Escalation Dynamics Analyst

The structurally important fact in this window is what counts as evidence. The claim doing the most work — Saudi interceptor depletion — originates in a *Wall Street Journal* report and reaches us only through belligerent and allied media: *Al Mayadeen* [TG-720237], *Almasirah* in English [TG-720352], *Mehr* as a "revelation" [TG-720335], and *CIG* via *AMK Mapping* [TG-720189]. Each repetition increases apparent corroboration while adding zero independent sourcing. I want to flag my own temptation here: depletion is exactly the input an escalation model wants, because it generates clean predictions about defender capitulation. It should not be treated as a datapoint about Saudi magazines; it is a datapoint about which argument four ecosystems have decided to build together. Second, the Hormuz incidents display a deliberate asymmetry of authorship. The mine strikes are claimed — the IRGC Navy announces them, names the cargo, publishes video [TG-719687, TG-720106]. The drone attack is not: *UKMTO* reclassified the Sohar anchorage event from incident to attack [TG-719674, TG-719809], *Rybar MENA* names the vessel *Petro Dolphin* [TG-720529], and India's foreign ministry confirms two ships with Indian crew hit with two crew missing, per *Radio Farda* [TG-720586] and *Xinhua* [WEB-102198]. No actor claims it. The same pattern recurs in Iraqi Kurdistan, where *Al Jazeera*'s correspondent reports a two-drone strike on an Iranian Kurdish opposition camp near Koysinjaq [TG-720984] and *Fars* describes the targets as "separatist terrorists" without claiming the strike [TG-720950]. Claimed actions carry signaling value; unclaimed ones carry deniability. Running both simultaneously is a bargaining posture, not confusion. Third, the negotiation signals are being read through a single channel. Rezaei says Washington accepted some of Iran's seven conditions for reopening the strait but that the response is inadequate [TG-719664, TG-720007] — relayed by *Xinhua* [TG-720383] and headlined by *Haaretz* as "insufficient" [WEB-102284]. Parliament's national security spokesman Qashqavi tells *Tasnim*, via *Al Mayadeen*, that Qatar relayed a reply and that the American side "does not know what to answer" because its statements keep contradicting each other [TG-720753, TG-720754]. Every element of that — the partial acceptance, the inadequacy, the American incoherence — is Iranian-sourced. The one external constraint datapoint comes from *Anadolu*, reporting an Israeli opposition leader's expectation that Israel will not strike Iran before its elections [WEB-102281]. Election calendars are the only clock in this crisis anyone can verify.

Discipline first: nearly everything load-bearing today is a claim by a belligerent about its own performance. *IRGC Navy* says three tankers hit mines [TG-719622]. *Ansarallah*'s General Saree says four operations in 24 hours against Riyadh, Dammam, and two Aramco sites [TG-718930][TG-718933][TG-718935]. The Saudi-led coalition says it intercepted a cruise missile [TG-718691]. *UKMTO*, the nearest thing to a neutral observer in the corpus, says only that a tanker was damaged by an unidentified projectile [TG-718403] and later reclassified an incident off Sohar, Oman, as a drone attack [TG-719674]. Build models on the UKMTO line; treat the rest as signals. As signals, the structure is legible. Iran's choice of "naval mine" over "missile" is deliberate ambiguity — a mechanism that imposes costs while denying an engagement decision, the maritime equivalent of a deniable proxy. It is sub-threshold by construction. Against that, Rezaei's threat is explicitly horizontal rather than vertical: if war resumes, the region's cables, data centers and AI facilities — valued by him at over two trillion dollars — become targets [TG-718765][TG-718803], and American assets in the UAE and around the strait are "not yet" on the list [TG-718723]. Introducing a new target class while declining to strike it is a classic reputation-building move at low cost. Most important is what the diplomacy track shows. Rezaei states the Americans answered Iran's seven conditions, accepted some, and that Tehran judges the response insufficient [TG-718925][TG-718926], with approaches arriving through Pakistan, Qatar and Iraq [TG-718896] and the Qatari channel conveying the latest text [TG-719429]. *TRT World* [WEB-102074] and the *Jerusalem Post* [WEB-102104] both framed it as rejection. But a party that enumerates which of its conditions were accepted is negotiating, not refusing. Rezaei's simultaneous line that Iran will not return to its former diplomacy [TG-718833] is a commitment device aimed at his own side as much as at Washington. The escalation risk this window is not Hormuz. It is the convergence of fragments: *Tasnim* via *Almayadeen* reporting F-16s, F-35s and five KC-135s in Iraqi airspace launched from Jordanian and occupied territory [TG-719455][TG-719457], *Fotros* reporting Israeli jets exercising near Iran's western border [TG-719304], and Israeli Channel 13 via *Al Jazeera Arabic* reporting Netanyahu convening security consultations after Trump raised the possibility of resuming strikes on Iran [TG-719295]. Each is thin. Assembled, they constitute a mobilization narrative — and narratives of that shape tend to precede the thing they describe.

Two statements from the same American president in the same window, and the gap between them is the whole analytical problem. To *Middle East Spectator*, asked about his promise not to attack Iran: "We may. I mean, we'll see. But they're failing very badly" [TG-714113]. At a Syracuse rally hours later, per *ajanews* flashes: "Either Iran gives us everything we want, or it will no longer exist" [TG-714290], and "the Iran file will be settled very quickly, one way or another" [TG-714291]. *AbuAliExpress* carries the same line in Hebrew [TG-714945], *Solovievlive* in Russian [TG-714411], *Radio Farda* in Persian [TG-714817]. In signaling terms these are different instruments. "We'll see" is calculated ambiguity addressed to an adversary. "Cease to exist" at a campaign rally is a commitment device addressed to a domestic audience — and commitment devices are dangerous precisely because they transfer control of escalation to whoever is watching. *Dawn*'s editorial names the structural problem outright: Trump's promise to hold off attacking Iran until the midterms "does little to ease fears of a wider war" [WEB-101564], because a deadline tied to an election calendar tells Tehran exactly when the constraint expires. *BBC Persian* works the same seam, asking how the Iran war has complicated Trump's calculations ahead of the vote [TG-714867]. Against the maximalism, a concrete and much narrower demand surfaces. *Solovievlive*, citing Reuters, reports that Washington will require Iran to reduce enrichment capacity to end the war and remains willing to keep negotiating [TG-714426]. That is a bargaining position, not a regime-change position, and it sits oddly beside "everything we want." On the Iranian side, *Mehr* has the National Security Committee spokesman saying Iran's seven conditions in the "seven-day plan" remain in force and that "pressure on America is greater than on Iran" [TG-714837]. Both sides are signaling a deal exists; neither will be seen reaching for it. A discipline note on evidence, because this window offers a textbook case. *Al Mayadeen* ran a four-part breakdown of a *Bloomberg* report citing Western officials: Iran preserved its missile and drone factories through months of US and Israeli bombing, retains a solid ballistic stockpile and some anti-ship weapons, and "if Washington resumes airstrikes, Tehran has enough firepower to respond" [TG-714841, TG-714842, TG-714843, TG-714844]. *Al Jazeera Arabic* carries it too [WEB-101535]. This is useful because it is adversarial to the claimant — Western officials undercutting their own leadership's damage claims, which *Al Mayadeen* highlights explicitly [TG-714843]. It is still a report of an assessment, not an inventory. Build a capability model on it and flag the uncertainty. Finally, a trial balloon worth logging. *Middle East Spectator* quotes Tehran's Friday Prayer Imam: "The nation demands a reconsideration of nuclear doctrine. Why should we remain a signatory of the NPT?" [TG-714142]. One source, one pulpit, no official echo in this window. Attributed to the nation rather than the state — the classic architecture of a position being tested rather than taken.

