Bitcoin Daily Brief — September 9, 2026

Bitcoin price on September 9, 2026: BTC closed at $78,284, down 0.21%, in a $2,026 range between a $79,753 high and a $77,727 low — a session whose candle body measured just $165, and which ended a twenty-three-day run of closes above the 20-day average while Bollinger band width contracted for an eleventh straight day to 5.57%, the narrowest since August 18 and within a rounding of the level this brief projected a day earlier. 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 8, 2026. About this series: Bitcoin Daily Brief.

Bitcoin price on September 9, 2026, at a glance

The charts

The 30-second read

Bitcoin fell 0.21% on Wednesday to close at $78,284 — a $163 decline on a candle whose body measured only $165, the smallest relative to its own range since September 2 — after a $79,753 high was sold back to $77,727 and volume came in at 5,256 BTC, 0.74 times the twenty-day average, while WTI crude rose 3.9% to $96.69 and Brent passed $101 on US strikes against Iranian tankers, the S&P 500 fell 0.48% to 7,636.36, and Treasury yields set fresh fifty-two-week highs. Underneath the price, the compression this brief has tracked for three days completed itself: Bollinger band width fell an eleventh straight session to 5.57% — against the roughly 5.6% Tuesday’s edition projected — and, for the first time, the fall was not mechanical, because the twenty-day window has now shed every outlier it contained, so a flat close on Thursday would leave band width almost exactly where it is; price meanwhile closed $409 below its own 20-day average, ending a twenty-three-session run above it that began on August 17. The calendar takes over from here: August producer prices on Thursday, August consumer prices on Friday, and an FOMC meeting on September 15–16 at which CME futures were pricing roughly a 60% chance of a rate increase.

What happened in the last 24 hours

• Price. Opened $78,449, reached $79,753, fell to $77,727 and closed $78,284 (−0.21%, −$163) — a $2,026 range on volume of 5,256 BTC, 0.74 times the twenty-day average; the high was the highest since September 7 and the low the lowest since the session before, but the close held $2,065 above September 2’s $76,219 monthly floor.
• Flows. A fifth consecutive edition with no new US spot ETF figure: neither the September 8 nor the September 9 total had been published on the trackers this brief reads when it was written, so Friday’s +$174.6 million stands as the last print and September’s four reported sessions stay at +$770 million against August’s +$3,524 million.
• Positioning. CoinStats’ figures, stamped 00:08 UTC on September 9, cover the twenty-four hours into the start of Wednesday — that is, Tuesday’s session — and fill yesterday’s gap: $33.48 million of crypto liquidations, $28.08 million long (83.9%) against $5.40 million short, with funding quoted at +0.0043% per four hours and open interest near $53.27 billion. No liquidation total for Wednesday’s own session had been reported when this brief was written.
• On-chain. Cointelegraph reported that Bitcoin’s Spent Output Profit Ratio has held above its 1.0 breakeven line since August 19 — the longest such run of 2026, now in its fourth week and reading 1.002 — with Checkonchain calling the structure similar to “early bull-market recoveries” while ARK Invest’s David Puell said he still reads it “as a downside risk”; Bitfinex separately put 71% of circulating supply in profit and the average cost basis of active holders at about $76,350.
• Macro and policy. The S&P 500 closed at 7,636.36 (−0.48%), the Nasdaq Composite at 26,253.34 (−0.64%), the Dow at 52,380.66 (−0.77%) and the Russell 2000 at 2,921.72 (−1.30%), with gold barely changed at $4,442.30, WTI up 3.93% to $96.69 and Brent above $101 for the first time since late July; the ten-year Treasury yield rose about six basis points to 4.85% and the yen reached 153 per dollar, its weakest since February, after remarks by Treasury Secretary Scott Bessent.

Section summaries

Price and trend

Price gave up its 20-day average for the first time since August 16, closing $408.98 below it and ending a twenty-three-session run above the line, while still finishing 12.0% above the 200-day average for a twenty-second consecutive session. The candle itself said almost nothing: a $165.30 body inside a $2,026 range, with a $1,303 upper wick against a $557 lower one, is a market that tried $79,753, was refused, and came back to almost exactly where it opened. The golden cross that printed on Tuesday widened rather than reversed, the 50-day pulling to $272.36 above the 200-day from $88.65. Details ↓

Momentum and volatility

The MACD histogram widened a sixth day to −500.09 as the MACD line fell $229 to 2,615.14 against a signal line easing $125 to 3,115.23 — still 2,615 above zero, so a continued loss of upward momentum rather than a change of trend direction. RSI eased just 0.66 points to 59.45, and ADX fell a second straight session to 46.90 as the directional spread narrowed again, from 18.63 to 18.31. Bollinger band width contracted an eleventh day, by 2.87 points to 5.57%, and this is where the story changes: the twenty-day window has now expelled the last of the August breakout, so from Thursday the band width measures the market rather than the calendar. Details ↓

Flows and positioning

Two US business days have now passed without a published ETF total, which is one day longer than the ordinary reporting lag and leaves Friday’s +$174.6 million as the most recent print; September’s four reported sessions stand at +$770 million against August’s completed +$3,524 million. The derivatives readings available describe Tuesday rather than Wednesday, and they were larger than Monday’s: $33.48 million of liquidations, 83.9% of them long, with funding at +0.0043% per four hours on CoinStats’ quotation and open interest near $53.27 billion. Bitfinex, quoted by CoinTurk, described the tape as “being driven by spot buying” rather than by leverage. Details ↓

On-chain and network

The day’s on-chain reading was SOPR — the ratio of the price coins are sold at to the price they were acquired at — which has stayed above 1.0 since August 19, its longest run of 2026 and now in a fourth week at 1.002, meaning coins are on aggregate moving at a profit. Named readings of it point opposite ways: Checkonchain sees the shape of an early bull-market recovery, ARK Invest’s David Puell still calls it a downside risk. On the network, CoinWarz showed difficulty unchanged at 127.45T with the chain at block 966,271 and an average block time of 9 minutes 44 seconds, which implies a retarget of roughly +2.7% around September 19. Details ↓

Sentiment

alternative.me’s dated API could not be read on this run, so no reading from the index this brief normally uses is quoted; CoinStats’ separately constructed Fear & Greed index showed 70 (“Greed”) for the window into September 9, against 72 the day before on the same source. The news feed carried 28 Bitcoin-tagged articles, up from 23 on Tuesday, scored none bullish, eight somewhat bullish, fifteen neutral, four somewhat bearish and one bearish — and, as on Tuesday, the single bearish score attached to a security story rather than to anything about price. Published views split cleanly between Ted Pillows’ $83,000-weekly-close condition for a run at $100,000 and Michaël van de Poppe’s $72,000–$74,000 as the zone he would watch on a deeper decline. Details ↓

