Will the ceasefire between Iran and the United States continue to hold 7 days from today?

TL;DR

My forecast is 79% that the U.S.–Iran ceasefire is still treated as in effect on June 7, 2026.

Context

The forecast date is May 31, 2026, so the resolution date is June 7, 2026. I treat a ceasefire as in effect today. The April 7-8 pause was publicly announced as a two-week U.S.–Iran ceasefire tied to reopening the Strait of Hormuz, and the White House later said the February 28 hostilities had terminated with the April 7 ceasefire (Axios, Apr. 7; White House OMB, May 20).

The live issue is not whether there will be zero violence. There has already been direct violence inside the ceasefire period. The question is whether reliable sources will say by June 7 that direct U.S.–Iran hostilities have resumed in a sustained or large-scale way. As of the latest major reports I found on May 31, officials and AP were still using the frame of a fragile ceasefire that remains in effect, not a war that has restarted (CENTCOM, May 30; AP, May 30).

Evidence

The historical backbone points to high one-week survival, but not a safe ceasefire. The best broad dataset I found is the ETH/PRIO Civil Conflict CeaseFire dataset, first published online in 2022, covering 2,202 ceasefires in 109 civil conflicts across 66 countries from 1989 to 2020; its units are unique ceasefire IDs and actor-target-ceasefire observations, not interstate wars (ETH/PRIO dataset article). In that dataset, simple cessations of hostilities are 1,670 of 2,202 cases, or 75.8%; monitored cessations are 374 cases, or 17.0%; definitive ceasefires are 158 cases, or 7.2% (ETH/PRIO dataset article). The same article says failed nationwide ceasefires had a median duration of 65 days to a 25-battle-death threshold and 193 days to a 100-battle-death threshold; using the 65-day median as a crude exponential hazard gives $e^{-\ln(2)\cdot7/65}=92.8\%$ survival for one more week (ETH/PRIO dataset article). I discount that hard because this case is a direct interstate confrontation, has an active blockade, has no public monitoring architecture, and is inside a live nuclear negotiation. I also do not discount it all the way to 50-50 because the resolution threshold here is higher than “any violation.”

The ceasefire label has already absorbed serious incidents. On May 25, AP reported U.S. “self-defense” strikes in southern Iran, including missile launch sites and minelaying boats, while CENTCOM said it was acting with restraint during the ongoing ceasefire (AP, May 25). On May 28, CENTCOM said Iran launched a ballistic missile toward Kuwait and five one-way attack drones near the Strait of Hormuz, calling this an egregious ceasefire violation rather than the end of the ceasefire (CENTCOM, May 28). On May 30, CENTCOM said U.S. forces disabled the Gambia-flagged M/V Lian Star with a Hellfire missile and had disabled five commercial vessels and redirected 116 while enforcing the blockade “as a ceasefire with Iran remains in effect” (CENTCOM, May 30). AP’s May 30 account put the operational count at six stopped ships, one allowed through, and 116 redirected, while still writing that the fragile ceasefire had held since April 7 (AP, May 30).

The diplomatic track is the main pro-YES evidence. AP reported on May 28 that U.S. and Iranian negotiators had reached a tentative agreement to extend the ceasefire by 60 days and begin nuclear talks, while Vice President JD Vance said it was unclear whether President Trump would sign it (AP, May 28). AP reported on May 29 that Trump’s Situation Room meeting ended without a decision, that Iran said the agreement was not finalized, and that nuclear terms remained unresolved (AP, May 29). Axios then reported on May 31 that Trump wanted the deal but requested changes, especially around Iran’s nuclear material, starting another back-and-forth that could last several days (Axios, May 31). That puts the stress point inside this forecast window, but it is still a negotiation channel, not a rupture.

The main anti-YES evidence is that the unresolved issues are not cosmetic. Reuters reported on May 29 that Trump demanded immediate unrestricted Hormuz transit and dismantling of Iran’s nuclear-weapons capacity, while Iranian sources said nuclear issues were not part of the potential ceasefire-stage deal and that strait management had to be decided by Iran and Oman (Reuters via Investing.com, May 29). Reuters also reported that Trump faces pressure to reopen Hormuz and lower gasoline prices before the November 2026 congressional elections, while Iran wants sanctions relief, frozen funds, U.S. withdrawal from the region, and coverage of Lebanon in any peace deal (Reuters via Investing.com, May 29). AP reported on May 31 that Israel’s deepest push into Lebanon in more than 25 years complicates the Iran ceasefire extension because Tehran wants Lebanon included too (AP, May 31). Defense Secretary Pete Hegseth said on May 30 that the U.S. was ready to restart attacks if no deal could be reached, according to Reuters coverage carried by GMA (Reuters via GMA, May 30).

My scenario model treats the next week as four public-status paths. These are judgmental buckets, anchored by the historical one-week survival prior and adjusted for the live blockade, the near-but-not-signed MOU, and the repeated labeling of incidents as violations rather than a resumed war.

This gives $P(\text{NO})=21.1\%$, so $P(\text{YES})=78.9\%$. I round that to 79% in prose. The number is above the simple “fragile ceasefire” gut read because the resolution criteria require a public, sustained, large-scale breakdown. It is below the generic 90%+ one-week ceasefire base rate because the parties are operating in the same maritime battlespace, the deal is still unsigned, and the nuclear/Hormuz/Lebanon terms can still snap.

What's non-obvious

The non-obvious point is that this ceasefire is a political label more than a clean stop in firing. The U.S. has struck targets in Iran, Iran has launched missiles and drones, and the U.S. has fired on blockade-running vessels, yet the dominant official and wire-service language still calls these violations, self-defense actions, or tests inside an ongoing ceasefire (AP, May 25; CENTCOM, May 28; AP, May 30). For this question, that label matters. Another limited exchange next week does not automatically make the answer NO.

The second non-obvious point cuts the other way. The blockade is both leverage and a fuse. CENTCOM’s May 30 count of five disabled commercial vessels and 116 redirected ships shows regular coercive contact, not a quiet ceasefire (CENTCOM, May 30). The Strait of Hormuz also carried about 20 million barrels per day of crude oil and oil products in 2025 and just over 112 bcm of LNG, close to 20% of global LNG trade, so both sides have strong reasons to keep talking and strong incentives to use pressure around the strait (IEA Hormuz factsheet).

Limitations

The biggest gap is that no final 60-day MOU text was public as of May 31 at 22:51 UTC. The latest major reporting said the draft was tentative, Trump had requested edits, and Iran had not finalized approval (AP, May 29; Axios, May 31). That makes the forecast unusually sensitive to a few decisions expected within days.

The historical base rate is adjacent rather than exact. ETH/PRIO is a civil-conflict dataset from 1989 to 2020, while this case is a direct U.S.–Iran interstate war embedded in an energy chokepoint and a regional war (ETH/PRIO dataset article). The current question’s resolution rule is also a reporting-and-classification rule. If major outlets or the resolver treat a future “self-defense” strike as the public end of the ceasefire, the forecast would be too high; if leaders keep calling even serious exchanges “violations” while talks continue, the forecast would be too low.