Nov 2, 2026
Korea, Cyprus, Transnistria and Karabakh show that a freeze holds only as long as an outside power is willing to be shot at to keep it. The guarantee Ukraine has been offered so far is built on the wrong model.
On 22 October, Volodymyr Zelensky said the sentence Kyiv had avoided for four years. “Yes, right now, we have to freeze,” he told reporters, calling a halt along current lines “the minimum of what we need from Russia today.” Within days the Kremlin said Vladimir Putin had agreed “in principle” to restart trilateral talks with Washington and Kyiv, most likely in Abu Dhabi in mid-November. After a year in which, as Modern Diplomacy argued in March, a deal looked out of reach, Western capitals are already arguing about where the line should run.
That is the wrong argument. The history of frozen wars says the line matters far less than who stands behind it, and what they do when it is tested. A freeze holds only while a guarantor stays committed — and the guarantee on offer to Ukraine looks less like Korea than like Karabakh.
Four Lines, Four Patrons
Stay ahead of the geopolitical week.
MD Briefing delivers expert analysis across five global fronts — the Indo-Pacific, energy, geoeconomics, European security, and the Middle East — every Monday morning. Free.
The precedents usually cited are not one model but four. In Korea, the armistice of 27 July 1953 stopped the shooting without a peace treaty; ten weeks later Washington signed a Mutual Defense Treaty with Seoul, and roughly 28,500 US troops are still there. In Cyprus, Turkey’s 1974 invasion left the island partitioned along a buffer zone patrolled by the UN force UNFICYP, with a large Turkish garrison in the north. In Transnistria, a 1992 ceasefire left a Russian-backed enclave inside Moldova, guarded by about 1,500 Russian soldiers. In Nagorno-Karabakh, a 1994 ceasefire left Armenian forces in control until Azerbaijan retook most of the territory in 2020 and the rest in September 2023 — three years into a Russian peacekeeping mission.
For Ukraine, the offer on the table is the Paris Declaration of January 2026: a French- and British-led multinational force deployed “away from the contact line,” a US-led sensor network to monitor a ceasefire, and an Article 5-style assistance pledge whose activation would be a political decision and whose American version would need Congress.
Why the Line Is the Least Important Part
Start with the line that has held longest. The Korean front has barely moved in 73 years, through the 2010 sinking of the South Korean corvette Cheonan and Pyongyang’s acquisition of nuclear weapons. What has stayed constant is that an attack across it would kill American soldiers on the first day. US forces in Korea are not monitors. They are a tripwire: a commitment made credible by placing the guarantor’s own people in the path of an attack, so that Washington never has to decide, in the moment, whether to keep its word.
Cyprus is a less comfortable case, because its guarantor is also the occupying power. Turkey invoked the 1960 Treaty of Guarantee to intervene and has never left. The line holds because Ankara’s commitment to the north has never wavered, and because the Greek Cypriot side, which lost, had no military route back and found a better one in EU membership in 2004. The freeze has bought half a century without war. It has not bought a settlement — the UN’s Annan Plan died in a referendum that same year — because the side that could reopen the question has no reason to.
Transnistria shows what happens when a guarantor’s commitment fades but nobody wants to fight. Since 2022, Russia’s garrison has been stranded behind Ukrainian territory; since 1 January 2026, Ukraine and Moldova have closed the remaining supply routes by land, rail and air. Moscow can no longer reinforce or resupply its troops. The line has not collapsed, because Chișinău has chosen customs posts and gas bills over tanks. But the freeze is thawing on Moldova’s terms. The guarantor’s retreat is deciding the outcome; only the method is peaceful.