Three structural observations, with a warning attached to each. First, the Hormuz dispute has become a pure signalling exchange detached from any verifiable state of the world. *BBC Persian* [TG-706055] reports Rubio in Athens saying Iran has 'completely lost control' of the strait and its economy is in freefall; *Mehr* [TG-705981] renders the same remarks as 'Hormuz is open!' with the exclamation mark doing the editorial work. Opposite it, *ISNA* [TG-705948] and *Fars* [TG-706123] carry an IRGC adviser saying the strait is closed and under full armed-forces control until Iran's legitimate demands are met, and *Al Jazeera* [TG-707124] has the Basij chief adding that Hormuz 'was and will remain Iranian.' Both sides are making a claim about physical reality that the other's claim cannot be reconciled with. Neither is a datapoint about capability. The only third-party metric in the window is the *Reuters* tanker count [TG-707128, WEB-100757] — and note it supports *neither* framing cleanly: traffic is flowing and being attacked. Second, and this is the escalation-theoretic event of the window: *Al Jazeera Arabic* [TG-707670, TG-707671], *TASS* [TG-707751] and *Jerusalem Post* [WEB-100830], all citing *The Atlantic*, report the White House requested Iran strike options executable before the 3 November midterms, with *Al Jazeera* [TG-707672] noting the plans reflect Trump's desire to lower fuel prices. If accurate, this inverts the standard escalation model. Escalation driven by an electoral calendar is not responsive to the adversary's behaviour — which means Iranian restraint cannot purchase de-escalation. That is a dangerous property. It also means Tehran's incentive to concede before 3 November is near zero. Third, watch the pre-positioning. Iran's foreign ministry spokesman released a historical explainer on American and Israeli false-flag operations — *IRNA* [TG-706787], *Press TV* [WEB-100774] — and *teleSUR* reproduced the frame in Spanish within hours [TG-707956]. A belligerent building an advance interpretive frame for an incident that has not yet occurred is a recognizable move: it lowers the cost of denying attribution later. Set beside *Fars*'s report [TG-707680] of unexplained explosions off Kuhestak and Iranian state outlets relaying unverified 'Hebrew sources' on a blast in Haifa [TG-707199, TG-707393], the information environment is being prepared for contested attribution on both sides.

The structurally interesting thing this window is not an escalation but a signaling divergence inside one government, surfaced by an ecosystem actor rather than by the government itself. The US Vice President told *Reuters* that ending the war requires Iran to make a "meaningful" reduction in enrichment capacity [TG-704944], and said Washington remains open to a deal conditioned on tangible nuclear concessions [TG-704732]. *Middle East Spectator* flagged the diction: "reduce," not "give up," and noted that this departs from the prior US position [TG-704706]. Within hours the Secretary of State said Iran will "never" have a nuclear program [TG-705793], and the House Speaker said Iranians cannot be trusted as negotiating partners [TG-705991]. On the arms-control ladder, the distance between "reduce capacity" and "never" is the entire bargaining space. Three principals occupying three positions in one news cycle is either deliberate ambiguity or coordination failure; from Tehran's side the two are indistinguishable, which raises the risk of misreading a real offer as a feint. Second, a methodological note I want to insist on. The volume of operational claims in this window — Houthi strikes on Riyadh, Abha, Khamis Mushait, three camps in Jizan and Asir with "dozens dead and wounded" [TG-705086] — comes overwhelmingly from the belligerent's own outlet, *Al Masirah*, in rapid-fire identical bursts [TG-705081][TG-705082][TG-705084], with the coalition claiming interception [TG-704941]. These are competing announcements, not capability measurements. The exception worth isolating is the AESSCO hangar damage, where *Al Masirah* cites third-party satellite imagery [TG-705079][TG-705080] — that is a different evidentiary class, and the fact that the Houthi outlet reached for it suggests awareness that its own word has depreciated. Third, a self-contradiction inside the Iranian ecosystem that matters for any future bargaining model. An IRGC adviser told *Al-Alam* that Hormuz is closed and only small boats smuggle oil out [TG-705948]; the outgoing oil minister insisted production and export continued despite blockade [TG-705577]. Those are opposite claims. Each is coherent for its audience — deterrence credibility for one, regime resilience for the other — but together they mean we cannot infer Iran's actual reservation price from its public statements, because its public statements are not addressed to us. Fourth, precedent. *Mehr* relayed a *Foreign Affairs* warning that the belief short heavy strikes can force capitulation has repeatedly started or prolonged wars [TG-704670]. That an Iranian state outlet chose to amplify a Western academic-journal argument is a signal in itself: Tehran is building a case that time, not escalation, is the operative variable. Trump's framing — finish it "the pleasant way or the less pleasant way," plus the claim that nobody knows who runs Iran [TG-705471][WEB-100584] — is the inverse premise. Two actors with incompatible theories of how this ends are the precondition for a long war, not a short one.

This window offers a textbook case of two incompatible battlefield maps circulating simultaneously, and the discipline question is which inputs an analyst is entitled to use. On one side: *Al Jazeera English* reporting Saudi-backed Yemeni forces confirming they retook Mocha [WEB-100131], *Al Arabiya* and *Al Hadath* carrying 250 Houthi casualties on the Taiz axis [TG-701501, TG-701502], *Anadolu* reporting 100 Houthis killed and five drones downed [WEB-100189], and *Daily Sabah* headlining the recapture of areas near Bab al-Mandab [WEB-100199]. On the other: *Almasirah* and *Al Jazeera Arabic* carrying the Ansarallah spokesman's claim of full control over all field gains and dismissal of "fictitious heroics" [WEB-100150, TG-701922], plus more than 200 casualties inflicted at Ras al-Arah and al-Suqya [WEB-100178, TG-701908]. The structurally interesting actor is *Middle East Spectator*, which did something neither belligerent did: it checked. It reported that Al Jazeera's Mokha capture footage was filmed 56 kilometers south of the city [TG-701396], then tracked the PLC advance and complete withdrawal back to Ras al-Arah within a single day [TG-701595]. It also documented the rhetorical adjustment on the other side — pro-Saudi commentators moving from "we have captured Mokha" to "don't rush the battle, it requires patience" within six hours [TG-701485]. That shift is evidence about confidence, not about terrain, and it is the kind of input I can actually use. A note on my own discipline: *Rybar*'s Middle East desk, no friend of Riyadh, nonetheless treats the Houthi recovery claims conditionally — "if we believe the Houthis" [TG-702416]. When a partisan channel hedges and a wire service does not, the hedge is the better signal. On escalation structure, the genuinely new variable is the airspace denial claim [TG-702685, TG-701902]. This is a threat to third parties, not to a belligerent, and it is designed to be self-enforcing through insurance and scheduling decisions rather than through further strikes. That makes it difficult to de-escalate: there is no obvious reciprocal concession, and the cost continues accruing after the launches stop. *Almasirah* already cites flight-tracking sites as corroboration [TG-702598] — the belligerent outsourcing verification to third-party data it does not control. Meanwhile both principals are signaling exit. Trump says the Iran war ends soon "one way or another" and prices will fall [TG-701567]; Pezeshkian, per *Al Jazeera Arabic*, blames Washington for collapsing the talks [WEB-100085] and tells *Naharnet*'s readers negotiations are "meaningless" [WEB-100235]. *IRNA* relays a Council on Foreign Relations assessment that any final deal will resemble the Islamabad memorandum [TG-701886]. Mutual exit signaling without a channel is where accidents live.

The structural development here is that both parties have begun publishing the existence of a negotiating channel while disputing what the channel is about — and I want to be careful that I treat these as signals rather than facts. Ghalibaf, in a pre-agenda speech to parliament, states that "Americans who say otherwise in the media have recently raised proposals through a mediator," that the era of running down the clock is over, and that the Strait of Hormuz will not reopen until Iran's seven conditions under the Islamabad memorandum are met [TG-696085, TG-696099, TG-696100, TG-696110, TG-696115, WEB-99523, WEB-99537]. Baqaei adds the mediator's identity and the direction of travel: the Qatari side conveyed the American view during a stopover in Doha, Iran is studying its reply, and the American proposal "resembles previous ones and focuses on the nuclear file" [TG-696399, TG-696436, TG-696460]. Iran's counter-offer is described as a seven-day plan under which Hormuz reopens within a week if accepted [TG-696510, TG-696208]. This is a classic commitment-credibility problem played out in public. Ghalibaf's formula — "we both fight and negotiate" [TG-696067, TG-696142] — is an attempt to make a conditional concession credible without appearing to concede. Araghchi's version is cleaner: there is no military solution and no solution based on more sanctions, the only route is negotiation grounded in justice and diplomatic dignity [TG-696227, TG-696261, WEB-99547, WEB-99555]. Note the file mismatch, which is the real structural obstacle: Washington's proposal is reportedly about enrichment; Tehran says its current focus is exclusively Hormuz [TG-696446, WEB-99563] and explicitly denies that it offered IAEA inspections in exchange for sanctions relief [TG-696445, WEB-99561]. Two parties negotiating over different agendas will generate the appearance of bad faith even where none exists. The counterparty's voice reaches us only through reflection. Trump, per *BBC Persian*, said the conflict can be settled "the hard way or the easy way" and that a decision is his personally [TG-695601, TG-695552], and per *AJA* that he will force Iran to abandon its program [WEB-99536]; per *Dawn*, that the war will end "very soon" [WEB-99474]. *Mehr* relays Hegseth repeating that the Hormuz blockade is firm [TG-695494]. I would caution against building anything load-bearing on these: they are campaign-rally statements as rendered by hostile or sympathetic intermediaries, and the one US datapoint we can actually verify in this window is procedural — a sanctions waiver permitting Iraqi–Iranian flights to resume [TG-695820, WEB-99455].