Macro and cross-asset

Wednesday repeated Tuesday’s shape and deepened it: US equities fell across the board, with the Russell 2000 down 1.30% and the S&P 500 down 0.48% to 7,636.36, while WTI crude rose 3.93% to $96.69 and Brent passed $101 for the first time since late July after US strikes on Iranian tankers. Treasury yields set fresh fifty-two-week highs, the ten-year adding about six basis points to 4.85%, and the yen weakened to 153 per dollar. Bitcoin’s 0.21% decline was the mildest of the risk assets on the day, and CME futures were pricing roughly a 60% chance that the Federal Reserve raises rates on September 16. Details ↓

Calendar

Thursday September 10: August producer prices, weekly jobless claims and the European Central Bank decision. Friday September 11: August consumer prices, the report the coverage has described both Bitcoin and the S&P 500 as waiting on. Sunday September 13: the weekly candle closes, the frame Rekt Capital’s $78,300 breakdown condition is defined on. September 15–16: the FOMC, decision Wednesday at 2 p.m. ET; and the Senate cloture vote on the CLARITY Act. September 18: the September options expiry. September 19: the next difficulty adjustment. Details ↓

Reading the combination

Wednesday is the day the squeeze stopped being an artefact and became a fact. For three sessions this brief has argued that the collapse in band width was mechanical — old outliers leaving a twenty-day window, not a market going quiet — and that argument now expires, because the window contains nothing left to expel: its widest and narrowest closes are $81,264 and $77,054, and a flat Thursday would leave band width at about 5.58% against Wednesday’s 5.57%. That makes the next reading informative in a way the last eleven were not. Against that, the momentum layer keeps deteriorating quietly — a sixth widening in the MACD histogram, a second fall in ADX, price under its 20-day average for the first time in twenty-three sessions — while the trend layer keeps improving, the golden-cross gap tripling in a day. The bracketing levels are the tightest of the run: Rekt Capital’s $78,300 retest is $16 above the close, with Wednesday’s $79,753 high and the $80,000 round number beyond it, and $77,727 down to $77,100 below; the whole span from Wednesday’s low to the round number runs $2,273, about 1.01 times a single day’s ATR. 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 9 close)	Zone	Change since Sep 8	What this zone has historically meant
Bollinger Bands (20, 2)	%B 0.41; band width 5.57%	Squeeze — narrowest since August 18	%B 0.50 → 0.41; width 8.44% → 5.57%	An eleventh straight contraction, of 2.87 points, in a run whose daily falls read 0.70, 1.03, 1.72, 2.16, 2.16, 3.30, 4.43, 4.52, 6.74, 4.84 and 2.87. Band width is 13.9% of the 40.03% peak of August 29, eleven sessions ago, and 5.57% is the lowest reading since August 18. It is also the point at which the mechanism changes. For three days the contraction was driven by old outliers leaving the twenty-day window; that window now runs from August 21 to September 9 and its extremes are $81,263.99 and $77,054.44, a spread of $4,210, with a standard deviation of $1,096.59. Rolling it forward on an unchanged close puts band width at about 5.58% — effectively flat. Two base rates apply, and one methodological note applies to both and to every other base rate on this board: the instance count is the number of earlier days matching the filter, while the percentage is struck on the subset of those days for which the full forward window is available, so the two differ by a handful. The state is mildly constructive: of 185 earlier days with band width at or below 6% and price above its 200-day average, the close was higher ten sessions later 69% of the time (median +1.68%) and twenty sessions later 73% (median +5.87%). The entry is not: of the 21 earlier occasions when a close first took band width below 6% after a wider spell, with price above the 200-day, the close was higher ten sessions later 52% of the time (median +0.84%) and twenty sessions later 57% (median +1.69%). Today is the twenty-second such entry.
Price vs 20-day SMA	−0.52% below ($78,693)	Below, first time since August 16	+0.02% → −0.52%	The close finished $408.98 under the 20-day average, ending a run of twenty-three consecutive closes above it that began on August 17. Yesterday the gap was $17.76, the smallest of 2026 as a percentage of price; today the average kept rising, by $263.61 to $78,692.96, while price fell $163.13, and they crossed. A short average crossing under price on a $163 move is a reminder of what these lines are: a 20-day average moves whenever a new close replaces a twenty-day-old one, so it can cross price without price doing anything much at all.
50-day vs 200-day SMA	$70,192 vs $69,919 (gap +$272.36)	Golden cross — 50-day above, second day	Gap +$89 → +$272	The crossing that printed on Tuesday did not un-cross. The 50-day rose $235.36 to $70,191.82 while the 200-day rose $51.64 to $69,919.46, tripling the gap in a single session and marking a twenty-ninth consecutive day of the 50-day average rising. That matters mainly because Tuesday’s initial margin of $88.65 was the narrowest first-day gap of the last six golden crosses, and a margin that thin can be undone by a handful of ordinary sessions; $272 is still thin, but it is three times less thin. The base rates are unchanged from yesterday because they are struck on the cross date: of the twelve prior golden crosses in this series since the 200-day average first became computable in February 2011, price was higher thirty sessions later six times (median −2.62%), sixty sessions later eight times (median +12.60%) and ninety sessions later eight times (median +10.67%) — a coin flip at one month and a positive tilt at two and three, on a sample of twelve.
Price vs 200-day SMA	$78,284 vs $69,919 (+11.96%)	Above (22nd straight close, since Aug 19)	+12.28% → +11.96%	Price above the 200-day average is the most common definition of a long-term uptrend, and this is a twenty-second consecutive close above it. The distance narrowed by a third of a point, with the average doing most of the work this time: the 200-day rose $52 while price fell $163. The 200-day has now risen on each of the last nineteen sessions, since August 22 — it was still edging lower on the first three days of this run. What actually pulled the 50-day above it on Tuesday is the shorter average’s own twenty-nine-session climb.
RSI (14)	59.45	Upper neutral zone (30–70)	60.11 → 59.45	Down 0.66 points, a third consecutive decline, and still the lowest reading since August 18’s 55.58. RSI compares the average size of up-days with down-days over fourteen sessions, so a $163 fall barely registers; the reading is drifting down because the large up-days of the August breakout are ageing out of the window, not because selling has intensified. Of 436 earlier days in this series with RSI between 57 and 62 and price above its 200-day average, the close was higher ten sessions later 55% of the time (median +1.20%) and twenty sessions later 54% (median +2.09%). Readings in the high fifties are the ordinary resting state of an uptrend, not a signal in either direction.
MACD (12, 26, 9)	2,615 vs signal 3,115; histogram −500.09	Below signal, far above zero	Histogram −396.55 → −500.09 (widening)	A sixth day below the signal line and a sixth consecutive widening, the sequence now reading −49.50, −108.96, −139.48, −261.93, −396.55, −500.09. The MACD line fell $228.56 while the signal line eased $125.02. The line is still 2,615 above zero, so this describes a loss of upward momentum rather than a change of trend direction — and the gap between the two lines, $500, is the widest of this leg. Of 637 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)	46.90 (+DI 30.5, −DI 12.2)	Strongly trending (above 25), falling for a second day	47.22 → 46.90	A second consecutive fall, of 0.32 points, with the internals still narrowing: +DI eased from 31.70 to 30.54 while −DI fell from 13.07 to 12.23, taking the spread from 18.63 to 18.31 points. That is a third straight narrowing, and 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 (240 earlier instances) were followed by a higher close ten sessions later 72% of the time (median +5.05%) and twenty sessions later 71% (median +5.90%) — 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 240 independent observations.
