On 30 July I set out a framework for this war and called it the Consequence Ceiling. What it named was a class of target rather than a list of sites — infrastructure defined not by its size or its ownership but by a single property, that destroying it removes something the destroying party's own position depends on. Every party in this war has threatened that class. None has executed against it. All have manoeuvred around it. The assertion was simple: the omission was deliberate structure, not coincidence.

Twenty days on, the ceiling has held.

The war has spent those twenty days doing nearly everything else. The 60 day memorandum lapsed on 17 August with no deal, no extension and no scheduled talks; asked whether he would seek one, the President said no. The UAE severed all trade, commercial exchanges and financial transactions with Iran until further notice — the Gulf's principal re-export and settlement channel for Iranian trade. Mojtaba Khamenei's appointment decrees to newly named commanders ordered a shift from defensive to offensive posture, the IRGC's political deputy stating on state television that it is now the enemy which must anticipate strategic surprises. And the Wall Street Journal published an IRGC escalation menu running to undersea cable sabotage, political subversion inside Kuwait and Bahrain, and ground operations in Kuwait.

None of it inside the class.

The framework described what the parties cannot do. It did not describe what they do instead, and that is the more consequential half.

Pressure that cannot move vertically moves laterally, into domains where the terminal class consequences do not apply. The instruments are not new. What is new is that they have become the whole of the escalation.

The American record is the shorter and the more recent. On 12 August the President posted of the Strait of Hormuz that he thought the United States would keep it, on the 14th he said he would be declaring it a territory of the United States, on the 18th the claim was posted as an AI-generated graphic captioned NEW U.S. TERRITORY. Iran and Oman share jurisdiction over that water. The United States has none, and no mechanism was offered.

Set that against what it replaced — a massive attack announced on 23 July and never ordered, and a published formula committing the United States to destroy one bridge or power plant for every ship struck in the Strait, Tehran explicitly included and Tehran not hit. Within a month the escalation has migrated from the kinetic register to the semiotic one.

Alongside it: a threat to bomb the mediator, Oman having in the President's account not behaved very well, with a senator now preparing a resolution to bar military action against them. A land blockade requiring the cooperation of seven of Iran's neighbours. Seized Iranian assets committed to ship damage claims not yet incurred. And the Emirati severance was announced by the UAE's own foreign ministry after a call with Washington. The most consequential economic instrument deployed against Iran this month was executed by somebody else.

The Iranian record is longer and considerably more effective. Transit fees at 5% to 7% of declared cargo value with a penalty tier reaching 20%, and a parliamentary draft barring designated vessels outright — a permissioning regime rather than a toll regime. A refusal to let any third country de-mine, which means no reopening can be executed by any party but Iran, on Iran's schedule. Route determination separated from reopening, so that technical progress in Muscat cannot deliver the outcome it appears to be about. The waterway reclassified from bargaining chip to declared national security asset, a category from which nothing is traded. A coercion campaign through four Gulf capitals that produced the 1 August stand down by moving Gulf consent rather than American will. And a cyber operation, assessed by US intelligence as likely Iranian, against more than thirty municipal water systems in Minnesota.

Not one of those is a strike. Together they have done more damage to the American position than the campaign Iran's own published target list advertised in August.

The claim is directional, and has to be — states have always held non kinetic instruments and pointing at a sanctions regime proves nothing. What the reading predicts is a relationship: activity in these domains rises as kinetic options narrow, and falls away if the ceiling breaks and vertical escalation becomes available again. Checkable both ways, and the second half is the part that would hurt.

On 12 August a senior adviser to the IRGC commander in chief said on American television that one way is to prolong the war to the next presidential term and cause attrition, so that anyone contemplating an attack on Iran knows there is a cost. Six days later Washington was saying the same thing in its own vocabulary.

The 60 day memorandum lapsed on Monday 17 August with no deal, no extension and no scheduled talks. Asked whether he would seek one, the President said no, and he is not in a hurry. Envoys have been instructed to halt their conversations, allies have been told the strategy has changed, and the framing communicated to political allies is a shift from hammering Iran as soon as possible to strangling it over time.

There is a face saving element in that: a president who declines an extension six days after his adversary announces it intends to outlast him is partly describing his own position in terms that make continuation look chosen. But the vocabulary is naming a policy that was already running, on instruments that were already operating.

Which produces a condition this war has not been in before. Escalations get walked back, and this war has walked back several already. A stated strategy of exhausting the opponent does not get walked back, because abandoning it is indistinguishable from losing. Neither party is now describing a path to settlement.

The honest qualification is that negotiation is still happening. The Oman track is live, and Iran's foreign ministry describes the route talks as complex but continuing. What that track cannot do is deliver the outcome it appears to be about, because Araghchi has said so directly: once the route is determined, whether the Strait reopens depends on other conditions the United States must meet. Route design and reopening are two questions in Iran's account, which means progress on the first is not evidence about the second. A technical negotiation both sides can point at, and which neither expects to produce a reopening, is more stable than deadlock. It is also the mechanism by which a war with a stated end date runs for years without either party appearing to refuse to talk.

If both parties are waiting, the question becomes what the waiting costs, and one series answers it more directly than any battlefield figure. The US Strategic Petroleum Reserve fell 6.1m barrels in a week to just under 300m — its lowest since 1983, on a reported eighteen consecutive weeks of decline. Iran's stated strategy is to outlast the American position to the end of the presidential term. The reserve is the instrument on which that strategy works or fails, and it is currently moving in Tehran's favour.

The reading here is testable within weeks not years. If strangulation is a reframe, no new instrument appears and the word simply describes what was already running. If it is a genuine reorientation, new mechanisms should be observable, and quickly. That distinction resolves on its own.