My discipline requires me to say first what this window does not contain: verified capability data. It contains claims, denials, and imagery, and the structural question is what each actor's choice of claim reveals about its assessed position. Start with the Saudi denial. The coalition told *Asharq al-Awsat*, carried by *Al Arabiya* and *Al Hadath*, that Houthi claims of targeting Riyadh are 'misleading' [TG-695442] [TG-695433], and *Xinhua* carried the same line [WEB-99425]. In signaling terms, denial is a costly choice when counter-evidence is cheap to produce — and *Al Mayadeen* had already published what it described as satellite imagery of fire at the Aramco site [TG-694326], with *Fars* claiming twelve hours of continuous burning per *Reuters* [TG-694820]. A state that denies a strike it cannot conceal is prioritizing domestic reassurance over deterrent credibility. That is a defensive signal, and the escalation literature treats it as a predictor of either disproportionate retaliation or quiet accommodation — rarely of steady state. Second, the ultimatum. *Mehr* and *Fars* both report Yemeni forces giving Saudi-backed units in Taizz 48 hours to withdraw [TG-694693] [TG-695177]. A published deadline is a commitment device: it transfers the cost of inaction onto the issuer. Issuing one while simultaneously striking the adversary's capital is the classic combination of coercive diplomacy — but the same window shows *Al Jazeera Arabic* carrying government claims of recapturing Al-Maqarima, Al-Zakirah and Jabal al-Ja'sha [TG-695560], which, if believed, makes the deadline unenforceable. I cannot adjudicate the ground truth, and that is precisely the point: both sides are issuing commitments calibrated to information their own audiences cannot check. Third, and most important for escalation forecasting: an authoritative actor has stated that it treats war as inevitable. *Al Mayadeen* carries Hezbollah MP Hassan Fadlallah saying Iran assesses that Netanyahu's interest lies in continuing the war, and that Tehran operates on the assumption that war's occurrence is certain [TG-695396]; he adds that Iran has achieved a deterrent balance unseen since the Second World War [TG-695397]. Set aside whether the second claim is true. The first describes a belief structure in which de-escalatory signals from the adversary are pre-coded as deception. Under that belief, confidence-building measures do not reduce risk; they increase it, because each one is read as cover for preparation. Against that, two de-escalatory data points sit awkwardly. *Al Jazeera Arabic*, citing the Iranian presidency, reports the national security council secretary telling the economic coordination body that current conditions are the most difficult — and, in the same readout, that the negotiation track is serious [TG-694919] [TG-694920]. And Trump, per *AJA*, says Iran has abandoned many of its nuclear weapons plans and that he will make a decision on Iran [TG-695486]. These are the two sides each telling a domestic audience something slightly softer than their declaratory positions. Historically — Iraq 2003 is the cautionary case — that pattern is compatible with both a settlement and a run-up. What distinguishes them is whether either side's media ecosystem builds the frame to receive a settlement. Neither did in this window. One methodological note for my colleagues: the Qeshm air-defense event reached us as sound, then as OSINT identification [TG-694502], then as state confirmation via *Tasnim* [TG-695073]. *Radio Farda* observed that *Mehr* published without citing sources [TG-694893]. Build no capability inference on that chain.

Four claim-objects in this window are being treated as facts by the ecosystems carrying them, and all four fail the sourcing test. First: Yemeni government forces announce 474 attacks killing or wounding 1,540 Houthis [WEB-99200, WEB-99141] and 71 separate targetings in Taiz [TG-693058]. Second: the Houthi military spokesman announces 94 Saudi airstrikes and missile attacks in 24 hours [TG-693027, TG-693362]. These are mirror-image belligerent tallies, each carried onward by the aligned outlets — *Al Jazeera*'s Arabic and English services for the first, *Press TV* and *Radio Farda* for the second — and neither is independently corroborated anywhere in the corpus. Third: an IRGC Aerospace adviser claims long-range ballistic capability against moving targets [TG-693284]. Fourth: *Fars* reports a US Navy MQ-4C Triton overrunning a runway [TG-693716, TG-693733]. Each is a signal about what its sender wants believed, not a measurement. The structurally important development is a termination signal. Trump, at a rally in Alabama, says the war will end 'one way or another,' possibly immediately after the midterm elections [TG-692975, TG-692999, TG-693168]; that oil prices will fall sharply once it ends [TG-692976]; that Iran will never possess a nuclear weapon [TG-692958]; that the blockade is now called 'a wall of steel' [TG-692977]. Binding war termination to an electoral date is a costly-signal problem in reverse: it tells the adversary the clock is political, which rewards waiting. *IRNA* read it exactly that way within hours, publishing a party secretary-general's analysis that Trump is 'besieged by electoral calculations' [TG-693418] and flagging the internal contradiction of simultaneously conceding Iran has no weapon [TG-693138]. More revealing is the line almost nobody amplified. *Al Jazeera*'s Arabic flash carried Trump saying 'part of our problem is I do not know who I am dealing with in Iran, because there is nobody who wants to lead' [TG-692974]. In a crisis whose central unresolved variable is Iranian succession, that is the most consequential sentence of the rally — and in this window it travels through one Arabic channel and does not appear in the Iranian state material we collect. Absences in amplification are data. Finally, the deliberate leak. *Axios* reporting of a secret Camp David session on Iran and Yemen chaired by the Vice President [TG-693187, TG-693177, WEB-99270] is a signal with a known function: it communicates deliberation without committing to an option. Every ecosystem reprinted it; none added to it.

Discipline first: almost everything load-bearing in this window is a claim, and several claims are explicitly flagged as unverifiable by the claimants' own allies. The cleanest structural datapoint is the commitment problem being stated out loud. Al Jazeera Arabic, citing Axios citing US officials, reports Washington cannot guarantee the IRGC or hardline currents would abide by any agreement [TG-691776]. That is one side publicly conceding it cannot identify a counterparty capable of binding itself. Pair it with Tehran Times quoting Khatibzadeh saying Iran seeks concrete US guarantees to end the war [WEB-99105], and you have both parties announcing that the obstacle is assurance, not terms. Theory says that is where mediation enters — and Iraq's foreign minister offered exactly that this window [TG-692025], as did Baghdad's broader 'we oppose war' positioning in his Radio Farda interview [TG-691128]. Against that, the public signaling is almost pure noise, and it is calendar-driven. Trump told a campaign rally that Iran is ready to surrender and that victory comes right after the election [TG-691877], claimed Iran is in a difficult position [TG-692617, TG-692618], and promised oil below pre-war levels once the war ends [TG-692620, TG-692635]; BBC Persian reports he also claimed for the first time that the bombed nuclear facilities were drug production sites [TG-691514], a line Middle East Spectator quoted verbatim [TG-691916]. Mohsen Rezaei's reply is the mirror image — 'reality is shaped by Iran, by its people, its soldiers, its missiles' [WEB-99064, TG-691888]. The IRGC's True Promise 2 anniversary statement declaring the era of cost-free threats over [WEB-99104, TG-691518] is deterrence-restoration rhetoric pinned to a commemorative date. Anniversary signaling carries very little information about intent; it is addressed inward. Two clocks are visible. Reuters via Daily Maverick describes Trump's 32-day midterm blitz [WEB-99013]; Anadolu reports Israeli polling at 51 seats for the Netanyahu bloc against 53 for the opposition before the October 27 vote [WEB-98981]. Both principals now have an electorate-facing incentive to convert ambiguity into a claimed outcome. And a warning on inputs: the Yemen numbers — 1,540 Houthis neutralised in 474 strikes [TG-692405, TG-692407] against 94 Saudi raids and 1,350 cumulative [TG-692285] — are not capability data. They are messaging artifacts, and no structural model should eat them.

