PLA forces land on Taiwan

How I read the question

For YES, Taipei or Washington must officially state that People's Liberation Army personnel physically reached Taiwan's main island in a real operation. Blockades, outlying islands, coast guard actions and drills do not count. In practice this means a PLA assault or raid on Taiwan itself.

The other qualifying paths are negligible:

• Relief or evacuation with Taiwan's consent: both main parties have refused PLA help in past disasters.
• A covert team that Taiwan confirms is PLA: Taiwan has recorded 29 landings by Chinese nationals in small boats since 2025, and none was identified as PLA (Taiwan News, 11 Aug 2026).

If a real landing happened, Taiwan's defence ministry or presidential office, or the US Defense Department or White House, would almost certainly name the landing sites within days. I put the chance of a qualifying confirmation, given a landing, at about 97%.

Current status (as of 13 Oct 2026)

• No landing has occurred. This week's PLA live-fire drills are along China's own coast, far from Taiwan (Focus Taiwan, 12 Oct). Incursions into Taiwan's air defence zone averaged about 160 a month in 2026, against about 300 in 2024–25 (ISW/AEI, 6 Oct).
• US–China détente: Trump visited Beijing in May 2026; Xi visited Washington on 22–24 Sep; an Asia-Pacific summit in Shenzhen follows in November. A $14bn US arms sale to Taiwan is on hold (ISW/AEI, 25 Sep).
• US intelligence: the March 2026 annual threat assessment said China does not currently plan to invade in 2027 (Al Jazeera, 19 Mar). Reuters reported on 1 Oct that US officials see an invasion before 2028 as increasingly unlikely, with 2028 "sporty" if Lai or another candidate from his ruling party wins re-election (via Taipei Times, 2 Oct).
• PLA leadership gutted: Zhang Youxia and Liu Zhenli were expelled on 21 Sep, leaving China's top military commission with two active members (Modern Diplomacy, 30 Sep). CSIS (24 Feb) found about 52% of senior leadership posts hit by purges; Fravel says rebuilding will take years. Jamestown (Oct 2026) also expects near-term risk to fall but long-run risk to rise.

Base rates and other forecasts

• History: the PLA has not landed on Taiwan in 77 years. A simple rule-of-succession estimate gives about 1.3% a year.
• Expert surveys run higher:
  • A 2024 Forecasting Research Institute survey put a Chinese invasion before 2030 at 25% (expert median) and 19% (superforecaster median, only 11 respondents).
  • In an Atlantic Council survey (Feb 2026), 70% of respondents expect an attempt to take Taiwan by force within ten years.
  • Good Judgment superforecasters (Jul 2026) give only 12% to either side accusing the other of using a weapon before July 2028.
• Polymarket (pulled from its API), for invading any Taiwan-held territory: 2.2% by end-2026, 7.5% by June 2027, 10.5% by end-2027. I computed its price history:
  • The 2025 contract stood at 12.5% in January 2025 and resolved No.
  • The 2026 contract stood at 11.5% on 1 Jan 2026 and is now 2.2%.
  • This looks like a recurring floor of roughly 10% a year driven by long-shot betting, so I discount it heavily.
• Thin Kalshi markets on a US "do not travel" advisory for Taiwan are higher still; I give them little weight.

Model

I built yearly hazards for a landing on Taiwan proper, spread across quarters with mild weighting toward the April and October sailing windows, then multiplied by 0.97 for confirmation:

Period	Hazard
Rest of 2026	0.08%
2027	0.8%
2028	1.2%
2029	1.9%
2030	2.1%
2031	2.2%

The 2028–31 figures come from a two-scenario mix of Taiwan's January 2028 election:

• Ruling party wins (about 45%): about 12% chance of a landing by 2031.
• Opposition Nationalist Party wins (about 55%): about 1.5%.

Accumulated, this gives about 0.85% by end-2027, 2% by end-2028, 3.8% by end-2029 and about 7.8% by end-2031.

Strongest case against

Crowd and expert forecasts are two to three times higher. Other arguments for a higher number:

• The purges may be removing cautious generals.
• The US is distracted by the Iran war and Trump is ambivalent about defending Taiwan.
• Xi, now 73, may want unification during his rule.

I moved my 2029–31 hazards up somewhat for these reasons. I did not go further for four reasons:

• Recent US intelligence points to delay.
• Experts expect a blockade, which does not count here, before any invasion (CSIS 2024 survey).
• Beijing's political strategy toward Washington and Taiwan's opposition is currently working.
• The betting-market prices look inflated on the evidence above.

What would raise the forecast most: the ruling party winning in January 2028, visible mobilization (such as requisitioning civilian ferries), or a blockade.

Question. By the last day of each quarter, has Taiwan's government or the US government confirmed that People's Liberation Army forces have landed on the main island of Taiwan in a military operation?

Definitions.

• People's Liberation Army (PLA) forces: serving personnel of any PLA service or branch acting under military orders. Coast guard, police, maritime militia, civilian personnel, or unidentified Chinese personnel alone do not count unless the confirming government identifies participating personnel as PLA forces.
• Main island of Taiwan: the island of Taiwan itself, including its beaches, ports, and airfields. Landings on Penghu, Kinmen, Matsu, Dongsha/Pratas, Taiping/Itu Aba, or any other separate island do not count.
• Landed: PLA personnel have physically reached land on the main island from the sea or air, including amphibious landings, disembarkation at a port, parachute or helicopter insertions, and arrival aboard aircraft that land at an airfield. No minimum force size, duration, territorial control, or successful follow-on operation is required.
• Military operation: an actual operational deployment of PLA forces under military orders, including an assault, raid, reconnaissance insertion, occupation, evacuation, or relief operation. It need not be opposed, involve combat, or lack Taiwan's consent. Exercises, ceremonial visits, military exchanges, individual travel, defections, accidental arrivals, and emergency landings unrelated to an operational insertion do not count.
• Confirm: an official statement affirmatively says the landing occurred, rather than merely reporting allegations, warning of a possible landing, or describing a plan, simulation, or attempted landing.

Resolves YES if either government publishes a qualifying confirmation establishing all the elements above. Confirmation by both governments is not required, and a denial by Beijing or by the other confirming-eligible government does not veto a qualifying confirmation. An operation described as an exercise by Beijing counts if Taiwan's or the US government's confirmation identifies it as an actual operational landing meeting these criteria.

Exclusions (resolve NO). No qualifying confirmation by the quarter-end; landings only on outlying islands; offshore or airborne activity without personnel reaching the main island; blockade or quarantine alone; missile strikes; drones, unmanned vehicles, or equipment reaching land without PLA personnel; and forces attempting a landing but being stopped before reaching land. Reports saying only that forces landed “in Taiwan,” without identifying the main island or a location on it, are insufficient.

Resolving source. Official publications, releases, or government-published transcripts or recordings from Taiwan's central government or the US federal executive government only. Eligible issuers include Taiwan's Ministry of National Defense (PLA Activities and press releases), Office of the President, Executive Yuan, and Ministry of Foreign Affairs; and the US White House, Department of Defense and military commands including US Indo-Pacific Command, Department of State, and intelligence agencies. Their institutional successors and other central/federal executive agencies may also confirm. Statements must be issued institutionally or by an official speaking in an official capacity and published on an official government channel. Legislators' personal statements, campaign statements, media paraphrases, anonymous officials, and Chinese government statements alone do not count. This is a narrower resolving-source rule than Public confirmation (A2). Apply the Resolution window and Resolution date (A3) conventions; the date of the qualifying government publication, not the landing date, determines resolution.