United States forces struck and sank five Iranian oil tankers on Tuesday night, and on Wednesday Iran said it had attacked ten ships near the Strait of Hormuz and fired twenty ballistic missiles at a base in Jordan, of which Jordan said it intercepted eighteen, Al Jazeera reported on 9 September. Central Command named the tankers — Kaviz, Charminar, Horizon 1 and Riesco in the Gulf of Oman, Derya off Kharg Island — and said crews had been directed to abandon ship before the strikes. Brent crude, which settled at $97.92 on Tuesday, ended Wednesday at about $101, its first settlement above $100 since July, per TheStreet’s market wrap. The S&P 500 closed at 7,636.36, down 0.48%; the Dow at 52,380.66, down 0.77%; the Nasdaq at 26,253.34, down 0.64%; the VIX at 16.46. On the Treasury’s constant-maturity curve the 2-year closed at 4.43%, the 10-year at 4.83% and the 30-year at 5.28%, each up 3 to 4 basis points on the day from Tuesday’s 4.39, 4.80 and 5.25.

Bitcoin closed Wednesday 9 September at $78,281.65 on Bitstamp, down 0.213% on the day after Monday’s −1.554% and Tuesday’s −0.811%: three consecutive declines, the twenty-first run of three or more this year, for a three-day fall of $2,057.33 or 2.561% from Sunday’s $80,338.98. Tuesday’s low of $77,602.59 was the lowest print since 3 September’s $76,940.96. Wednesday’s range of $1,981.64 — high $79,741.64, low $77,760.00 — ranks 112th-narrowest of 252 completed sessions in 2026, the middle of the distribution; its volume of 1,787.66 BTC was 146th-lowest. The month is now −0.368% against 31 August’s $78,571.17 and the year −10.531% against $87,496.00. At the desk’s 06:10 UTC partial on Thursday the price was $78,382.35, up 0.129%. The desk did not publish on Tuesday or Wednesday; this piece covers both days.

The angle the mainstream tape offers, tested

CoinDesk wrote at 10:47 UTC on Wednesday that bitcoin was trading “with gold, not equities” as Brent broke $100, with bitcoin up intraday alongside gold while European stocks fell — the inverse, it noted, of 2 September, when strikes sent bitcoin down with equities. On the closes the desk grades on, the claim is weaker than the intraday snapshot. Bitcoin’s Wednesday close was −0.213%; the S&P’s was −0.48%; gold futures, per the Yahoo Finance quote strip TheStreet cites, finished about +0.16% at $4,446. That is bitcoin between the two, closer to gold than to stocks on the day, but on Monday and Tuesday — −1.554% and −0.811% against the S&P’s −0.58% Tuesday — it fell harder than equities. Three sessions do not establish a regime, and the desk’s five-month study found none; the honest reading of this week is that bitcoin has tracked the front end of the Treasury curve, which has repriced for a hike, and not oil or the war directly.

That front end is the story. The 2-year has moved from 4.37% on 4 September to 4.43% on Wednesday; Central Bank Watch’s futures-implied probability of a 25-basis-point hike on 16 September stood at 59.4% on its 9 September page, and CoinDesk’s Monday live blog put the CME FedWatch figure at 60% after Friday’s payrolls and a 9.5% rise in oil this month. The Federal Reserve is in its pre-meeting blackout until 17 September, so nothing official moves the number before the decision; producer prices at 12:30 UTC today — consensus +0.4% headline and +0.3% core for August per TradingKey’s preview — and Friday’s CPI are what can. Marker N1 requires 60% or better on Friday evening from two major outlets; at 59.4% on one source it is inside a percentage point.

Leverage: the window low, and a basis that keeps rising while funding falls

Binance BTCUSDT perpetual open interest at 00:00 UTC on Thursday was 105,105.968 BTC, down 1,521.38 (−1.427%) from Wednesday’s 106,627.347. That is the lowest of the 31 daily snapshots from 11 August to 10 September, below the 105,531.279 printed on 24 August that Monday’s piece named as the window minimum, and 7,611.58 coins (6.753%) below the 112,717.545 of 4 September. The path since Monday was not one-way: Tuesday’s snapshot rose 1,332.40 to 106,876.110, Wednesday’s slipped 248.76, and then Thursday’s fell through the floor. The dollar value of the position fell from $8,362.4 million to $8,227.2 million, an implied mark of about $78,275, consistent with the spot print. The live reading at 06:10 was 104,900.571. Marker G1 — a September 00:00 print below 100,000 — is 5.11% away, closer than at any point since it was set.