ATR (14)	$2,250 (2.9% of price)	Falling, fifth straight day	$2,267 → $2,250	A fifth consecutive fall, of $17.25, on a $2,025.73 high-to-low range that was 0.893 times Tuesday’s ATR and 2.59% of the closing price. That range ranks 114th narrowest of 2026’s 252 sessions by dollars and 98th by percentage of the close, so it was a slightly-below-median day — and it was 7% wider than Tuesday’s $1,888. ATR and band width are telling mildly different stories here, which is worth noticing: the daily range is roughly where it has been, while the band has collapsed.
30-day realized volatility	46.8% annualized	Elevated; first fall in three days	47.25% → 46.75%	Down 0.50 points, ending two consecutive rises. This is the measure that has refused to confirm the squeeze all week: while band width fell 34.46 points from 40.03% to 5.57%, the actual dispersion of daily returns over thirty days went from 44.03% on August 29 to 46.75% now — it rose, dipping to 42.30% on August 31 and peaking at 47.25% on September 8, a five-point range against band width’s thirty-four-point one. The two use different windows on the same data, and only one of them was being rearranged by the calendar. Now that the rearranging is finished, the two are free to converge or diverge for real reasons — which is what makes the next few readings worth watching rather than explaining away.
Volume vs 20-day average	5,256 BTC vs 7,136 average (0.74×)	Below average	5,149 → 5,256 BTC	Up 2% on Tuesday and the 75th lightest of 2026’s 252 sessions. The twenty-day average itself fell again, from 7,681 to 7,136, as the heavy late-August days continue to roll out of it — the same window effect that drove the Bollinger collapse, seen on a different series. Of 572 earlier days in this series with volume below 0.75 times its twenty-day average and price above the 200-day average, the close was higher ten sessions later 55% of the time (median +0.88%) and twenty sessions later 56% (median +1.73%).
US spot ETF net flows	Not yet reported when this brief was written; last was Fri Sep 4 at +$174.6M; September ≈ +$770M; August +$3,524M	Fifth edition without a new figure	No published change	Two ordinary US business days — Tuesday and Wednesday — have now passed without a published total on the trackers this brief reads. One missing day is the normal reporting lag; two is a day longer than normal, and no explanation is available from the outside, so none is offered here. Friday’s +$174.6 million therefore remains the most recent print, with BlackRock’s IBIT at +$117.4 million of it and Fidelity’s FBTC at +$57.2 million. When the backlog clears it will carry four days of accumulated brokerage-channel orders into the two inflation reports that close this week. The medium-term picture reported earlier in the week — roughly $3.8 billion of net inflows over the three weeks to September 8, described by Crypto Briefing and Cryptonomist as the strongest three-week run of 2026 — is not affected by the gap.
Perpetual funding rate	+0.0043% per 4h (CoinStats’ quotation; ≈ +9.4% annualized)	Positive, moderate	Not comparable — the source changed its interval	Funding is the periodic payment between holders of perpetual futures that keeps the contract tethered to spot; positive means longs pay shorts. Two cautions attach to today’s figure and both matter. First, the CoinStats page carried an “09 Sep, 2026 00:08” stamp, so it describes the twenty-four hours into the start of Wednesday — Tuesday’s session — rather than Wednesday itself. Second, the same page quoted funding per four hours today, where yesterday it quoted per eight; on the ×6×365 convention that a four-hour interval implies, +0.0043% annualizes to about +9.4%, against the +2.30% derived from yesterday’s per-eight-hour figure. Because the interval changed, no day-over-day comparison is claimed here — the two numbers may or may not describe a real move in funding, and there is no way to tell from the outside. What can be said is that a reading near +9% annualized is positive but a long way below the levels usually described as crowded long positioning.
Open interest / liquidations	24-hour liquidations $33.48M across crypto; OI ≈ $53.27B on CoinStats	83.9% long; window covers Tuesday	Liquidations $5.75M → $33.48M on the same source	Liquidations are forced closures of leveraged positions by an exchange when margin runs out. This reading fills the gap yesterday’s edition flagged: the window ending at the start of Wednesday captures Tuesday’s $1,888 range and $645 fall, and it produced $33.48 million of liquidations, $28.08 million long (83.9%) against $5.40 million short — nearly six times Monday’s $5.75 million on the same source and the same convention, which is a genuine comparison rather than a cross-source one. It remains a small number in absolute terms; a heavily leveraged market produces multiples of this on a 2% down-day. No total for Wednesday’s own session had been published when this brief was written. Open interest is reported at about $53.27 billion, up from the $52.96–53.01 billion range the same tracker showed a day earlier, but for a seventh consecutive day no day-over-day change is claimed, because that tracker has been printing internally inconsistent levels all week.
Fear & Greed	CoinStats index 70 (“Greed”); alternative.me not available this run	Greed	Not comparable — different index from yesterday’s	The alternative.me dated API this brief normally reads could not be fetched on this run, so the series it has carried all week — 74, 73, 73, 71, 69 for September 4 through 8 — has no September 9 entry here, and none is invented. CoinStats’ own Fear & Greed index, which is constructed differently, showed 70 (“Greed”) for the window into September 9 against 72 the day before on the same page. Because the two indices are not the same measure, that 70 is not comparable with yesterday’s 69, and no change is claimed. One structural point carries over regardless of source: these indices blend volatility, momentum, dominance and survey inputs rather than tracking price, and a collapse in volatility of the kind chart 3 shows pushes them up mechanically — so any greed reading printed this week is flattered by the squeeze.
Hashrate / difficulty	Difficulty 127.45T; block height 966,271; hashrate 1.07 ZH/s on CoinWarz	Between retargets; blocks running fast	Height 966,124 → 966,271	Difficulty is unchanged and will stay so until the next 2,016-block boundary at height 967,680, now 1,409 blocks away with the cycle 607 blocks in. CoinWarz did not publish a headline retarget estimate when the page was read for this edition, but it did show a recent average block time of 9 minutes 44 seconds against the ten-minute target, which implies an adjustment of roughly +2.7% — close to the +2.55% the same site was projecting on Tuesday, and consistent with a retarget around September 19. The hashrate figure the site displayed, 1.07 ZH/s, is again well away from the 932 EH/s its own chart page was showing a short time earlier; this brief has flagged these swings as estimation artefacts for three days and continues to treat them that way, because hashrate is inferred from block intervals rather than measured. The September 19 retarget will settle it with an actual number.