Two separate ecosystems converged on the same calendar date this window, which is the structurally significant development. Trump told *TIME* — carried in fragments by *Al Jazeera Arabic* — that renewed or intensified strikes on Iran after the US midterms are "possible" [TG-688435] [TG-688486], and that the Iran issue could hurt him electorally [TG-689464]. Within hours, *Fars* published economist Majid Shakeri arguing the Islamic Republic has a "36-day golden window" and should raise the level of tension through it [TG-688518]. Netanyahu, per *Al Jazeera Arabic*, said there are indications Iran and its proxies will try to attack during Israel's election season [TG-688547]. Three actors, three electoral clocks, one shared assumption that the window before voting is where leverage lives. In signalling terms, publicly announcing your own timing constraint is usually a commitment device. When all parties do it simultaneously, it becomes a coordination point for escalation rather than restraint. Methodologically, I want to flag the attribution cascade in this window, because it is where the analytic risk sits. Three incidents — the RAF Fairford security event, the Flydubai cockpit attack, and an alleged foiled suicide bombing in Mecca — were each publicly attributed or insinuated before investigation closed. *Fars* reports the UK prime minister claiming "strong evidence" of an Iranian role at Fairford [TG-688836]; *Al Jazeera English* asks the obvious question in its headline, "Is there evidence?" [WEB-98789]; *TASS* then reports police found no explosives in the suspects' vans [TG-689964], and a British-Iranian dual national was arrested as suspected organiser [TG-689611]. On Flydubai, Trump said Iran would be "hit very hard" if involved — we know this only through *abualiexpress* relaying his exchange with a reporter [TG-689607]. The Mecca claim exists in our corpus only via OSINT aggregators [TG-690183] [TG-690189]. One aggregator, *Tabz*, actually names the pattern: an unusually irregular week of non-state attack attempts [TG-690233]. That framing is more useful than any individual attribution. The discipline point: operational claims are not capability data. IRGC aerospace adviser Balali's statement that Iran can sustain its missile launch rate "for several years" [TG-688305] [TG-688410] [TG-688411] reached us only through *Fars* and *Al Mayadeen* — belligerent plus allied media. Trump's claim that munitions shortage reports were "mostly fake news" [TG-688437] is the mirror image, and *Mehr* inverted it into a headline about Trump *admitting* depleted stocks [TG-688678]. Both are signals about what each side wants believed. Neither is evidence about magazine depth. Finally: Iranian deputy FM Khatibzadeh saying "the ball is in America's court" [TG-689675] while *Reuters*, via *Al Hadath*, reports Iran preparing a broader response including targets in countries that assist a US attack [TG-688968] [TG-690026] describes a posture of deliberate ambiguity. Combined with declared electoral deadlines on three sides, the structural risk here is not a decision to escalate — it is an incident arriving inside a window everyone has announced they intend to use.

Two structural observations, and one methodological warning to my colleagues. First, the Iraq withdrawal produces three mutually exclusive victory constructions from the same event, which is the signature of a settlement nobody has actually agreed to. Trump's telling, which reaches us only through *Al Jazeera* and *Al Mayadeen* reflection, is a withdrawal he chose — an orderly exit from a costly quagmire, with credit claimed for the Iraqi prime minister's election win [TG-686346, TG-686387, TG-686421]. The Islamic Resistance in Iraq's telling, delivered at the Palestine Square rally, is expulsion under painful blows, with explicit refusal of any disarmament timetable [TG-685636, TG-685643]. The Iraqi state's telling is a sovereignty transition in which war and peace become an exclusive state prerogative [TG-685275, TG-685569]. Ambiguity of this kind is sometimes stabilising — everyone can claim to have won — but it is unstable when one party has publicly refused the central term. Kataib Hezbollah's evening order to halt military operations and mobile logistics [TG-686446] is the first datapoint suggesting the ambiguity is being managed rather than tested. Second, the escalation ladder is being described in mirror-image absolutes. Trump says the US has virtually total control of the Strait of Hormuz [TG-686406]; the IRGC spokesman says there are military confrontations every twenty-four hours and Washington has stopped responding [TG-684780, TG-685084]. Both cannot be true. What matters analytically is that both are addressed to domestic audiences, and that Trump's own framing — maybe we strike, maybe we make a deal, the war ends soon one way or another [TG-686887] — is the language of a decision not yet made, six days after Iran's seven-day proposal and one day after Washington's reply arrived [WEB-98264, WEB-98300]. The warning: today produced a textbook false signal. *Middle East Spectator* published initial reports of ballistic missile launches from Iran [TG-685984]; *Boris Rozhin* posted purported launch footage [TG-686375]; Tasnim, via *Al Mayadeen*, denied any launch toward Jordan or from Shiraz and contradicted Israeli Channel 14 [TG-686622, TG-686623]. Nobody should have built a capability inference on that chain.

Discipline first: nearly every operational claim in this window is single-sourced to a belligerent. The IRGC's "full control" of Hormuz [TG-678291] and CENTCOM's "we sank their navy" [TG-678781] are both claims; the Houthi spokesman's count of 1,123 Saudi strikes [TG-679655] and the Yemeni government's 356 operations with 476 Houthis killed or wounded [WEB-97704] are both claims. Treating any of them as capability data would corrupt the model. What can be modelled is signalling behaviour, and here the window is unusually legible. The pattern is *mutual deniability construction*. Mediated talks are confirmed by Reuters via *AJA News* [TG-678851] and *BBC Persian* [TG-678918]; Araghchi meets Qatari mediators with, per *ISNA* relayed by *AJA News*, no American representative present [TG-679113, TG-679115]. Both principals then deny the substance while conceding the process: Trump confirms his team held indirect talks [TG-680110] but denies the Axios account of concessions [TG-680299]; *Tasnim*'s source denies any agreement to negotiate [TG-679169] while Araghchi tells *IRNA* he discussed ideas with the Qataris and that a reply will be conveyed to the American side [TG-680324]. This is the signature of a negotiation that both sides want to be able to abandon without domestic cost. It is a positive escalation indicator only in the weak sense: parties building deniability are parties who expect to need it. The structural risk sits with the audience costs each side is generating. Iranian parliamentarians are raising the price of entry publicly — Moqtadai saying Washington must accept Iran's conditions before any talks and that Hormuz, not the nuclear file, is now the core [TG-680183, TG-680184]; Rezaei conditioning talks on US performance of the Islamabad understanding [TG-680188]; a three-urgency NPT withdrawal bill going to committee [TG-679934, WEB-97697] while Araghchi tells NBC that 60% enrichment is peaceful and the NPT intact [TG-679778, TG-679930]. On the American side, *Al Mayadeen* relays CNN quoting a US official who, asked what Republicans could do about the midterms, answered: go back in time and never start the war [TG-678626], and *The Telegraph*, also via *Al Mayadeen*, reports Virginia veterans planning to abstain [TG-678708, TG-678709]. Both leaderships are accumulating commitments that make the mediated framework harder to sign and easier to blame the other for losing.

Two structural observations, both about inputs rather than conclusions. First: this window contains a direct, unresolved contradiction about whether negotiations exist. *Xinhua* reports Trump expecting further US-Iran talks within the week, hours after rejecting Iran's Hormuz proposal [WEB-97333, WEB-97361]; *IRNA*, cited by *Axios* in the Iranian rendering, is carried by Iranian outlets as confirming a new round [TG-677914]. Then *IRNA* itself, via a source close to the delegation, states the Iranian mission in New York has no plan to negotiate with the United States — relayed by *BBC Persian* [TG-678298], *Naharnet* [WEB-97457], *L'Orient Today* [WEB-97458] and *Daily Sabah* [WEB-97488]. Both claims cannot be true. For escalation modelling this matters more than either claim's content: when the two principals cannot agree on whether a channel is open, third-party mediators lose the ability to calibrate, and every subsequent signal is read through the wrong frame. The *Jerusalem Post* reports mediators pushing nuclear-file concessions as the entry price [WEB-97418] — a sequencing demand that, if real, explains the contradiction without resolving it. Second, and this is the discipline point: almost every operational claim this window is single-sourced to a belligerent. *Barantchik* reports Iran striking nineteen merchant vessels in Hormuz over two days, sourced to *Fars* citing its own sources [TG-677638]. Nineteen vessels is a blockade-scale event. No other channel in this corpus carries it — not the Gulf outlets, not the OSINT aggregators, not the shipping press. A claim of that magnitude with that sourcing profile is a signal about Iranian messaging intent, not a datapoint about Iranian capability. Treat it accordingly. The same applies in the other direction. Trump's assertion that Iran's nuclear capabilities are destroyed and therefore no longer a concern [TG-678508 — carried by *Al Jazeera Arabic*] is a claim by a belligerent about its own campaign's success, offered without verification, at a moment when the same speaker is declining to rule out further strikes before the midterms [TG-677241, TG-677282]. The escalation-relevant domestic signal is Iranian: *ISNA* reports a group of former officials publicly urging withdrawal from the NPT [TG-678172]. Former officials are a low-cost trial balloon — the state can disavow. But it establishes the option in public discourse, which is what trial balloons are for. Historically, this is the step that precedes narrowing rather than widening the negotiating space.