Funding settled at 0.009012% at 08:00 on Wednesday, 0.004815% at 16:00 and 0.003562% at 00:00 Thursday — 3.900% annualised on the last print. In the 125 settlements from 16:00 UTC on 30 July, one has been negative (5 September 08:00), 26 sat at the exchange’s 0.0100% default, and the mean is 7.005% annualised. Meanwhile the September quarterly on COIN-M traded at $78,568.0 against an index of $78,355.13 at 06:10 — a 0.2717% premium over the 15.08 days to expiry on 25 September, or 6.577% annualised, up from 5.266% on Monday and 2.875% on Saturday; the December contract at $79,444.7 implies 4.785%. Funding is now 268 basis points below the September basis, wider than Monday’s 221. Field Guide #37 is the desk’s reading of that shape: leverage leaving the perpetual while the dated contract is bid is not a market being liquidated, it is one whose fast money is leaving and whose slower money is paying more for a fixed-date position. Marker G2 (September basis above 8.00% on 25 September) is 142 basis points short, from 273 on Monday. Global long/short accounts moved 1.0602 → 1.1608 → 1.2952 → 1.2707; the top-trader position ratio 2.0219 → 2.0050 → 2.1312 → 2.1911, the highest of the fifteen rows since 27 August.

The ETF table: two outflow days, and September still positive

Farside’s table now has Tuesday and Wednesday. Tuesday 8 September: IBIT +10.7, FBTC −17.1, BITB +14.5, ARKB +8.1, BTCO −4.7, MSBT +7.4, GBTC −65.5, the other five columns zero, total −$46.6 million. Wednesday 9 September: IBIT −19.5, ARKB −78.0, MSBT +4.5, GBTC −27.2, eight zeros, total −$120.2 million. The first back-to-back outflow days of September, whose only previous red row was 1 September’s −$236.5 million; the week so far is −$166.8 million, against +$986.7 million for the whole of last week. September’s running total is +$603.2 million over six of its twenty-one sessions (−236.5, +101.1, +730.8, +174.6, −46.6, −120.2), so marker H2 (a positive September) remains open and positive but has lost $166.8 million of its cushion. IBIT’s September is +$361.4 million.

Rates: H3 is passing, I1 is waiting, O1 has its base tonight

The Treasury file, absent on Monday, has two rows. Marker H3 requires the 30-year at or above 5.25% on 17 September; it printed 5.25% on Tuesday and 5.28% on Wednesday, so it is passing by 3 basis points with six sessions to go, after failing by 1 on 4 September. Marker I1 requires the 10-year real yield at or above 2.50% on any close to 16 September; the real curve shows 2.43% on Tuesday, unchanged from 4 September, and 2.46% on Wednesday, posted after the desk’s first pull. Run the split Field Guide #39 exists for and Wednesday’s 3-basis-point nominal rise was entirely real: the 10-year breakeven, nominal less real, was 2.37% on both days. I1 is 4 basis points away with four sessions to go. Marker O1 is defined as the 2-year’s Friday close being at least 5 basis points above its Thursday close; Thursday’s close, after the PPI print, is the base, and Wednesday’s 4.43% is not it. TheStreet called Wednesday’s 4.843% 10-year and 4.427% 2-year 52-week highs; the desk grades on the Treasury’s 4.83 and 4.43, as Field Guide #42 explains.