On 18 August the President posted that the Strait of Hormuz is open and operating, and that all water mines have been removed or detonated. Asked to confirm the clearance, US Central Command referred the question to the White House.

The day before, an outbound bulk carrier was struck in the Strait by an unknown projectile that damaged its engine room. INTERCARGO identified the Liberia flagged Minoan Dignity and confirmed a seafarer had died; the Joint Maritime Information Center placed the vessel southbound less than a nautical mile off the Omani coast. UKMTO attributed the projectile to nobody. The cumulative figure now stands at 17 mariners killed across 65 attacks on commercial shipping since 28 February.

Six commodity vessels transited on Monday, against a ten day average of 11 and a pre-war norm near 20 a day. No very large crude carriers, no LNG tankers. That series counts laden commodity traffic tracked by Kpler, a narrower population than the all vessel figure used in the previous edition — the two are not interchangeable, and the gap between them is definitional rather than physical.

What is not definitional is the direction. Traffic has fallen following each of the last three declaratory claims about the waterway.

The obvious explanation is that shipping is a physical object and rhetoric cannot move it, and it matters that this explanation is wrong. War risk premiums, charterer decisions and insurer behaviour are all responses to perceived risk; BIMCO's head of safety and security describes the crew decision as governed by fear rather than money. A credible declaration of American control over the Strait is precisely the kind of claim that would move transit behaviour. That is the channel through which it would have worked.

It did not work because the credibility has been spent. A massive attack announced and cancelled, a bridge per ship formula published and never executed against Tehran, a South Pars threat left standing after Ras Laffan was struck — an instrument whose entire power is the capacity to make an audience act as though a claim were true stops working when the audience stops doing that.

Which gives the reading a clean test, and the measure is already running. If the count were simply deaf to words, no declaration of any credibility would ever move it. If the mechanism is exhausted credibility, the count stays flat against announcements while responding to verifiable action — a confirmed de-mining, a functioning escort regime, an insurer reclassification. Gharibabadi put the same point better: the Strait cannot be taken over by a tweet, nor by an aircraft carrier, nor by issuing a decree, nor by an election speech.

In late July I expected Beijing to move toward de-escalation. The reasoning was that China's energy cost had risen past whatever it was learning from watching, and that when the cost of waiting exceeds the benefit, a state acts. Three weeks on, nothing has moved. I priced two options where there were three. Absorb the cost, or end the war — and the third, pay the freight premium and stay out.

That is what has happened, twice. Rosatom confirmed seven Chinese permits for Northern Sea Route transits on 8 August, naming the Hormuz closure as the cause. And state controlled COSCO Shipping Energy Transportation and China Merchants Energy Shipping have kept their tankers out of both Hormuz and Bab al-Mandab since late July, following communications with central authorities — together more than 100 VLCCs, roughly 200m barrels of capacity, handling about half of China's Middle East crude before the war. The ships remain engaged on much longer routes, loading at Fujairah and Omani ports by ship to ship transfer outside the Gulf. Daily tanker margins have risen to around $110,000 from $30,000–40,000 pre-war. Beijing is buying its way around the chokepoint rather than acting on it.

The reason it need not act is a ledger. An American president has publicly asserted that an international strait can be owned, and posted a graphic to that effect through the official White House channel — against eighty years of American argument that no state may claim an international strait, which is the legal foundation of the American position in the South China Sea. Alongside it, at zero cost to Beijing: 48 THAAD interceptors shifted off the Korean peninsula, a Marine Expeditionary Unit redeployed from Japan, the Pacific left without a carrier presence, joint exercises with Seoul cut on the stated ground of a personal relationship with the North Korean leader days after Pyongyang fired ballistic missiles, five Gulf states questioning American basing, and a mutual defence pact contracted by three partners hedging against the American guarantee.

No adversary operation could have extracted that list. It was produced without one.

The class is unchanged: Gulf hydrocarbon processing at a scale that prices everyone's energy, Iranian export architecture, and desalination at population scale, where the dependency is a mass population one and the failure timescale is weekly rather than quarterly. A strike on any of it that produces the consequences the class was defined by falsifies nothing — that is the ceiling breaking, and the framework said in advance that a ceiling enforced by mutual vulnerability rather than by agreement holds only while every party reads it identically. A strike on the class that produces none of those consequences would mean the targets were mispriced, not that the restraint was misread.

Three additions since July, each with its own way of failing.

The displacement reading fails if novel domain activity is unrelated to kinetic constraint. If the ceiling breaks and vertical escalation becomes available again, the lateral instruments should thin out; if they continue at the same rate or increase, the two are independent and I have described a coincidence. Equally, a sustained fall in that activity while the ceiling holds would kill it outright.

The strangulation claim resolves faster than anything else here. If it is a reframe, no new instrument appears and the word describes what was already running. If it is a reorientation, new mechanisms are observable within weeks. That is a matter of days rather than quarters.

The exhausted instrument reading fails if the transit count is simply deaf to words. Watch what happens on verifiable action rather than announcement — a confirmed de-mining, a functioning escort regime, an insurer reclassification. If the count moves for those and not for declarations, the mechanism is credibility. If it moves for neither, I have mistaken a physical constraint for a perceptual one.

And there is an instability in the finding itself that is worth stating, because it cuts against the reassurance a reader might take from six months of restraint. A ceiling that has held through the memorandum lapsing, through the Gulf's principal settlement channel closing to Iran, and through both parties adopting attrition as declared strategy, is a real constraint. It is also a constraint with no agreement underneath it, no mechanism enforcing it, and no party committed to it beyond self interest. It has been approached and not taken, repeatedly. Nothing in that record says it cannot be taken tomorrow.

What it does say is that the reason nobody has is structural, and that the pressure which cannot go up has been going sideways for a month.