Structurally, this window changed the subject of the dispute. Trump's rejection of Iran's seven-day proposal is confirmed across ecosystems — *Xinhua* [WEB-97066], *Guancha* [WEB-97058], *Fars* [TG-674572], *solovievlive* [TG-674905] — but the substantive US position, as reported by *BBC Persian* citing Axios and by *Al Jazeera Arabic*, is that Iran lacks standing to set any conditions because it does not control the strait [TG-674555] [WEB-97069]. That is not haggling over terms. It is a dispute about who is a legitimate party to the transaction. Disputes over standing resolve far more slowly than disputes over price, because concession implies recognition. Both sides are now building record for that argument rather than for a deal. Araghchi's repeated formulation — conditions conveyed fully to mediators, awaiting the final word, "no one can accuse Iran of not testing diplomacy" [TG-674567] [WEB-97053] [WEB-97133] — is standing-construction, aimed at third parties. Saudi Arabia's "pre-February-28, no fees or restrictions" formula [WEB-97051] and Oman's offer of continued mediation [WEB-97080] are the third parties responding. A sourcing discipline note for my colleagues. The night's dramatic material — *Middle East Spectator* reporting 10+ anti-ship cruise missiles and mine-laying [TG-674670] [TG-674671], *IntelSlava* citing unnamed "Iranian sources" for at least six launches [TG-674648] — is a claim from aggregators, not a capability datapoint. No Iranian state channel in our corpus confirmed it; *Jerusalem Post* put it in a live-blog headline [WEB-97082]. Treat it as signal about what the OSINT layer will carry, not as evidence about Iranian inventory. The genuinely interesting de-escalatory signal came packaged as a denial: Iran's SNSC explicitly rejected speculation it would strike regional airports over sanctions compliance [WEB-97034] and said its options are non-military [WEB-97095]. Ruling something out in public is a costly signal. Against that, Hatami's "the war is not over" [WEB-97156] and Akraminia's warning of renewed US aggression [WEB-97154] read as domestically-addressed, not as ladder-climbing.

The structural feature of this window is that the most consequential claim is one we cannot see directly. The Wall Street Journal is not in our corpus. Everything we know about Trump's rejection of the seven-day proposal reaches us through mirrors: Al Jazeera Arabic in a six-post breakout [TG-671998, TG-671999, TG-672000, TG-672002, TG-672004], intelslava [TG-672006], Middle East Spectator [TG-672018], BBC Persian [TG-672007], Xinhua [WEB-96770], Haaretz [WEB-96820], Jerusalem Post [WEB-96772], L'Orient Today [WEB-96828], SABC [WEB-96867]. Al Mayadeen carries a parallel Washington Post version [TG-672008]. That is a claim about presidential intent, sourced to anonymous officials, now functioning as a settled fact across at least nine ecosystems within three hours. Before anyone builds an escalation model on it, note that we are modelling a leak, not a decision. The sequencing deserves attention. Araghchi announces the plan at roughly 21:00 [TG-671708]. Within forty minutes a US official tells Reuters the indirect talks have been 'constructive' [TG-671758, TG-671776, TG-671874]. Around 21:35, an informed source speaking to IRGC-affiliated Fars denies Axios and Al Jazeera reports of a new negotiating round entirely, calling them false [TG-671755]. At 00:29, the rejection lands. Four incompatible signals in three and a half hours, three of them anonymous. The information environment is not tracking a negotiation; it is tracking the residue of one. On signalling: the midterm framing is the structurally interesting part. If the reported expectation is that strikes resume after November [TG-671999] and that Washington believes economic pressure will force a signature [TG-672023], then the ecosystems are collectively constructing a model in which the war has an electoral clock. Tehran's own messaging answers that clock in kind — the IRGC saying punishment continues until seven conditions are met [WEB-96798], Pezeshkian telling CBS that Washington's object is not the nuclear file but regime change [TG-672675]. Both sides are now publicly arguing that the other cannot be satisfied, which is what a commitment problem looks like from the outside. One caution for my own discipline: the Houthi launches on Riyadh and Khamis Mushait are claims-and-counterclaims, not capability data. The Saudi-led coalition announced interceptions [TG-672040, TG-672076, TG-672107]; OSINT channels reported explosions and no sirens [TG-671972]; Iranian state outlets reported 'explosions heard' and attributed them to unofficial sources [TG-671815]. Nobody in this corpus observed an impact.

The structural question in this window is whether we are looking at one Iranian negotiating position or several, and the honest answer is that the information environment cannot tell us. Track the claims. Araghchi, speaking at the UN Security Council and then to reporters, says Iran has conveyed a specific seven-day plan to Washington via Qatar; if conditions are accepted, Hormuz reopens and normal traffic resumes within a week, and the clock starts the day the US accepts [TG-671634, TG-671636, TG-671698, WEB-96718]. *Naharnet* and *Haaretz* both report the conditions include a Lebanon ceasefire [WEB-96527, WEB-96643]. Qatar's PM tells CNN, per *Al Jazeera Arabic*, that indirect talks the previous day made progress and both sides showed willingness [TG-671275, TG-671276]. Now the counter-signals, all from the same state. *Reuters*, via *Middle East Spectator* and *IntelSlava*, quotes Iranian officials saying Tehran will make no nuclear concessions even if a Hormuz deal is reached — and *Middle East Spectator* explicitly flags that this contradicts Pezeshkian [TG-671104, TG-671439]. *Fars*, citing an informed source, denies the Axios and Al Jazeera reporting of a new negotiating round outright and says it is intended to move oil prices [TG-671376, TG-671389]. *Al Mayadeen* carries a senior Iranian security source saying the same and adding that the seven conditions were delivered and "the ball is in the American court" [TG-671646, TG-671694]. And Pezeshkian tells Fox News, per *ISNA* and *BBC Persian*, that Iran is willing to dilute its 60 percent stockpile under NPT verification [TG-670957, TG-671034]. Signaling theory would call this deliberate ambiguity. But there is a cheaper explanation that the data also supports: factional publics. *Fars* ran a piece asking whether reformists would "impose a third war" through renewed negotiation [TG-670373]; *Fars* also carried a former official saying Pezeshkian should have been more careful in foreign media interviews [TG-670804]; the resistance-aligned *Fotros* channel complained outright about him talking to Fox and CBS [TG-671509]. When a state's negotiating signal is being publicly litigated by its own media, the recipient cannot distinguish strategy from disagreement — and that is exactly when miscalculation happens. One caution for my own side of the ledger: the US position is also arriving through a single channel. An unnamed American official told *Al Jazeera* that Washington controls Hormuz, is "in no hurry," and is holding constructive discussions through mediators [TG-671127, TG-671129]. Anonymous, uncorroborated, and perfectly designed to lower the price of waiting.

The structurally important event in this window is that a concession was made in a venue where the concession's domestic audience would not see it. Pezeshkian told Fox News he would relinquish 60% enriched uranium 'within the framework of international law' and the NPT — carried in breaking-flash form by *Al Jazeera Arabic* [TG-668883, TG-668884], by *TASS* [TG-668875, TG-668926], and translated into a numbered Hebrew list by *AbuAliExpress* [TG-669641]. In the Persian-language state feed of the same hours, the uranium line is essentially absent; *Mehr*, *Fars* and *Press TV* run 'we did not choose war' [TG-668929], 'the gift was guided missiles and destruction' [TG-668952] and 'we will keep resisting until the very end' [TG-669438]. Signaling theory has a name for this: it is a costly signal deliberately stripped of its domestic cost. If the audience that could punish you never hears the concession, the signal is cheaper — and therefore, to the receiving side, less credible. The Iraq 2003 comparison is instructive in the opposite direction: there the problem was a regime unable to prove a negative. Here Tehran is able to make a positive offer but is structuring its transmission so that it is deniable at home. The second structural point concerns Araghchi's seven-day Hormuz plan [TG-669127, TG-669129, TG-669137, WEB-96454, WEB-96520]. *Naharnet* reports the conditions include a Lebanon ceasefire [WEB-96527]. This is issue-linkage — bundling a maritime concession with a theater that Iran does not formally control. Linkage expands the bargaining space but also multiplies veto players: any actor capable of continuing fire in south Lebanon now holds an option on the Hormuz deal. Third, and most important methodologically: I want to register that almost everything in this window that looks like a capability datapoint is a claim. The IRGC's '400 kilometers' [TG-668795], the Army deputy's assertion that the Arash drone has been upgraded and that Iran reached technology capable of downing enemy UAVs [TG-669659, TG-669672], the Houthi claim of a 'sensitive target' in Riyadh [TG-669855] — none of these are corroborated in our corpus by anything other than the claimant and its allied media. Read as signals they are informative: Iran is advertising continued capability development *while* offering a seven-day de-escalation plan. Read as facts they are worthless. Finally, the Senate's 49–50 rejection of the War Powers Resolution [TG-669226] removes a domestic constraint that Tehran might have been pricing in. *Al Mayadeen* carries Sanders saying almost all Republicans voted to keep the war going [TG-669079], and *IRNA*, citing Time, reports Republican candidates campaigning to end it [TG-669647]. Both can be true. But an adversary reading only the second would badly misprice how long American latitude lasts.

