The Bestseller Code: Anatomy of the Blockbuster Novel

　　暢銷書的背後，總存在某種引人上鉤的共通魅力。
　　我們恨不得翻遍上萬本小說找出答案，但人類的能力終究有限。
　　本書兩位作者來自史丹佛大學的文學實驗室，
　　他們歷時五年，運用文字探勘技術，
　　訓練電腦「閱讀」兩萬部小說，
　　全面破解《紐約時報》暢銷書在題材、情節、角色和風格上的共同模式，
　　告訴你人類如何閱讀，以及小說引人入勝的終極關鍵。

　　原來──

　　◎《格雷的五十道陰影》和《達文西密碼》竟然擁有完全相同的說故事節奏！
　　◎ 動詞運用洩露出《龍紋身的女孩》和《列車上的女孩》主角擁有共同特質！
　　◎ 暢銷作家都有兩至三個拿手題材，而且占據了每部作品三分之一的篇幅！
　　◎ 標點符號具有預測能力，暢銷書出現問號的比例特別高！
　　◎ 電腦判定的暢銷作家以女性居多，而史蒂芬．金被認定為女性！

　　不僅如此，電腦還可以進一步運用研究結果預測暢銷書，
　　只憑書稿，不論作者名氣或行銷運算，
　　預測準確度就已超過八成──

　　◎ 麥可．康納利的《林肯律師》暢銷書指數99.2%
　　◎ 丹．布朗的《地獄》暢銷書指數95.7%
　　◎ 安迪．威爾的《火星任務》暢銷書指數 93.4%
　　◎ 希維雅．黛的《謎情柯洛斯》暢銷書指數 91.2%

　　史無前例，《暢銷書密碼》是第一本「電腦觀點」的小說分析，
　　運用新科技帶我們重新理解好故事如何打動人心。
　　無論你熱愛小說、寫作或資料科學，
　　都可以從本書得到¬前所未有的樂趣和啟發。

各界推薦

　　顏擇雅 出版人
　　陳昇瑋 中央研究院資訊科學所研究員
　　好評推薦

　　為出版界帶來重大革命的一本書！──《衛報》

　　一個故事能吸引數百萬人閱讀，那是魔法。魔法能被描述嗎？這本書做到了！──《紐約時報》暢銷作家 希維雅．黛

　　實在出色！亞契和賈克斯用前所未見的視角洞悉暢銷小說的DNA，並轉換成引人入勝的文字，告訴我們人類究竟如何閱讀。──《大數據》作者 麥爾．荀伯格

　　《暢銷書密碼》是用新科技說故事的先驅。
　　兩位作者創造了一個閱讀機器人，教導讀者、作家、出版人所有關於暢銷小說的運作。──《故事如何改變你的大腦》作者 哥德夏

　　想寫小說的人都恨不得把這本書翻爛，徹底研究一番。──《公共讀書雜誌》

　　---取自中譯版《暢銷書密碼：人工智慧帶我們重新理解小說創作》雲夢千里出版

"When a story captures the imagination of millions, that's magic. Can you qualify magic? Archer and Jockers just may have done so."―Sylvia Day, New York Times bestselling author

　　Ask most book people about massive success in the world of fiction, and you’ll typically hear that it’s a game of hazy crystal balls. The sales figures of E. L. James or Dan Brown, they’ll say, are freakish―random occurrences in an unpredictable market. But what if there were an algorithm that could predict mega-bestsellers with stunning accuracy? What if it knew, just from reading an unpublished manuscript, not just that genre writers like John Grisham and Danielle Steel would sell in huge numbers, but also that authors such as Junot Diaz, Jodi Picoult, and Donna Tartt had signs ofNew York Times bestselling all over their pages?

　　Thanks to Jodie Archer and Matthew Jockers, the algorithm exists, the code has been cracked, and the results are stunning. Fine-tuned on over 20,000 contemporary novels, the system analyzes themes, plot, character, setting, and also the frequencies of tiny but amazingly significant markers of style. The “bestseller-ometer” then makes predictions, with fascinating detail, about which specific combinations of these features will resonate with readers. Somehow, in all genres, it is right over eighty percent of the time.

　　This book explains groundbreaking text mining research in accessible terms, but its real story is in what the algorithm reveals about reading and writing and how successful authorship works. It offers a new theory on the success of Fifty Shades of Grey. It explains why Gone Girl sold millions of copies. It reveals the most important theme in bestselling fiction and which topics just won’t sell. And then there’s “The One,” the single most paradigmatic bestseller of the past thirty years that a computer picked from among thousands. The result is surprising, a bit ironic, and delightfully unorthodox.

　　The project will be compelling and provocative for all book lovers and writers. It is an investigation into our intellectual and emotional responses to stories, as well as a big idea book about the relationship between creativity and technology. It turns conventional wisdom about book publishing on its head.The Bestseller Code will appeal to fiction lovers, data nerds, and those people who have enjoyed books by Malcolm Gladwell and Nassim Taleb.