Price and trend (chart 1)

Wednesday was, in the arithmetic, a session that did not happen. Bitcoin opened at $78,449.28 — Tuesday’s close, as the daily candles here are continuous — rose to $79,752.73, fell to $77,727.00, and closed at $78,283.98, a loss of $163.13 or 0.208%. The range was $2,025.73, or 2.59% of the closing price and 0.893 times Tuesday’s ATR, which ranks 114th narrowest of 2026’s 252 sessions by dollars and 98th by percentage of the close: a slightly-below-median day, and 7% wider than Tuesday’s. But the body — the distance between open and close — was only $165.30, just 8.16% of the range, with a $1,303.45 upper wick and a $556.98 lower wick (a wick is the thin line above or below the candle body, marking where price traded but did not close), and the close finished 27.5% of the way up the range. Only 28 of 2026’s 252 sessions have had a smaller body relative to their range; the most recent was September 2. A candle shaped like this is the technical definition of indecision: the market found a price $1,303 above its open, rejected it, found one $722 below, rejected that too, and closed $165 under where it started. Volume was 5,256 BTC against a twenty-day average of 7,136, a ratio of 0.736, up 2% on Tuesday and the 75th lightest session of the year.

The one boundary that did move is the short average. The close of $78,283.98 finished $408.98 below the 20-day simple moving average of $78,692.96 — the first close beneath that line since August 16, ending a run of twenty-three consecutive closes above it that began on August 17 and spanned the entire breakout. It is worth being precise about how little it took: the average rose $263.61 while price fell $163.13, so more than three-fifths of the $427 swing in their relationship came from the average moving, not from price. That is what a rising short-term average does when the closes entering its window are larger than the ones leaving it. Elsewhere the day changed nothing structural. The high of $79,752.73 was the highest since September 7’s $80,462.23 and the low of $77,727.00 the lowest since Tuesday’s $77,589.00, so the September range is intact at both ends: the month’s high remains September 3’s $82,283.00 and its low September 2’s $76,219.18, $2,065 below Wednesday’s close. Against longer reference points the close sits 38.0% below the all-time high of $126,296, 10.5% below the 2025 year-end close of $87,497.94, and 33.8% above the June 30 low close of $58,523.93. September is 0.35% lower across its first nine sessions. 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 50-week simple moving average computed here — using Wednesday’s close as the current, still-open weekly bar — stands at $79,869.67, $1,586 above the close; the weekly candle settles at 00:00 UTC on Monday September 14.

The golden cross widened. The 50-day average rose $235.36 to $70,191.82 while the 200-day rose $51.64 to $69,919.46, taking Tuesday’s $88.65 gap to $272.36 — a relative move of $183.72 and a twenty-ninth consecutive session of the 50-day average rising. This is worth a line because Tuesday’s initial margin was the narrowest first-day gap of the last six golden crosses, and this brief said then that a margin that thin can be reversed by a handful of ordinary down-days. It was not reversed; it tripled, on a day price fell. The mechanism is the same one that produced the crossing: the 50-day window is now 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 in a flat market. Bitcoin closed 11.96% above the 200-day average, 11.53% above the 50-day, 17.47% above the 100-day and 0.52% below the 20-day — a twenty-second consecutive close above the 200-day in the run that began on August 19.

One further point on the crossing, because it is the sort of thing that gets miscounted. Tuesday’s was the thirteenth golden cross this series contains, and the base rates quoted on the signal board rest on the twelve prior ones — 2012, 2014, twice in 2015, 2019, twice in 2020, 2021, twice in 2023, 2024 and 2025. The series itself begins on July 17, 2010, but the 200-day average cannot be computed until February 1, 2011, and on that first date the 50-day is already above the 200-day; the record therefore opens mid-uptrend rather than at a crossing, which is why a count taken from this data will differ by one from counts taken from series that start elsewhere. CoinDesk’s own tally, published on Tuesday, counts twelve occurrences, finds an average three-month gain of 24.9% across nine measurable instances, and notes that only three of the twelve stayed valid a full year — its summary being that the signal has “triggered premature bull traps 3x more often than sustained multi-year rallies.” Nothing about Wednesday changes that; a gap widening from $89 to $272 tells you the two averages are separating, not that the next quarter is decided.

Levels being discussed (attributed)