Three signalling artifacts in this window, and I want to be disciplined about what each one is. First: Safavi's Indian Ocean remark. *Al Jazeera Arabic*'s wire feed carried it in stages — if Washington implements Iran's proposals, transit mechanisms through Hormuz could be studied differently [TG-666807]; a mechanism was reached with Oman but Washington obstructs it [TG-666808]; Iran will always control and manage Hormuz [TG-666809]. *Al Hadath* compressed the whole sequence into one headline about the Khamenei adviser waving expansion of the conflict to the Indian Ocean [TG-666903]. Same speech, and the Saudi-owned outlet selects the maximalist fragment while the Qatari-owned one preserves the conditional structure. The conditional structure is the signal: Iran is publicly advertising a de-escalation mechanism and publicly blaming Washington for blocking it. That is bargaining language dressed as threat language. Second: the negotiation reporting is almost entirely reflected, and it contradicts itself. *Middle East Spectator*, citing Yonhap, says the US rejected the conditions Araqchi conveyed to Witkoff [TG-665616]. Iran's foreign ministry spokesman, via *Press TV* [TG-666045], confirms two hours of message exchange and clarification through Qatari mediation. *Fotros* [TG-666824] reports Fars News dismissing Kyodo and Reuters reporting that Iran wants Hormuz reopened within seven days as unreliable. We are watching a denial of a report we cannot read, about a negotiation we see only in third-party fragments. The honest reading is that we do not know the state of the talks — and that Fars bothering to deny a specific seven-day figure is itself a datapoint. Third, a caution to my own colleagues. Iran's acting defence minister declaring that the defence doctrine has changed [TG-666777, TG-666760] is a claim about posture made from a podium, not evidence of posture. *Al Mayadeen*, citing *The Economist* [TG-666350], reports a prevailing view in Washington, Tehran and the Gulf that the conflict runs to year's end, and [TG-666353] that Iran is in no hurry for a deal that would lower oil prices and help Trump before the midterms. That is a claim about incentives sourced to unnamed officials. Useful as a frame. Not a fact.

Structurally, this window contains two Iranian signals emitted within four hours of each other, and they are not compatible — which is the point. Signal one, roughly 13:40–14:10 UTC: SNSC Secretary Rezaei, carried by *Fars* [TG-664067, TG-664068, TG-664069], *ISNA* [TG-664042] and *Mehr* [TG-664088], states seven conditions, declares there will be no return to prior diplomacy, and says that absent implementation "neither the strait opens nor do negotiations begin." *Al Jazeera Arabic*'s wire treatment added a clock the Persian versions did not foreground: a five-day deadline [TG-664051]. Signal two, 14:12–14:37 UTC: President Pezeshkian at the UNGA — "we believe in diplomacy and will not surrender under military pressure" [TG-664245, TG-664307], "we are ready for dialogue" [TG-664423]. The textbook reading is bad-cop/good-cop. But the corpus suggests something messier. *Radio Farda* [TG-663558, TG-664093] and *abualiexpress* [TG-664616] both report IRGC-linked agencies attacking FM Araghchi for meeting Witkoff "without authority"; *Al Arabiya* headlines a threatened parliamentary summons [TG-664925]. *Fars* itself ran "Three questions for Araghchi" [TG-663340]. Genuine factional contestation and deliberate two-track signalling look identical from outside — and that ambiguity is precisely what makes commitment problems unsolvable. Washington cannot tell whether the seven conditions are Iran's position or one faction's bid. A methodological note I want on the record. Rezaei's claims about the carrier *George Washington* [TG-664021] and a reverse-engineered American submarine [TG-664022] are being treated in some coverage as evidence of a shifted regional balance [TG-664043]. They are not data about capability; they are data about what Tehran wants believed. The only sources are the claimant and its allied media. Feeding them into a capability model corrupts the model. The genuinely novel structural item is the reciprocal-hostage framing. Pezeshkian to the Arab states: "either we live in security and harmony together, or we will make sure that you and us experience insecurity together" [TG-664320, TG-664377], paired with Rezaei's specific threat to regional airports [TG-664141]. This converts Gulf neutrality from a preference into a coerced choice, and it does so publicly — which raises the domestic cost for any Gulf state that complies with the US aviation ban and is later seen to have done so. Compare the Iraqi government's careful deflection: per *Al Mayadeen* citing INA, the refusals come from private ground-handling companies fearing sanctions, not from Baghdad [TG-663584, TG-663585]. That is a state building an escape hatch in advance. On the American side, Rubio supplied the most consequential signalling line: Iran believes the Democrats will win the midterms and that Trump cannot act against it [TG-664333]. Whether or not that is Tehran's actual calculus, publicly asserting it commits Washington to demonstrating the opposite. That is a self-binding statement, and self-binding statements are how crises acquire momentum.

Structurally, this window contains a textbook case of costly-signal ambiguity resolved through source laundering — and analysts should be careful not to treat any stage of it as capability data. Stage one: *Al Jazeera Arabic*, citing *Kyodo*, reported an unnamed senior Iranian official offering to open the Strait of Hormuz within seven days if Washington eased military pressure, adding that the initiative had the Supreme Leader's and the Supreme National Security Council's approval, and that a presidential meeting was ruled out [TG-659770, TG-659771, TG-659772]. Stage two: a parallel *Reuters* version, also anonymous, adding that the proposal was delivered through mediators on 16 September and that the New York delegation has full authority [TG-659846, TG-659849]. Two agencies, two anonymous sourcings, one message — the architecture of a deliberate trial balloon. Stage three is the interesting one: denial from within. *fotrosresistancee* reported that *Fars* called the Kyodo and Reuters accounts unreliable and lacking credibility [TG-660057], and *Al Mayadeen* carried an IRGC officer flatly asserting that the Strait is not open and America is lying [TG-660498]. Stage four: Iranian state television confirmed an Araghchi–Witkoff meeting had occurred — framed as held at the American representative's insistence, with Iran's conditions delivered [TG-661748, TG-661749]. *Trump* separately claimed a three-hour productive meeting [TG-661465], with *Xinhua* naming Witkoff and Kushner [WEB-95635]. The signalling logic is legible. Anonymous placement permits an offer to be made without domestic cost; hardline denial preserves the escalation posture; official confirmation arrives only once it can be framed as the adversary's supplication. Each stage is aimed at a different audience, and 'who asked for the meeting' is doing more work than the meeting's content. One caution against over-reading Trump's UNGA framing as a ladder rung. *AJA* carried him saying the crisis could end after the midterms [TG-661772] and also saying he must decide whether to make a deal or annihilate the Islamic Republic [TG-660773]. Simultaneous maximal and conditional framings are not a ladder; they are optionality preservation. What actually changed materially in this window was not the rhetoric but the aviation cordon — a coercive instrument applied to third parties, which is much harder to reverse than a speech.

The structurally interesting feature of this window is that the same actor issued a maximal threat and a de-escalatory signal roughly three hours apart, and the information environment sorted them into separate stories. The threat: at the General Assembly, Trump framed a binary — a deal that lets Iran become a far greater country, or annihilation of the Islamic Republic, driving it into hell with no chance of survival [TG-660773, TG-661204, TG-660804, WEB-95540]. He paired it with a call for all states to join a complete economic blockade [TG-660764] and asked ICC members to withdraw immediately [TG-660834]. The signal: later the same afternoon he told reporters that US officials had held a three-hour meeting with an Iranian delegation that went well and was productive [TG-661459, TG-661465, TG-661510], that Iran wants to talk and had made contact that day [TG-661511], and that a settlement is coming [TG-661513]. *AbuAliExpress* identifies the American side as Witkoff and Kushner, citing Trump [TG-661506]. Signaling theory predicts exactly this: a coercive public message for the domestic and allied audience, a private channel for the adversary, and an ambiguity that the sender controls. What the theory does not predict is the sender advertising the private channel himself within hours of the threat. That converts a costly signal into a cheap one, because it tells Tehran the annihilation language is priced for the room, not for them. Trump also volunteered the timing marker — a deal likely after the midterms [TG-660706] — while insisting the election does not affect his decision [TG-660778, TG-660779]. Those two statements cannot both be load-bearing. On the input-discipline point I keep making: I would not build any capability inference on the IRGC statements in this window. Mohebbi announcing the Corps has prepared for a long war [TG-659639], Izadi claiming the ability to target large enemy vessels [TG-659940], Ghalibaf claiming missiles can reach any chosen target without restriction [TG-659673, TG-660693] — these are all belligerent claims made during a UNGA week in which Iran's president is in transit to New York [TG-660211]. They are positioning for a negotiation, and they should be read as bargaining posture, not as a capability disclosure. The one datum that is not a claim is the airline closure cascade, because third countries executed it.