Karabakh is the case that should worry Kyiv, and it is usually told wrongly. The front did not collapse the moment Moscow looked away. It collapsed after a long public test that Moscow failed. Under the November 2020 ceasefire statement, nearly 2,000 Russian peacekeepers were to guarantee the Lachin corridor, the only road between Armenia and Karabakh. From December 2022, Azerbaijani “eco-activists” and then a state checkpoint shut it. Armenia’s prime minister told Putin to his face that the peacekeepers had lost control of the corridor; in February 2023 the International Court of Justice ordered the road reopened. Russian troops stood aside for nine months as food and fuel ran out. When Baku attacked on 19 September 2023, Karabakh’s forces gave up within about a day, and more than 100,000 people — almost the entire Armenian population — fled within a week. By June 2024 the last peacekeepers had gone home.
The lesson is not that guarantors vanish overnight. It is that a guarantor whose troops are present but not committed to fight is worse than none: it lulls the defender and shows the challenger exactly how much it can get away with. Azerbaijan probed, watched Moscow do nothing, and then moved.
The strongest objection is that Karabakh fell because the balance of power shifted, not because a guarantor failed. Azerbaijan’s oil revenue built a far larger army, armed with Turkish and Israeli drones, and in October 2022 Armenia itself recognised Azerbaijan’s territorial integrity, implicitly including Karabakh. All true. But balances of power always shift over the life of a frozen conflict; that is what the guarantor is there to absorb. A North Korean nuclear arsenal is a bigger shift than any drone fleet, and the Korean line did not move, because Washington’s commitment was built not to depend on its mood in any given year. Karabakh’s depended entirely on Moscow’s mood — and by 2023, as Modern Diplomacy noted at the time, Moscow needed Baku and Ankara more than it needed Yerevan.
By that test, Ukraine’s offer has a Karabakh shape. A reassurance force kept away from the contact line is not a tripwire; a probe at the line would never touch it. Sensors that “verify breaches and assign responsibility” describe a referee, not a guarantor. An assistance pledge that must be triggered by a political decision is a promise to deliberate later. Russia, whose foreign ministry says it will not accept foreign troops in Ukraine “in any format”, will read those gaps the way Baku read the Lachin road.
Three Ways the Freeze Plays Out
Base case — a monitored, untested freeze (roughly 50%). Talks after 10 November produce a ceasefire along current lines in the first half of 2027, with European hubs in central and western Ukraine and US-led monitoring, but no ratified American obligation. The line holds for a year or more because both armies are exhausted. The key assumption is that Moscow needs time to rebuild more than it needs another offensive. This buys Ukraine time, not security; its value depends on how fast Kyiv’s own army becomes the real deterrent.
Downside — the Lachin pattern (roughly 30%). The same freeze, followed within 12 to 24 months by a calibrated Russian probe below the threshold of war: a “humanitarian” blockade of a frontline town, the seizure of a contested village, an energy cut-off dressed up as a technical dispute. The European force, stationed far behind, does not engage; the monitors assign responsibility and nothing follows. Each unanswered probe lowers the price of the next. For investors and insurers pricing Ukrainian reconstruction, this is the scenario not yet in the numbers: the risk is less a return to full-scale war than erosion by increments.
Upside — the Korean pattern (roughly 20%). Washington sends a binding assistance obligation to the Senate, and at least one allied contingent is stationed close enough to the line to be in harm’s way. Moscow objects but accepts a ceasefire anyway because its economy needs relief from sanctions. This is the only version in which today’s line is likely to still be the line in 2040.
Watch the Senate, Not the Map
A frozen front line buys time. Whether that time belongs to the defender or the attacker depends on one thing: whether the power behind the line has made its commitment automatic and costly to break. Korea’s guarantor did, and the line has lasted 73 years. Karabakh’s did not, and the 2020 line lasted less than three. Ukraine is being offered monitors, rear-area hubs and a promise to decide later.
So the thing to watch is not the map that comes out of Abu Dhabi. It is whether the White House submits any US assistance obligation to Congress for ratification, and whether any allied soldier is placed where a Russian advance would have to go through them. If neither happens, the ceasefire will be a pause in which Moscow chooses the time and place of the next test. Lines do not hold wars back. People willing to be shot at do.
Probabilities reflect the author’s analytical judgment based on currently available information and are not statistical forecasts.
—