Mining: the projection is now quotable

The difficulty period that began at block 965,664 on Saturday evening had 648 of its 2,016 blocks at our pull, 32.14%, at an average interval of 9.78 minutes; mempool.space projects a retarget of +2.404% at block 967,680, estimated for about 13:10 UTC on 19 September. That is above the desk’s quoting threshold of a quarter of the period, so marker P1 (a positive retarget) is on course. The hashrate series, whose rows are stamped at the end of the day they measure, prints 881.29 EH/s on the 8 September row, 857.91 on the 9 September row and 1,018.89 on the row stamped today; the 31-row mean is 913.25 and the high remains 1,053.07 on the 31 August stamp. Difficulty stands at 127.45 trillion. Over the last week Foundry mined 266 of 1,044 blocks (25.5%), AntPool 193 (18.5%), F2Pool 156 (14.9%), SpiderPool 100 (9.6%), ViaBTC 81 (7.8%), SECPOOL 60 (5.7%) and MARA Pool 50 (4.8%); every fee tier on mempool.space read 1 sat/vB at the pull. The Fear & Greed index printed 69 on Thursday after 66 on Wednesday, 69 on Tuesday and 71 on Monday: the twenty-second consecutive Greed reading since 20 August, the day after 19 August’s 46.

The markers, walked

Nineteen were open on Monday. M1 is graded: fail. Strategy’s 8-K of Monday 8 September states that between 31 August and 7 September the company “did not purchase or sell any bitcoin”; the bar was at least one coin bought in the week of 31 August to 6 September. Today’s update has the filing. T1, the Liquid return, has 3,400 of its 3,500 BTC and settles Sunday; today’s lead grades it early as failing unless 100 more coins move. V1, a Liquid restart with a federation peg-out by 30 September, is harder after Wednesday’s three-phase plan, which puts peg operations last and conditions them on “fund returns.” U1, Friday’s close at or above $80,318, needs +2.60% from Wednesday’s close in two sessions that contain PPI and CPI; the desk set it when the price was $20.98 above the bar and it is now $2,036.35 below. L1, $81,265.00 on 16 September, is $2,983.35 away. W1, HYPE at or above $82.60 at 00:00 UTC Sunday, stood at $83.56 on CoinGecko’s Thursday daily print, $0.96 above the bar and down from $87.95 on Monday. H2 positive, H3 passing, G1 5.11% away, G2 142 basis points short, I1 4 basis points away, N1 within a point, O1 bases tonight, P1 on course, I2 (Bank of Japan, 18 September), K1 and Q1 (the Senate’s 15 September cloture vote, which Cointelegraph reported Wednesday still needs “a handful of Democrats”), R1 and S1 (Trezor) unchanged. Eighteen remain open after M1.

The marker

As standing practice we mark one falsifiable claim on the day’s numbers. X1: Binance BTCUSDT perpetual open interest at 00:00 UTC on Thursday 17 September 2026 — the morning after the Fed’s decision — is above 105,105.968 BTC, today’s print. The reasoning for: the perpetual has already shed 6.75% from its 4 September high and sits at a window low, which leaves less to shed; the top-trader position ratio is at its window high, which is large accounts adding exposure, not removing it; and the dated basis at 6.6% says demand for exposure has moved rather than disappeared. The reasoning against: a hike priced at 60% is not a hike priced at 100%, and the three sessions after 4 September showed how quickly the perpetual can shed 7,000 coins when the print goes the wrong way; G1 is 5.1% away and a bad CPI could deliver it. If X1 passes, the September drain was rotation; if it fails alongside G1, it was exit.

Method: prices, funding, open interest, basis and on-chain figures in this article are pulled directly by Bitcoin Mastery at the timestamp stated — Bitstamp BTC/USD daily candles for closes, Binance BTCUSDT spot and USDT-margined perpetual for intraday, open interest, funding and account ratios, Binance COIN-M quarterly contracts for basis, mempool.space for difficulty, hashrate, pool shares, fees and individual transactions, alternative.me for the Fear & Greed series, Farside Investors’ table for ETF flows and US Treasury CMT par yields for rates. Where a third-party figure is cited we name the source and its date; where two sources disagree we print both. Every streak or extreme figure is published with the first date of its series in the same sentence.

Disclaimer: This article is for informational purposes only and does not constitute investment advice. Cryptocurrencies are volatile and you can lose money. Nothing here is a recommendation to buy or sell any security, digital asset or exchange-traded fund, including MSTR, STRC, L-BTC or HYPE. Do your own research and consult a licensed financial advisor before making investment decisions.