Discipline note before anything else: almost every operational claim in this window is single-sourced to a belligerent. The El Gaia mine strike is an IRGC Navy statement [WEB-92940] contradicted by CENTCOM [WEB-92984]. The strike on Aramco's Yanbu refinery is an IntelSlava assertion [TG-637075] with no corroboration in our corpus. The MQ-1 shootdowns are IRGC Public Relations statements [TG-637115, TG-637639]. The thirteen Saudi civilian injuries are a coalition spokesman figure [TG-636823]. None of these are facts; all are signals, and the signalling is what is analysable. The structural event of the window is a negotiation-claim war in which both sides are performing reluctance. Trump says Iran wants a deal «quickly and badly» and he is open to it [WEB-93025, WEB-92935], while — per Xinhua, the only outlet in our corpus carrying this — also demanding that countries pay the US «escort compensation» for Hormuz transit [TG-637166]. Rezaei responds with a single X post: no talks until Iran's conditions are met, do not be distracted by contradictory signals [TG-636680, TG-636695]. That post saturates every ecosystem we track within three hours. What is analytically interesting is that Al Jazeera Arabic stopped covering the exchange and started covering the coverage: «Tehran and Washington — who is eager to negotiate and who is refusing?» [WEB-93075, WEB-93081] and «Who is fighting negotiations inside Iran? An internal battle that goes beyond Washington» [WEB-93022]. When a major regional outlet's lead product becomes meta-analysis of competing reluctance performances, it is signalling that neither side's stated position is being read as its actual position. On escalation architecture, the Iranian Armed Forces General Staff circular is the single most revealing document of the window: a passive-defence directive written for three explicit scenarios — war concluded, in-between (neither war nor peace, e.g. ceasefire), and continuing war [TG-637422]. TRT picked it up and internationalised it [WEB-93064]. A bureaucracy that publishes contingency taxonomy for its own termination conditions is a bureaucracy that has priced termination. That is a more meaningful de-escalation indicator than any spokesman's sentence — and it went almost entirely unread outside the Iranian and Turkish ecosystems.

Discipline first: almost nothing in this window is a datapoint about capability. The King Khalid Air Base strike is a *Yahya Saree announcement* [TG-634508, TG-634509], reproduced by *Xinhua* explicitly as a Houthi claim [WEB-92648], by *Naharnet* [WEB-92662], by *Jerusalem Post* [WEB-92682]. The MQ-1 shootdown is an *IRGC public-relations statement* [TG-634202], reproduced verbatim by *IRNA* [TG-634189], *Fars* [TG-634180], *ISNA* [TG-634225], *Al Manar* [TG-634383], *Al Jazeera Arabic* [WEB-92610] and *TASS* [TG-634241]. Wide reproduction is not corroboration; it is one source with many mouths. What is structurally interesting is the Salalah postponement, because the signalling is contested in a way that changes what the event means. Version one: Oman and Iran jointly postpone "in the interest of regional consensus" — *Press TV* [TG-633614], *Xinhua* [WEB-92499], *Rudaw* [WEB-92490], *TRT World* [WEB-92501]. Version two, twelve hours later: Saudi Arabia asked for it — Baghaei, carried by *Al Mayadeen* [TG-634394], *Mehr* [TG-634453], *Naharnet* [WEB-92657]. Version one preserves the option of resumption; version two assigns blame and forecloses it. That Iran published both — *Fars* still carrying the consensus formulation from its own MFA Gulf director [TG-634427] — suggests an unresolved internal argument about whether to keep the channel open. The substantive question underneath is what a Hormuz arrangement would even be. Oman's energy minister says the strait will open and the situation may be short-lived [TG-634127]. *Al Manar* headlines that the Oman deal does not mean reopening, attributed to Araghchi [TG-634211]. *Al Jazeera Arabic* reports an Iranian official saying Hormuz will not open unless Iranian conditions are met [WEB-92593], and separately that a transit-fee mechanism remains rejected [WEB-92548]. Baghaei's own framing — that the corridor arrangement concerns the sovereign rights of the two littoral states [TG-634393, TG-634448] — is the escalatory move hiding inside the de-escalatory one: it converts an international waterway into a bilateral administration. One more precedent worth flagging. *Bloomberg*, via *Mehr* [TG-634088] and *Press TV* [TG-634335], reports Iran's nuclear chief barred from the IAEA General Conference. Whatever one thinks of the merits, excluding a state's technical delegation from the forum designed to reduce uncertainty about its program removes the last low-cost channel. Historically, that is where miscalculation lives.

Structurally, this window contains one genuine signal and a great deal of noise dressed as signal. The signal is the postponement of the Salalah ministerial, and the interesting thing is that three actors gave three different accounts of causality. *Oman's foreign minister* framed it as serving consensus [TG-633405][TG-633406]. *Iran's MFA Persian Gulf director*, via *IRNA*, said it came at the request of some regional states and by joint Iran-Oman decision [TG-633451], with *ajanews* carrying the same attribution [TG-633445]. *Fars*, per *TASS*, called it an agreement between Tehran and Muscat [TG-633474]. Each version distributes agency differently, and each is doing work: Oman preserves mediator neutrality, Iran avoids looking rebuffed, and nobody names who objected. That matters because hours earlier *Araghchi*, speaking to *Al-Araby Al-Jadeed* and relayed across Iranian state channels, had set the agenda as a new maritime route through Hormuz [TG-632149][TG-632166][TG-632517] while stating via *ajanews* that Iran's condition for reopening the strait is a US return to its Islamabad memorandum obligations [TG-632130]. A precondition addressed to a party not at the table is not a negotiating position; it is a signal to a domestic audience about what will not be conceded. The postponement followed the *Bloomberg* report, carried by *ajanews*, in which the US Energy Secretary called a consensual arrangement with Iran over Hormuz *not a good bet* [TG-633119]. Sequence is not causation, but the sequence is legible. On methodology, I want to flag a discipline problem in how this window's maritime incident is being processed. *UKMTO*, relayed through Iranian state media, reported a vessel struck by an unknown projectile with a fire aboard and the crew evacuated [TG-632213][TG-632169][TG-632071]. *Press TV* converted that into a US attack by a self-propelled underwater missile [TG-632564]. *India's MEA* condemned the attack and named no actor [TG-632878][WEB-92419]. The specificity in the second account is not corroborated in the first or third. A claim that gains precision as it travels away from the reporting authority is a claim to distrust, and I would not use the weapon-type detail as an input to any capability model. The third structural item is deterrence bookkeeping. *abualiexpress* notes that if Iranian state TV's account is accurate, then by the IRGC spokesman's own published equation Iran owes ten American targets in response [TG-633424]. That is an adversary ecosystem holding a belligerent to its own declared retaliation rule — and it identifies the real escalation risk in this window. Published proportionality formulas create commitments that are politically expensive to abandon. The unexplained explosions at Sirik in Hormozgan late in the window, which *IRNA* reported without identifying a source [TG-633598][TG-633591], are exactly the kind of ambiguous event such a formula can be attached to after the fact. Finally, *Trump*'s timeline — the war ending before or shortly after the midterms, Iran wanting a deal badly [TG-632631][TG-632632][WEB-92403] — is an electoral clock, not a diplomatic one, and *Mike Johnson*'s claim that the war is not continuing [TG-633129] is a third variant of the same domestic message. Three American framings, one audience.

