"Sarah Paine — Why wars are so difficult to end" — Dwarkesh Patel

Why this is in the vault

This is the closing lecture in Dwarkesh's Sarah Paine series and gives the sharpest single framework yet for reasoning about how wars actually end (culminating points, limited vs. unlimited objectives, value of the object) rather than how they start or who's "winning" a given battle. It's a direct analytical input for reading the Russia-Ukraine war's endgame, which sits upstream of the geopolitical-risk assumptions in RDCO's Markov capital-cycle thesis.

Episode summary

Sarah Paine, military historian and 24-year Naval War College faculty member, delivers a lecture on war termination: why wars are easy to start, hard to end, and even harder to end on the terms you wanted. She builds her framework from Clausewitz (culminating point of attack, culminating point of victory, the "blown flask" metaphor for regime collapse, limited vs. unlimited objectives) and Sun Tzu (death ground), then applies it across five case studies: the Russo-Japanese War as a model of successful limited-war termination, the 1991 Gulf War as an even cleaner success (the US didn't even pay for it), the Korean War as a case of exceeding the culminating point twice, the American Revolution and WWI as regional wars that escalated into global ones, and Japan's occupation of China plus the 2003 Iraq War as cases of "victory denial" — conventional victory followed by an insurgency that never lets the winner stabilize what it won. She closes by proposing a checklist for reading any ongoing war (Ukraine included): who has limited vs. unlimited objectives against whom, how do allied objectives align internally, and what is each side's relative "value of the object."

Key arguments / segments

• [00:00:00] How Japan beat a far bigger Russia. Sets up the Russo-Japanese War (1904-05) as the model case: Japan fought for limited objectives (Russia out of Korea/Manchuria) and got everything it wanted plus more, because it never let its war aims exceed what it could sustain.
• [00:08:18] Clausewitz's culminating points of victory. Introduces culminating point of attack (a battle/campaign concept — go too far and "the scale turns") and culminating point of victory (the whole-war version); a key warning sign you've exceeded it is third-party intervention.
• [00:14:24] What finally made Russia give up? Walks five competing explanations for Russia's 1905 capitulation — the domestic revolution, the Battle of Mukden, Russia running out of money/credit, the naval destruction at Tsushima, or the loss of Sakhalin Island — as a teaching exercise in evidence-based argument, not a single answer.
• [00:26:54] Gulf War: the war the US didn't pay for. Bush Sr.'s explicit, narrow objective (National Security Directive 54) plus a 34-country coalition meant the US netted more in allied payments ($50B+$14B) than it spent, for under 300 coalition deaths.
• [00:30:31] Korea and the cost of shifting war aims. The US exceeded its culminating point of victory twice — first Kim Il-sung invading the South, then MacArthur pushing to the Chinese border and triggering Chinese intervention — turning a plausible quick war into a permanent, expensive stalemate.
• [00:34:48] Sun Tzu's "death ground." Unlimited objectives (regime change) put the losing side on death ground, where surrender means annihilation, so they fight to the death — the mechanism behind Ukrainian resistance to Russia's unlimited objective, and behind failed occupations generally.
• [00:39:32] How regional wars go global: WWI to Pearl Harbor. Large theaters + protracted timelines give third parties room to pile in (French/Spanish entry into the American Revolution; misaligned allied objectives turning a regional Balkans dispute into WWI; Smoot-Hawley and Japan's resource desperation cascading into the Pacific War).
• [00:46:38] Why insurgencies deny victory: China to Iraq. "Victory denial" strategy — Japan in China, the US in Vietnam/Iraq/Afghanistan — is a cost-imposition play that's ruinously expensive for both sides but especially for the occupier, because "value of the object" (how much the fight matters to you) is almost always higher for the people who live there than for the intervener.
• [00:51:06] Questions to ask before any war. Closing framework for reading the Ukraine war and any live conflict: map limited vs. unlimited objectives on all sides, check alignment of allied war aims, compare relative value-of-the-object, and assess geographic sanctuary/access.

Notable claims

• Japan lost ~45,000 men (an entire army's worth) in frontal assaults on Port Arthur before taking 203 Meter Hill, the position that let it destroy Russia's fleet in harbor — a costly but ultimately correctly-timed culminating point of attack.
• The 1991 Gulf War coalition paid the US over $50B directly plus $14B to other coalition partners; US/coalition deaths totaled 392 against 20,000-35,000 Iraqi military deaths — an unusually lopsided and cheap war for the intervening side.
• Only three European states lacked a legislature in 1904: Russia, the Ottoman Empire, and Montenegro — part of Paine's case that Tsarist illegitimacy, not just battlefield losses, drove Russia's 1905 collapse.
• Trans-Siberian Railway troop-transport capacity grew roughly 5x over the war's course (20-40k/month early on to 100k/month by the Battle of Mukden) — had that capacity existed from the start, Russia would likely have won the Russo-Japanese War outright.
• Paine explicitly frames Ukraine as fighting an unlimited-objective war forced on it by Russia, which is why Ukrainian resistance doesn't fit typical expectations for a smaller power facing a larger one — it has been put on death ground.

Guests

Sarah Paine — military historian, 24-year faculty member at the US Naval War College, specializing in East Asian military history and grand strategy (China, Japan, Russia, the Pacific War). Author of The Japanese Empire and The Wars for Asia 1911-1949. This is the final installment of Dwarkesh's multi-part lecture series with her, building on the continental/maritime framework introduced in the 2026-06-09 episode.

Mapping against Ray Data Co

Mapping: medium

• Geopolitical risk timing for the Markov capital-cycle thesis. The prior Paine notes in this vault map to why US-China/Russia tension is structurally durable; this lecture is the complementary piece on how and when such conflicts terminate. Paine's framework for reading Ukraine (limited vs. unlimited objectives, value-of-the-object asymmetry, culminating-point signals like third-party intervention) is a direct lens for assessing when/whether the war resolves — a live input to energy, defense, and European-equity exposure inside the capital-cycle thesis, not just background color.
• Negotiation and conflict-escalation mental model. "Value of the object" (the side closer to the stakes will absorb more cost) and "culminating point" (the point past which further pushing degrades your position) generalize past war — useful frames for reading how far to push in any adversarial negotiation or competitive standoff, including business ones.
• No direct RDCO product/service tie. Same caveat as the earlier Paine notes: this is worldview/strategic-thinking material, not an operational or client-facing input. File as a mental-model anchor, not a project deliverable.

Related

• [[2026-06-09-dwarkesh-sarah-paine-russia-china-geography]] — the foundational continental/maritime framework this closing lecture builds on
• [[2025-12-19-dwarkesh-sarah-paine-why-russia-lost-cold-war]] — same speaker/series, prior installment mapped to the Markov capital-cycle geopolitical-risk thread
• [[2025-10-31-dwarkesh-sarah-paine-russia-sabotaged-china]] — chip-fab/industrial-capacity mapping from the same series
• [[2025-09-05-dwarkesh-sarah-paine-hitler-starved-britain]] — resource-scarcity/strategic-overreach mapping from the same series