Level	Why it is being watched	Who cites it
$100,000	Named as a possible destination if a weekly close above $83,000 prints first; the condition is the whole of the claim	Ted Pillows, via CoinCentral (Sep 9)
$90,000	Described as the higher objective above the immediate resistance band	CoinStats daily analysis (Sep 9)
$83,000	The weekly-close level that the $100,000 projection is conditioned on	Ted Pillows, via CoinCentral (Sep 9)
$82,800	The May peak from which price retreated to $78,300 — the reference point for the current retest	CoinTurk (Sep 9)
$82,283	The September 3 high and the month’s high, the highest print since May 11	Alpha Vantage feed
$80,500–$82,000	Primary resistance band on the day’s published analysis	CoinStats daily analysis (Sep 9)
$81,719	The 365-day moving average, computed here	This brief’s calculation
$80,886	Upper Bollinger Band, 3.3% above the close and $853 lower than yesterday	This brief’s calculation
$80,000	Round number, now given back on a closing basis for a third consecutive session; the level Wednesday’s coverage framed the day around	Widely cited; Cointelegraph (Sep 9)
$79,870	50-week simple moving average, computed here on the current, still-open weekly bar; the close is $1,586 below it	This brief’s calculation
$79,753	Wednesday’s high, the highest since September 7	Alpha Vantage feed
$79,500–$80,000	“The region bulls aim to recover”	CoinTurk (Sep 9)
$78,693	20-day simple moving average, $409 above the close — crossed on Wednesday for the first time since August 16	This brief’s calculation
$78,300	“The retest of ~$78,300 is now in progress”; a weekly close below it followed by a bearish retest “would likely confirm a breakdown”	Rekt Capital, via CoinCentral and CoinTurk (Sep 9)
$78,000–$78,900	Primary support band on the day’s published analysis	CoinStats daily analysis (Sep 9)
$77,727	Wednesday’s low on the 00:00 UTC candle used throughout this brief	Alpha Vantage feed
$77,600	The intraday low quoted in exchange-feed coverage, $127 below the candle low above; a reminder that “the low” depends on which feed is read	CoinTurk (Sep 9)
$77,500	Described as “critical support”	CoinTurk (Sep 9)
$77,100	“Support area that attracts new demand”	Bitfinex, via CoinTurk (Sep 9)
$77,054	The August 22 close — the lowest inside the current 20-day Bollinger window, and one of the two numbers now setting band width	This brief’s calculation
$76,500	Lower Bollinger Band, $1,380 higher than yesterday	This brief’s calculation
$76,350	The average cost basis of active Bitcoin holders; the same analysis puts 71% of circulating supply in profit	Bitfinex, via CoinTurk (Sep 9)
$76,219	The September 2 low — the floor of the month’s range, $2,065 below Wednesday’s close	Alpha Vantage feed
$76,000	Secondary support on the day’s published analysis	CoinStats daily analysis (Sep 9)
$72,817	200-day exponential moving average	This brief’s calculation
$72,000–$74,000	The zone named as where a deeper decline would be watched; also the “major support” band in the day’s published analysis	Michaël van de Poppe, via CoinCentral (Sep 9); CoinStats (Sep 9)
$70,192 / $69,919	50-day and 200-day simple moving averages — the two lines that crossed on Tuesday, now $272 apart	This brief’s calculation
$65,185	Long-term moving-average support described as preserving the bullish structure	Jesse Olson, developer of Markets Sniper, via CoinTurk (Sep 8)

One note on the table. Fifteen of these twenty-eight entries are dated September 9 — eight more are this brief’s own calculations and four come straight off the price feed — and the day’s commentary clustered even more tightly around the price than Tuesday’s did. Rekt Capital’s $78,300 is now an active retest rather than a warning, and the close landed $16.02 underneath it — near enough that the level is neither support nor resistance so much as the price itself, $573 above Wednesday’s low; Bitfinex’s $77,100 demand area and $77,500 “critical support” bracket the low from underneath; and CoinStats’ $78,000–$78,900 band contains the close outright. Stripping out the conditional $100,000 projection and the long-horizon entries, the working band analysts actually named on Wednesday runs from $77,100 to $82,800 — Bitfinex’s floor and CoinTurk’s May-peak ceiling — which is $5,700 wide, or about 2.53 times Wednesday’s ATR. The bracket Wednesday’s own trade defined is much narrower: $77,727 to $79,753, a span of $2,026, or 0.90 times ATR. And the tightest bracket of all is the one the Bollinger Bands now draw, $76,500 to $80,886, a $4,386 envelope against $6,619 a day ago and $10,351 two days ago.

Momentum and volatility (charts 2 and 3)

Momentum kept leaking at the same steady rate it has all week. The MACD line fell to 2,615.14 from 2,843.71 while the signal line eased to 3,115.23 from 3,240.25, widening the histogram to −500.09 from −396.55 — a sixth session below signal and a sixth consecutive widening, the sequence now running −49.50, −108.96, −139.48, −261.93, −396.55, −500.09. The $500 gap between line and signal is the widest of this leg down, and the MACD line remains 2,615 above zero, so this is a loss of upward momentum rather than a change of trend direction. The mechanism is unchanged and worth restating once: the twelve-day exponential average still carries the memory of September 3’s $81,264 close and is shedding it faster than the twenty-six-day average is, so the two keep separating even on days when price barely moves. Of 637 earlier days in this series with the histogram negative and widening, 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%) — which is to say, almost exactly nothing. RSI fell 0.66 points to 59.45, a third consecutive decline and still the lowest since August 18. ADX fell a second straight session, to 46.90 from 47.22, with +DI easing from 31.70 to 30.54 and −DI from 13.07 to 12.23, narrowing the directional spread from 18.63 to 18.31 — its third consecutive narrowing. Because ADX is a fourteen-period smoothing of that spread, it turns last, and it has now turned. A falling ADX from a high level describes a strong trend losing intensity, not a trend reversing.

The volatility reading landed exactly where Tuesday’s edition said it would, and the landing is the news. Bollinger band width contracted for an eleventh straight day, to 5.57% from 8.44% — a fall of 2.87 points, the smallest of the last five contractions, whose daily falls read 0.70, 1.03, 1.72, 2.16, 2.16, 3.30, 4.43, 4.52, 6.74, 4.84 and 2.87. From the 40.03% peak of August 29, eleven sessions ago, band width is down to 13.9% of its value, a fall of 34.46 points. Tuesday’s brief projected that August 20’s $73,011.87 close leaving the twenty-day window would put band width near 5.6% on a flat session; it printed 5.57%, and the small overshoot is explained by price not staying flat. That projection has now been made three days running and has been right three times, and it is worth saying plainly why it stops working here. The twenty-day window now runs from August 21 to September 9. Its lowest close is August 22’s $77,054.44 and its highest is September 3’s $81,263.99, a spread of $4,210 — narrow enough that no single member of the window is doing the work any more. The standard deviation of those twenty closes is $1,096.59, and rolling the window forward on an unchanged close puts it at $1,097.32 and band width at about 5.58%. There is nothing left to expel. From Thursday, in other words, a move in band width means the market moved, which has not been true since August 29 — and that makes the reading useful again in a way it has not been for eleven sessions. The other volatility measures stayed quiet. ATR fell a fifth day, by $17.25 to $2,250.03, or 2.87% of price, on a range 0.893 times Tuesday’s. %B fell to 0.4068 from 0.5027, its lowest since August 16, in a band now running $76,499.78 to $80,886.14 — a $4,386 envelope against $6,619 a day ago. Thirty-day realized volatility fell 0.50 points to 46.75%, its first decline in three days and still, notably, higher than it was when band width was seven times as wide.

Since a squeeze this pronounced invites a particular claim — that a big move must follow — it is worth testing that claim against this series rather than repeating it. Take every occasion since 2016 on which a close first pushed band width below 6% after a wider spell: there are 24 such entries with ten sessions of data behind them, and the median absolute move over the following ten sessions was 2.49%, against 5.53% for all days in the same period. Only a third of them moved more than 5% either way, against just over half of all days. In other words, squeezes in this record have been followed by calmer ten-day periods than average, not more violent ones. Extend the horizon to twenty sessions and the picture reverses: the median absolute move is 10.12% against a 8.93% baseline, and 62% exceed 5%. The honest reading of that is not “the squeeze predicts a big move” but something narrower and more useful — compression has historically persisted for a while before it resolves, and the resolution, when it comes, has tended to be larger than usual. It also says nothing at all about direction, which is the part the popular version of the claim quietly assumes. The last six squeeze entries in this series, most recent first, went −0.1%, +1.6%, +3.3%, −1.1%, −10.4% and −6.7% over the following ten sessions, in no discernible pattern.

