{"rewrite":{"id":"r_b066c45a7280aac09f0a6cd3","clusterId":"c_a0e952b90abec499bc69b500","slug":"ataraxos-ai-beats-stratego-world-champion-15-1-with-four-draws","model":"deepseek-v4-1-flash","headline":"Ataraxos AI Beats Stratego World Champion 15-1 With Four Draws","summary":"A research team from Carnegie Mellon University, MIT, and other institutions built an AI called Ataraxos that beat four-time world champion Pim Nijmeijer at Stratego, winning 15 games, losing 1, and drawing 4. The team says this is superhuman performance in a game of imperfect information. Ataraxos learns through self-play and infers the opponent's hidden pieces during play.","whyItMatters":"Ataraxos adds an inference step that DeepNash lacked, searching the opponent's hidden setup during play rather than only selecting moves from a self-play network.","webCardHtml":"\u003cp\u003eStratego hides piece identities from both players, so the AI must infer what it is facing while it plays. The team\u0026#39;s approach pairs a policy-value network built through self-play with a belief network that reasons about the opponent\u0026#39;s hidden pieces. DeepNash, by contrast, did not search future developments during play. The reported match against Nijmeijer ended with fifteen wins, one loss, and four draws.\u003c/p\u003e","blueskyPost":"Ataraxos learned Stratego through self-play and infers hidden pieces mid-game, the skill that separates imperfect-information play from chess-style search. Pim Nijmeijer is the benchmark that skill was measured against.","twitterPost":"Ataraxos infers hidden pieces during play, not from a database of openings. Stratego has no perfect-information tree to search.","threadsPost":"Ataraxos won 15 games by reading Pim Nijmeijer's hidden pieces as they moved, the part of Stratego that self-play cannot memorize. Chess engines solved a game where both sides see everything; this one solved a game where neither does.","newsletterBlurb":"A research team from Carnegie Mellon, MIT, and other institutions reports that its Ataraxos AI defeated four-time Stratego world champion Pim Nijmeijer with 15 wins, 1 loss, and 4 draws. The team says this is superhuman performance in Stratego. The system combines a self-play policy-value network with a belief network that infers the opponent's hidden pieces during play, an approach DeepNash did not take.","attributionJson":"[{\"source\":\"GIGAZINE\",\"url\":\"https://gigazine.net/news/20261005-stratego-ai-ataraxos/\",\"title\":\"AI 'Ataraxos' wins decisively against world-class Stratego player with 15 wins, 1 loss, 4 draws\"}]","lintFlagsJson":null,"lintHits":0,"costUsd":0,"inputTokens":4696,"outputTokens":565,"status":"published","repairAttempts":0,"nextRepairAt":null,"factsAttemptedAt":1791263241,"createdAt":"2026-10-06T05:05:41.000Z","publishedAt":"2026-10-06T05:06:12.000Z","updatedAt":"2026-10-06T05:06:12.000Z"},"cluster":{"id":"c_a0e952b90abec499bc69b500","canonicalTitle":"AI「Ataraxos」が軍人将棋型ボドゲ「ストラテゴ」で世界最強クラスのプレイヤーに15勝1敗4分で大勝","representativeArticleId":"a_629c0c814d3da3c926d7d7eb","sourceCount":1,"writtenSourceCount":1,"writeAttempts":0,"isSolo":true,"entitiesJson":"{\"anime_titles\":[],\"manga_titles\":[],\"work_titles\":[\"Stratego\"],\"studios\":[],\"people\":[\"Pim Nijmeijer\"],\"type\":\"news\",\"domain\":\"games\",\"is_roundup\":false}","contentType":"news","status":"published","firstSeenAt":"2026-10-05T04:08:00.000Z","lastSeenAt":"2026-10-05T04:08:00.000Z","updatedAt":"2026-10-06T05:06:07.000Z"},"attribution":[{"source":"GIGAZINE","url":"https://gigazine.net/news/20261005-stratego-ai-ataraxos/","title":"AI「Ataraxos」が軍人将棋型ボドゲ「ストラテゴ」で世界最強クラスのプレイヤーに15勝1敗4分で大勝"}],"entities":{"anime_titles":[],"manga_titles":[],"work_titles":["Stratego"],"studios":[],"people":["Pim Nijmeijer"],"type":"news","domain":"games","is_roundup":false},"keyFacts":null}