Rock Paper Scissors — credits and provenance ============================================ THE GAME -------- Rock Paper Scissors is a folk game in the public domain. It has no author, publisher or copyright holder to credit, and this build claims none. The earliest known written mention is in the Ming-dynasty book "Wuzazu" by Xie Zhaozhe (fl. c. 1600), who records the game under the name shoushiling and dates it to the Han dynasty (206 BCE - 220 CE). Li Rihua's "Note of Liuyanzhai" gives the names shoushiling, huozhitou and huaquan. The game reached Japan and joined the family of sansukumi-ken, "fist games of the three who are afraid of one another". The earliest was mushi-ken, in which frog beats slug, slug beats snake and snake beats frog; kitsune-ken pits fox against village head against hunter and needs both hands. The three-sign version is jan-ken, a seventeenth-century import from China, standardised between the Edo and Meiji periods in the late nineteenth century, and it is the direct ancestor of the modern game. Names in use include Rochambeau, Roshambo, Ro-sham-bo, Bato Bato Pik and Jan-ken-pon; gawi-bawi-bo in Korea; jak-en-poy and bato-bato-pick in the Philippines. The Japanese is written as three kana: - "ja" then "n" then "ke" then "n" (rendered in the app's About tab as jan-ken) The order of the three words in the game's name differs from country to country. A NAMING CLAIM THAT DOES NOT HOLD --------------------------------- "Roshambo" is very widely explained as a tribute to Count Rochambeau, said to have played the game during the American Revolutionary War. This is apocryphal. All evidence places the game's arrival in the United States after 1910, more than a century later. The explanations that survive are a phonetic drift from jan-ken-pon, or the Rochambeau statue in Washington DC. The encyclopaedia that popularised the Rochambeau story says so in the same article. This build reports the disagreement rather than repeating the legend. THE OPPONENT MODEL — the work this app is actually built on ----------------------------------------------------------- Zhijian Wang, Bin Xu and Hai-Jun Zhou, "Social cycling and conditional responses in the Rock-Paper-Scissors game", arXiv:1404.5199v1 [physics.soc-ph], 21 April 2014. Later published as Scientific Reports 4, 5830 (2014). Experimental Social Science Laboratory, Zhejiang University; Zhejiang Gongshang University; Institute of Theoretical Physics, Chinese Academy of Sciences. 360 students, 60 populations of six, 300 rounds each under random pairwise matching, at winner's payoffs a = 1.1, 2, 4, 9 and 100. One population was discarded, leaving 59 populations and 354 players. Taken from that paper and used here: the win / tie / loss conditional-response model and its nine parameters; the social-state formulation and the pairing-pattern distribution; the rotation-angle definition; the payoff identity g = g0 + (a-2)(1/6 - tau/2); the measured action marginals (0.36 / 0.33 / 0.32); the table of 59 per-population cycling frequencies; and the five printed sets of response parameters. A DISAGREEMENT WITH THE PUBLISHED VALUES ----------------------------------------- This build reimplements the conditional-response model independently, as an exact 28-state Markov chain and, separately, as a simulation sharing no code with it. Against the five parameter sets the paper prints: - the payoff g_cr reproduces for all five sets, to the printed precision; - the cycling frequency f_cr reproduces for four of the five; - for set 2 this build obtains f_cr = -0.007 where the paper prints +0.007. The magnitude agrees; the sign does not. The exact chain and the simulation agree with each other, and the paper's own stated rules for which parameters drive which direction also point clockwise for that set, since it has T- = 0.154 against T+ = 0.048 and W0 marginally below L0. This is recorded as an unresolved disagreement, not as a correction. Only the authors can settle it. The most likely account is a sign lost in typesetting. Two further observations of the same kind, offered as notes rather than criticisms: the paper's published action marginals (0.36, 0.33, 0.32) sum to 1.01, and two of the five printed response parameter sets have a triple summing to 0.999. Both are ordinary rounding of three-decimal figures, and both change results measurably if used as printed. This build renormalises and says so. HISTORICAL AND NAMING DETAIL ---------------------------- Wikipedia, "Rock paper scissors", retrieved via the MediaWiki API, 2026. Used for the history, the name variants, the regional versions, the Iocaine Powder result from the First International RoShamBo Programming Competition (1999) and the 2012 University of Tokyo high-speed robot hand. The World Rock Paper Scissors Society is mentioned there as having run tournaments from 2002 to 2009; its rulebook was not reachable from this build, so no "official" rule is claimed anywhere in this app. The rules the app implements are stated plainly in the About tab as the rules the app implements. WHAT IS ORIGINAL TO THIS BUILD ------------------------------ Everything else. This is an independent reimplementation; nothing is ported, decompiled or copied from any program. - All code: engine, bots, opponent models, the exponential-weights ensemble, the simplex linear program and its minimax certificate, the Markov chain, the renderer and the UI. - All artwork: the three hands, the icons and the social card are drawn in code by this build. Nothing is traced, and no image file is fetched. - Every opponent other than the uniform one and the conditional-response one is an original design, not a reproduction of any published program. WHAT DIFFERS FROM THE STUDY --------------------------- - The study ran six humans in a randomly repaired population. This app is one human against one program. - The study's measured response parameters are published only as bar charts, with no numeric table anywhere in the paper or its supplement, so none is reproduced here. The parameter sets offered in the Lab are the ones the paper prints as numbers, which are its own sampled optima rather than its measured human values. - Scores in play are wins minus losses per round. The paper's a-parameterised payoffs appear only in the equilibrium solver. NOT INCLUDED ------------ No five-throw variant is shipped. The well-known lizard-and-Spock expansion is a named third-party work and is deliberately absent.