Flows and positioning (chart 4)

For a fifth consecutive edition chart 4 is unchanged, and the reason has now crossed from routine into mildly unusual. The Saturday, Sunday and Labor Day Monday gaps were absences of a session. Tuesday’s gap was the ordinary one-day publication lag. But Wednesday’s brief finds neither Tuesday’s nor Wednesday’s total published on the trackers it reads — Farside Investors’ per-fund table, TFTC’s dataset and SoSoValue’s dashboard were each either unreachable or showed no figure past September 4 when this was written — and two business days without a print is a day longer than the normal cadence. No explanation for that is available from outside, so none is offered; it is recorded as a gap rather than dressed up as a signal. Friday’s +$174.6 million therefore remains the most recent print, with BlackRock’s IBIT accounting for +$117.4 million of it, about 67%, Fidelity’s FBTC for +$57.2 million, and every other US spot Bitcoin ETF for exactly zero. September’s four reported sessions stand at about +$770 million against August’s completed +$3,524 million. When the backlog clears it will land as several days of accumulated brokerage-channel orders in one or two figures, arriving into the two inflation reports that close this week, which will make those numbers harder to read rather than easier. The medium-term flow picture reported earlier in the week is unaffected by the reporting gap: Crypto Briefing and Cryptonomist both put net inflows over the three weeks to September 8 at roughly $3.8 billion, the strongest three-week stretch of 2026, with one week approaching $987 million. Set against a market 0.35% lower on the month, that remains the tension worth holding — heavy net demand meeting a price that has gone sideways.

The derivatives picture improved in one respect and got murkier in another. The improvement is that Tuesday’s missing liquidation total has arrived: CoinStats’ page, stamped “09 Sep, 2026 00:08”, covers the twenty-four hours into the start of Wednesday, which is Tuesday’s UTC session, and reports $33.48 million of crypto liquidations, $28.08 million long (83.9%) against $5.40 million short. That is nearly six times the $5.75 million the same source reported for the equivalent window a day earlier, and because it is the same source, the same convention and the same window length, it is a genuine comparison rather than one of the cross-tracker apples-to-oranges figures this brief has had to caveat all week. It is also, in absolute terms, still small: $33 million of forced closures on a day with a $1,888 range says leverage was tested but not stressed. No total for Wednesday’s own session had been published anywhere reachable when this brief was written. The murk is in funding. The same page quoted +0.0043% per four hours, where a day earlier it quoted +0.0021% per eight. On the ×6×365 convention a four-hour interval implies, +0.0043% annualizes to about +9.4%; on the ×3×365 convention used for the eight-hour figure, yesterday’s annualized to +2.30%. Whether funding actually rose fourfold or the page simply changed how it labels the interval cannot be determined from outside, so no change is claimed and the reading is reported with its stated interval attached. Open interest is quoted at about $53.27 billion against the $52.96–53.01 billion range shown a day earlier, and for a seventh consecutive day no day-over-day change is claimed for the same reason: that tracker’s own numbers have not been internally consistent this week. On the wider market, Bitcoin’s share of total crypto value was about 59% on CoinStats’ measure — a different construction from the tv-hub.org figure of 58.3% carried yesterday, and quoted here as its own reading rather than as a continuation of that series.

On-chain and network

Wednesday’s on-chain story was a single metric read two ways, which makes it a good one to explain properly. SOPR — the Spent Output Profit Ratio — divides the price at which coins move on-chain by the price at which those same coins last moved. Above 1.0 means the average coin changing hands is doing so at a profit; below 1.0 means at a loss. Cointelegraph, using CryptoQuant data, reported that Bitcoin’s aggregate SOPR has held above 1.0 continuously since August 19 — the day of the breakout — which is the longest such run of 2026 and is now in its fourth week, currently reading 1.002. The constructive reading came from Checkonchain: “In bull markets, short sharp moves below break-even tend to become buy-the-dip setups. The current structure is starting to look more like those early bull-market recoveries.” The cautious reading came from David Puell, a portfolio manager at ARK Invest, who was unmoved: “In our view, as of now, we leave it as a downside risk,” adding that he would want SOPR to hold above 1 for longer, alongside a series of higher highs and higher lows on the weekly chart, before changing his position. CryptoQuant’s chief executive Ki Young Ju has previously described the bear market as over on the firm’s own indicators. All three are looking at the same series; what separates them is how much of a streak counts as evidence, which is a judgment rather than a measurement. Bitfinex’s contribution to the same picture is arithmetic rather than interpretive: it put 71% of circulating supply in profit and the average cost basis of active holders at about $76,350 — $1,934 below Wednesday’s close, and inside the band the month’s low already tested.

On the network itself, difficulty is unchanged at 127.45T and will stay there until the next 2,016-block boundary at height 967,680, now 1,409 blocks away with the chain at 966,271 and the cycle 607 blocks in. CoinWarz did not display a headline retarget estimate when its page was read for this edition, so the +2.55% figure it was publishing on Tuesday is the most recent one this brief can attribute; what the site did show is a recent average block time of 9 minutes 44 seconds against the ten-minute target, which on its own implies an adjustment of about +2.7% and a retarget around September 19. A positive estimate means blocks are arriving faster than target, which is what a rising or steady hashrate produces. The hashrate figure itself remains unusable: the site displayed 1.07 zettahashes per second at block 966,271, while its own chart page was showing 932 exahashes per second at block 966,101 — a 15% difference across 170 blocks, which is not a real change in the network but an artefact of inferring hashrate from block intervals over short windows. This brief has declined to quote a hashrate reading for three editions now and continues to; the September 19 retarget will provide a number that is measured rather than inferred, because difficulty adjusts on what actually happened over 2,016 blocks. Elsewhere on the corporate and policy side, Block — Jack Dorsey’s company — applied to the Office of the Comptroller of the Currency for a national trust bank charter, proposing “Builders Bank” to provide federally supervised custody for Bitcoin and stablecoins, consolidating work it currently runs under more than fifty state money-transmitter licences; the entity would not take deposits or make loans. Germany’s finance ministry drafted a 25% tax on crypto gains from 2027, with anything bought before then keeping the existing twelve-month exemption, and Italy’s central bank ordered crypto service providers to screen transfers for sanctioned entities. None of these is a same-week price input; together they are the slow accumulation of regulated infrastructure and regulated cost that has characterised this cycle.