The methodological problem this window is attribution, and I want to be disciplined about it before drawing any escalation inference. Start with what is actually sourced. *UKMTO* — an advisory body, not a belligerent — reported a vessel struck by a projectile of "unknown origin" [TG-631170, TG-631207]. That is the epistemic floor. Everything built on top of it is claim. *IntelSlava* asserted the IRGC Navy struck the Riyan Star bulk carrier with an anti-ship cruise missile [TG-631086]; no other source in our corpus carries that vessel name or that attribution. On the second incident, the Qeshm governor's account moved within hours from "enemy" [TG-631437] to *Mehr*'s "details of the US attack" [TG-631756] and *Press TV*'s "US attack" [WEB-92299], with injuries drifting from three [TG-631382] to four [WEB-92299]. *Anadolu* preserved the hedge — "Iran says 'US attack'" [WEB-92257] — as did *Haaretz*, attributing throughout to Iranian media [WEB-92295]. *AbuAliExpress*, meanwhile, told its readers the US destroyed five Iranian tankers overnight [TG-631813], a figure absent from every other source we collect. So: four incompatible attributions of adjacent events, none independently corroborated. Any escalation model fed these inputs produces garbage. What can be modelled is signalling behaviour. Pezeshkian is running two registers simultaneously: "Iran will not surrender" for the domestic and resistance audiences [TG-630966, TG-630975], and to *India Today* "we are not at war with Saudi Arabia; the Houthis have their own issues" [TG-631684, TG-631023]. That is classic audience segmentation — deniability toward Riyadh, defiance toward Washington. Similarly the NPT signal splits: Pezeshkian argues membership proves non-weaponisation [WEB-92285] while a parliament presiding-board spokesman calls exit and doctrinal revision "easily achievable" [TG-631322]. Read structurally, these are not contradictions to be resolved but a bargaining range being advertised — one voice holds the floor, another holds the exit. The Salalah meeting is where that range gets tested.

The analytically interesting feature of this window is that the central military event has no claimant — and the information environment is filling that vacuum with attributions rather than evidence. Start with source discipline. The Islamic Resistance in Iraq's denial, carried by *Al Mayadeen* [TG-629285], *Al Jazeera Arabic* [TG-629357] and *Al Manar English* [WEB-92110], uses a formulation that is structurally unusual: the accusation is "an honour we do not claim." That is not a standard denial; it declines liability while retaining prestige, and it is paired with an offer to join a government inquiry [TG-629320]. Trump's counter-attribution — "probably Iran," per *Al Jazeera Arabic* [TG-629399], *Radio Farda* [TG-629490] and *BBC Persian* [TG-629652] — is offered without stated evidence. Neither is a datapoint about capability; both are signals. What *is* new evidence is procedural: *Reuters* via *Al Jazeera Arabic* reports Iraqi military sources closing the Tayyib border area after discovering launch sites [TG-630042], with traces near the Duwayrij river [TG-630044], fifteen platforms seized per *Al Arabiya* [TG-630885], and a police chief dismissed [TG-629463] [WEB-92066]. Baghdad then approves Iran's request for a *joint* investigation [TG-629918] [WEB-92088]. Game-theoretically that is elegant: a joint probe converts an attribution problem into a shared procedural commitment, which lets Baghdad avoid choosing between Riyadh and Tehran. Watch whether it produces findings or simply produces time. Second structural point: signalling through pre-emptive deflation. Ahead of Monday's Muscat meeting, Iranian officials are talking the agreement *down* — spokesman Baqaei says an understanding with Oman does not mean the strait is secure [TG-630162] [TG-630175], a parliamentary security committee member says a temporary corridor must not be confused with reopening [TG-630026], and *Mehr* headlines that Hormuz will not open under it [TG-630365]. Simultaneously Gulf outlets frame Pezeshkian as hinting at reopening if the blockade ends [TG-629546]. Deliberately lowering expectations before a deal preserves the coercive asset while banking the diplomatic credit — a costly-signal move, and a legible one. Third: absence as signal. Bahrain announces it will not attend any meeting including Iran before diplomatic relations resume [TG-629889] [WEB-92085]. Iraq announces it will attend [TG-630916]. The attendance list is itself the regional alignment map, published in real time.

The most analytically interesting event in this window is not a strike. It is the collision of two incompatible official timelines inside the same information space, within hours of each other. At the Dallas convention, Trump asserted the war ends immediately after the November midterms, with oil collapsing and gasoline at two dollars [TG-622368], that the US is winning and controls Hormuz [TG-622633], that Iran is collapsing [TG-622656], and that he should have the strait named after him [TG-622634]. Almost simultaneously the *Wall Street Journal*, reaching our corpus through *TASS* [TG-622525], *Almayadeen* [TG-622519], *IRNA* [TG-622613] and *L'Orient Today* [WEB-91326], reported that Vance, Rubio and other senior advisers privately assess the conflict may run to the end of the presidential term. *OSINTdefender* rendered this as a war potentially lasting until January 2029, with a sustained naval blockade [TG-622883]. Signaling theory has a straightforward reading here, and I want to resist the obvious one. The obvious reading is that the public claim is cheap talk and the leak is the true belief. But note that Trump himself supplied the bridging mechanism: he told reporters, per *BBC Persian*, that Iran has no political interest in ending the fighting before the American midterms [TG-623091]. That converts an electoral-calendar claim into a strategic-interaction claim — both sides waiting on a date. Whether it is true is unknowable from here. What matters is that it is a *legible* commitment device, and both the WSJ leak and the stage claim can be read as consistent with it. Second structural point, and a caution to my colleagues. Almost every operational input this window is a claim, not a datapoint. IRNA reported "enemy projectiles" striking Sirik [TG-622269] — and *Middle East Spectator* immediately noted only IRNA had reported it [TG-622293], while Hormozgan officials had said the sounds came from the sea [TG-622277]. The IRGC claims it hit fighter maintenance shelters in Jordan [TG-622758]. The Saudi coalition claims Houthi missiles at three cities [TG-622322, WEB-91111]; Ansarallah's spokesman claims Aramco and Khamis Mushait [TG-623237]. Each is a bid to be treated as capability evidence. Very few are corroborated. The one that *is* — the CBS damage report at Muwaffaq Salti [TG-622655, WEB-91201, WEB-91321] — is the one neither belligerent originated. Third: the escalation architecture is being rebuilt around a new object. Israel's defense minister Katz, via *AJA news*, threatened that any attack on Israel from Iran, for any reason and from any place, would meet an unprecedented response including Iran's main energy facilities [TG-623561, TG-623562, TG-623563]. Trump separately floated striking the site *Al Jazeera Arabic* renders as Jabal al-Fas — Pickaxe Mountain — citing observed activity [TG-622671, TG-622672, WEB-91208], with *Jerusalem Post* reporting analysts see a construction surge there [WEB-91206]. That is a target being publicly pre-announced. Pre-announcement is either deterrence or preparation of the record. It is never neutral.

Two mirror-image election theories are now on the record, and they are the most structurally interesting objects in this window. The first: *Trump*, speaking to reporters as relayed by *TRT World* [WEB-91124], *Al Jazeera* [TG-621867], *Radio Farda* [TG-621985] and *BBC Persian* [TG-622033], said the war with Iran will end "immediately after" the November midterms, because Tehran "won't be able to hold on," that gasoline will fall below two dollars a gallon afterward [TG-621997], that there will be further attacks on Iran [TG-621866], and that Washington is not currently seeking a deal because Iran has "very little left" [TG-622054]. He also asserted Iran is trying to influence the vote [TG-621908, TG-621996]. The second: per *Haaretz*'s summary of a New York Times report, US intelligence believes Iran wants to escalate before the midterms [WEB-90848]. Each side has publicly attributed to the other an election-timed escalation motive. In signalling terms this is a dangerous configuration: both parties now have a stated reason to read the other's next move as electorally motivated rather than militarily necessary, which is exactly the condition under which tit-for-tat rounds get misclassified. Announcing a terminal date is also a rare and costly move. It converts every Iranian action before November into evidence for or against Trump's prediction, and it gives Tehran a published deadline to price into its own decisions. Against that, the IRGC put a bargaining position in public — the first explicit one I have seen in this corpus. Ending the current situation requires the US to stop interfering with Iran's nuclear and missile capabilities, Israel to withdraw from Lebanon, the blockade of Yemen to end, and Iranian funds to be released [TG-620662, TG-620696, TG-620637, WEB-90911, WEB-90928]. Whatever one thinks of the terms, publishing them is a de-escalatory grammar: it converts an open-ended conflict into a list. Alongside it sits a commitment device — Mohebi's "if the enemy strikes two of our targets, we will strike twenty of theirs" [TG-620547, TG-621661]. Ratio doctrines announced in public are hard to walk back and easy to trigger. Methodological caution on the operational inputs. "Two US warships struck" is an IRGC claim, denied specifically by CENTCOM [TG-620408]; *fotrosresistancee* went further and attached hull numbers sourced to an Iranian television presenter's military contacts [TG-621697], which is a claim about a claim and should not be treated as capability evidence. The Sirik and Qeshm explosions late in the window are the cleanest illustration: IRNA reported enemy projectiles hitting areas of Sirik [TG-622269] while Hormozgan province said the sounds came from the sea [TG-622277], and *Middle East Spectator* posted an explicit note that US airstrikes on Sirik remain unconfirmed and single-sourced to IRNA [TG-622293]. Structural models are only as good as their inputs, and this window's inputs are mostly announcements.