex’s $76,350 active-holder cost basis sits between them. Four independent reference points inside $221, all within 0.5% of a close that finished near its low. Above, the nearest reference is the 20-day average $2,066.96 away. The indicators do not say which way it resolves, and Friday morning’s consumer price index is more likely to decide it than anything on this board.

Sources

Price, volume and all technical indicators: Alpha Vantage BTC/USD daily series (00:00 UTC candles), computed by the author; the September 10 candle was confirmed complete by the presence of a partial September 11 row. Thursday’s session narrative, the PPI overshoot and the thirty-year yield: Cointelegraph’s September 10 market report. Liquidations, the Bitstamp intraday low and the Polymarket, Kalshi and CME rate-hike odds: Bitcoin.com News. ETF flow detail for September 8 and 9: Crypto Briefing and Crypto Economy, with the same episode covered by Cointelegraph; flow history from Farside Investors and SoSoValue. Sell-side risk, long-term holder profit share, the $83,000–$86,000 supply band and spot CVD: CryptoSlate on Glassnode’s September 9 report, with Cointelegraph covering the same series. Ted Pillows’ 50-week EMA level, the scenario framing and Reuters’ retracement level: Cryptonews. Brian Armstrong’s long-horizon comments: Decrypt; the differing golden-cross timing: Decrypt. Support and resistance zones and the 71%-in-profit figure: usethebitcoin. Equity, oil, gold and volatility closes, Treasury yields and the CME FedWatch odds: TheStreet’s September 10 market blog. The 50-week average level, ETF asset total and the SGX perpetuals authorization: Bitcoin News Digest. Difficulty, hashrate and hashprice: Bitcoin.com News on mining and CoinWarz. Dominance, Ether’s price and total market capitalization: CoinGabbar, which also supplied the Fear & Greed reading; alternative.me’s dated API could not be fetched on this run. The economic calendar: FedRateCalc’s US economic calendar. News sentiment tally: Alpha Vantage NEWS_SENTIMENT feed for September 10.