Sentiment

The index this brief normally leads with is missing today, and saying so is more useful than substituting for it. 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 on this run, so there is no September 9 entry on that measure here and none has been estimated. CoinStats publishes its own Fear & Greed index, constructed differently, and it showed 70 (“Greed”) for the window into September 9 against 72 the day before on the same page. Two things follow. First, 70 is not comparable with yesterday’s 69: they are different indices, and reading a one-point rise into that pair would be an error of exactly the kind this brief spends its time trying to avoid. Second, the same page’s own history is internally odd — it shows 50 (“Fear”) for “last week”, where alternative.me’s dated series had 69 for September 1 — which is a further reason to treat the two as separate measures rather than as one series with a gap. The structural caveat that applies to both survives the source change: these indices blend volatility, momentum, market dominance and survey components rather than tracking price, and a collapse in volatility of the kind chart 3 shows raises them mechanically. Any greed reading printed during this squeeze is flattered by the squeeze itself.

Of the 28 Bitcoin-tagged articles in the Alpha Vantage news feed published between 00:00 UTC on September 9 and 04:00 UTC on September 10 — up from 23 on Tuesday — the feed’s sentiment model scored none as bullish, eight as somewhat bullish, fifteen as neutral, four as somewhat bearish and one as bearish. Sourcing was thirteen from Cointelegraph, eight from Decrypt, four from Motley Fool, two from CNBC and one from Zacks. The distribution is slightly more positive than Tuesday’s, but the composition undercuts a market reading of it in the same way it did then: the single bearish score attached to Decrypt’s report that hackers breached the email provider used by hardware-wallet maker Trezor and sent a fake security alert, and the somewhat-bearish four were a crypto-mysteries feature, a Mexican homicide case involving a cold wallet, a $245 million theft guilty plea and a former Silvergate chief executive’s account of his bank’s closure. Not one of the five negative scores is about market conditions. On the other side, the eight somewhat-bullish scores include two earnings-call transcripts and a memecoin explainer, which is a reminder that 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. The one item genuinely about price — Cointelegraph’s account of Wednesday’s session — scored neutral overall and somewhat-bearish on the Bitcoin ticker.

Published views, in both directions. The constructive case on Wednesday was conditional rather than declarative, which is itself worth noting after Tuesday’s golden-cross projections. Ted Pillows set an explicit trigger: “If Bitcoin manages a weekly close above $83,000, a pump to $100,000 could happen this year” — a claim whose entire content is the condition, since the close in question is $4,716 above where the market is. Checkonchain, on the SOPR data, said the structure “is starting to look more like those early bull-market recoveries.” Bitfinex described the tape as “being driven by spot buying” rather than leverage, and Avinash Shekhar, co-founder of the derivatives exchange Pi42, framed the same point as a market that “can absorb pressure, hold support, and recover without excessive leverage” — an argument about the quality of the bid rather than its size. Against that, David Puell of ARK Invest declined to change his reading of the on-chain data — “In our view, as of now, we leave it as a downside risk” — and Michaël van de Poppe named $72,000–$74,000 as the zone he would be watching on a deeper decline, roughly 5.5% to 8% below the close. Rekt Capital’s condition, first set out on Tuesday, moved from prospective to live: “The retest of ~$78,300 is now in progress,” with a weekly close below it followed by a bearish retest the thing that “would likely confirm a breakdown.” That is still the most falsifiable statement anyone has made this week, and Sunday’s weekly close is when it is tested. Beneath all of it the macro voices remain the bearish ones, on the argument that a Federal Reserve that raises rates next week lifts bond yields — which set fresh fifty-two-week highs on Wednesday — and redirects capital away from an asset that pays no yield.

Macro and cross-asset

Wednesday was Tuesday’s session with the volume turned up. The S&P 500 closed at 7,636.36, down 0.48%; the Nasdaq Composite at 26,253.34, down 0.64%; the Dow at 52,380.66, down 0.77%; and the Russell 2000 at 2,921.72, down 1.30%, the widest decline and the usual signature of a session where rising rates do the damage. The Cboe Volatility Index rose 4.58% to 16.44. Gold was effectively unchanged at $4,442.30 — the day’s market wrap quoted that as up 0.07%, but it is $3.30 below the $4,445.60 mark this brief carried for Tuesday, so the sign of the move depends on which of the two marks is right; the level is reported as published and no direction is claimed. The move was in energy and rates: US strikes on Iranian oil tankers took Brent above $101 a barrel for the first time since late July and October WTI up 3.93% to $96.69, while the ten-year Treasury yield rose about six basis points to 4.85% and yields across the curve set fresh fifty-two-week highs, helped along by a Treasury buyback announcement. The dollar leg of it ran through Japan: the yen weakened to 153 per dollar, its softest since February, after Treasury Secretary Scott Bessent said of intervention that “I am the house now” — remarks that Saxo’s chief investment strategist Charu Chanana read as adding to carry-trade risk, noting that “the carry trade is vulnerable because this unwind is happening before the BOJ has even delivered its expected hike.” A carry trade here means borrowing in a low-rate currency, the yen, to buy higher-yielding assets elsewhere; an unwind forces those positions to be sold and the borrowed currency bought back, which is why a yen move is a risk-asset event and not just a currency one.

Bitcoin’s own 0.21% decline was the mildest of the risk assets on the day, and that relative resilience — down less than small-cap equities, down less than the Nasdaq, on a day when oil rose 4% and yields hit year highs — is the most defensible bullish observation available about Wednesday. It is also a single session, and the same comparison ran the other way on Tuesday. At Wednesday’s marks Bitcoin buys 17.62 ounces of gold, against 17.65 on Tuesday — a ratio that fell because Bitcoin fell slightly more than gold did on the marks used here, subject to the caveat above about those marks. The policy question is unchanged and now four working days away: the FOMC meets September 15–16 with the target range at 3.50%–3.75%, and CME futures were pricing roughly a 60% chance of a 25 basis-point increase. Wednesday’s oil move feeds that directly, because higher energy prices push headline inflation up, and the two reports that stand between now and the meeting are Thursday’s August producer prices and Friday’s August consumer price index. The transmission to Bitcoin is the one the coverage keeps making — higher policy rates raise the yield on dollar cash and reduce the relative appeal of an asset that pays none — and it explains why a market whose trend indicators improved for a second straight day still closed lower.

Calendar: what’s ahead

Date (ET)	Event	Why Bitcoin traders watch it
Thu, Sep 10, morning	US spot ETF flow prints for Tuesday and Wednesday	Two business days of orders are now outstanding; the backlog will arrive compressed into one or two figures
Thu, Sep 10, ~8:15 a.m.	European Central Bank rate decision and press conference	The euro side of the dollar index, which has tracked Bitcoin inversely through this run
Thu, Sep 10, 8:30 a.m.	August producer prices and weekly jobless claims	The first of the two inflation prints before the FOMC, and the first read on whether a $96 oil price is reaching the pipeline
Thu, Sep 10, 8 p.m.	First Bollinger reading not distorted by the window	The 20-day frame has finished shedding the August breakout; from this session a move in band width means the market moved
Fri, Sep 11, 8:30 a.m.	August consumer price index	The report the coverage describes both Bitcoin and the S&P 500 as waiting on; four days before the FOMC convenes
Fri, Sep 11, 10:00 a.m.	University of Michigan consumer sentiment, preliminary	Carries the inflation-expectations series that a $96 oil price would show up in first
Sun, Sep 13, 8 p.m.	Weekly candle closes (00:00 UTC Monday)	The frame Rekt Capital’s $78,300 breakdown condition is defined on, and the next test of the 50-week average at about $79,870
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 Wednesday analysis set out what a failure would mean: the bill could be 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 roughly a 60% chance of a 25 basis-point increase on Wednesday; 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 is now the more interesting of the two, given Wednesday’s move to 153 yen per dollar
Sat, Sep 19	Next Bitcoin difficulty adjustment	CoinWarz’s most recent published estimate was +2.55%; its stated 9-minute-44-second average block time implies about +2.7%. A measured retarget also settles the week’s contradictory hashrate readings
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 0.35% lower so far

Reading the combination

The useful thing about Wednesday is that it retires an explanation. For three editions this brief has insisted that the collapse in Bollinger band width — 40.03% to 5.57% in eleven sessions — was mechanical rather than meteorological: old outliers leaving a twenty-day window rather than a market going quiet, with thirty-day realized volatility finishing the eleven sessions higher than it started them, at 46.75% against 44.03%, as proof. That explanation was correct, it was checkable, and it is now spent. The twenty-day window runs from August 21 to September 9, its extremes are $81,263.99 and $77,054.44, and rolling it forward on an unchanged close leaves band width at about 5.58% against Wednesday’s 5.57%. There is nothing left to expel, which means Thursday’s reading is the first in eleven sessions that carries information about the market rather than about the calendar. That is a small thing to be able to say precisely, and it is the whole value of having tracked the mechanism rather than the number. The layers otherwise pulled apart further. The trend layer improved for a second day: the golden cross widened from $88.65 to $272.36, the 50-day rose for a twenty-ninth consecutive session, and the close was a twenty-second straight above the 200-day average. The momentum layer deteriorated for a sixth: the MACD histogram widened to −500.09, the widest gap of this leg; ADX fell a second day as the directional spread narrowed a third; RSI fell a third; and price closed under its 20-day average for the first time since August 16, ending a twenty-three-session run. Both are true, and both are ordinary — a market can lose short-term momentum inside a structure that is still turning up on the long frame, and the two indicators are reading windows an order of magnitude apart. On the base rates the board is genuinely split, and the split is informative rather than annoying. For: 185 earlier days with band width at or below 6% and price above its 200-day average were higher ten sessions later 69% of the time (median +1.68%) and twenty sessions later 73% (median +5.87%); 240 days with ADX at or above 45, +DI above −DI and ADX falling were higher 72% of the time at ten sessions (median +5.05%). Against, or rather neutral: the 21 earlier entries into a sub-6% band width above the 200-day were higher ten sessions later 52% of the time (median +0.84%); 637 days with a widening negative MACD histogram above zero, 55% (median +0.84%); 436 days with RSI between 57 and 62 above the 200-day, 55% (median +1.20%). Every one of those is within a few points of a coin flip except the two that lean positive, and both of those cluster in strongly trending years, which is a real caveat and not a formality. Three things are worth holding separately. First, the popular claim that a squeeze predicts a big move is not what this record shows: since 2016, the twenty-four squeeze entries with ten sessions of data behind them produced a median absolute ten-day move of 2.49% against a 5.53% all-day baseline, with the excess only appearing by twenty sessions — compression has persisted before it resolved. Second, Wednesday’s decline was macro-shaped and mild: oil up 3.9%, Brent through $101, ten-year yields at fifty-two-week highs, the yen at 153, small-caps down 1.3%, and Bitcoin down 0.21% — the least of the risk assets, which is one session of evidence and not a pattern. Third, the level structure is now the tightest of this entire run, and the close sits almost exactly on the level everyone is watching: Rekt Capital’s $78,300 is $16.02 above it, with Wednesday’s $79,753 high, the 50-week average at $79,870, the $80,000 round number and the upper band at $80,886 stacked above that; below sit Wednesday’s $77,727 low, Bitfinex’s $77,100 demand area, the lower band at $76,500 and September 2’s $76,219 monthly floor. The bracket from Wednesday’s low to the round number runs $2,273, about 1.01 times a single day’s ATR. One ordinary session can resolve it in either direction, and nothing on this board says which.

Sources

Price, volume and all technical indicators: Alpha Vantage BTC/USD daily series (00:00 UTC candles), computed by the author; the September 9 candle was confirmed complete by the presence of a partial September 10 row. Wednesday’s session narrative, the oil and yen moves, and Saxo’s Charu Chanana: Cointelegraph’s September 9 market report. SOPR, Checkonchain and ARK Invest’s David Puell: Cointelegraph on the SOPR streak. Rekt Capital’s live retest, Ted Pillows’ $83,000 condition and Michaël van de Poppe’s $72,000–$74,000 zone: CoinCentral. Bitfinex’s cost-basis and supply-in-profit figures and Pi42’s Avinash Shekhar: CoinTurk. Funding, open interest, liquidations, dominance and the CoinStats Fear & Greed reading: CoinStats daily analysis and CoinStats’ index page, both used with the timing and construction caveats described above. Equity, oil, gold and volatility closes and the ten-year yield: the September 9 market close wrap, with intraday commodity levels and the fifty-two-week yield highs from TheStreet’s live blog. Fed rate-increase odds: Yahoo Finance, citing CME Group. Block’s trust-bank application: Decrypt and Cointelegraph. Germany’s draft crypto tax: Decrypt. Italy’s sanctions screening: Cointelegraph. The CLARITY Act scenarios: Cointelegraph. The Trezor email breach: Decrypt. Network data: CoinWarz hashrate. ETF flow history: Farside Investors and the TFTC-sourced flow tracker, neither of which showed a figure past September 4 when this brief was written. News sentiment tally: Alpha Vantage NEWS_SENTIMENT feed for September 9.