[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-82249-en":3,"doc-seo-82249-105":29,"detail-sidebar-cat-0-en-105":90},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":4,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":13,"seo_description":14,"update_tm":27,"read_time":28},82249,962075114765,"Quinn","https://ap-avatar.wpscdn.com/davatar_a8503ba1806abce46bf441b54a3ca4cd",8,"Research & Report","Application of Machine Learning to Monster Level Prediction in Tabletop RPG Game Design","Designing balanced adversaries is a key but costly task in tabletop role-playing game development. Each monster in systems like Pathfinder is characterized by many numerical attributes whose combined strength is summarized as an ordinal level. The study evaluates tabular ordinal regression approaches, comparing classical regressors with rounding schemes, dedicated ordinal models, and ordinal-aware neural networks. Models are assessed via chronological and expanding-window protocols with multiple metrics, and tree-based ensembles deliver near-perfect ordinal ranking.","arXiv :2607 .09 196v 1 [ cs .LG] 10 Jul 2026  \nApplication of machine learning to monster level prediction in  \ntabletop RPG game design  \nJolanta Śliwaa,1 , Jakub Adamczyka,1,∗  \na Faculty of Computer Science, AGH University of Krakow, Cracow, Poland  \nAbstract  \nDesigning balanced adversaries is a central but labor-intensive task in tabletop role-playing game (TTRPG) development. In systems such as Pathfinder, each monster is described by many numerical attributes that jointly determine its power, summarized as an ordinal level. We investigate whether machine learning can support designers by predicting this level from a monster’s attributes, framing the task as tabular ordinal regression. We introduce what is, to our knowledge, the first dataset built specifically for TTRPG monster-level prediction, derived from publicly available Pathfinder Second Edition data. Using it, we compare classical regression models with rounding schemes, dedicated tabular ordinal regression algorithms, and neural networks with ordinal-aware losses. To mirror real design workflows, we evaluate all models under chronological and expanding-window protocols with several complementary metrics. Results show that tree-based ensembles outperform linear models and neural approaches, achieving near-perfect ordinal ranking and high predictive accuracy. Explainable AI analyses, such as feature importance and error distributions, show that the model is aligned with human intuition and follows patterns grounded in game rules. Together, these results show that machine learning can reliably approximate designer judgments and serve asan effective computer-aided tool for monster balancing and broader TTRPG system design.  \nKeywords: machine learning, ordinal regression, computer-aided design, game design  \n2000 MSC: 62-04, 62J02, 62P25, 68-04, 68T05, 68T35, 68U07  \n1. Introduction  \nArtificial intelligence (AI) and machine learning (ML) methods have been used to play games since the inception of these fields, with breakthroughs such as alpha-beta pruning [1], Monte Carlo Tree Search (MCTS) [2], AlphaGo [3], and AlphaZero [4] . These applications are player-centered and assume an existing game environment together with its rules. In other words, they presuppose that the game itself has already been created and designed, which allows an abstract mathematical representation of its rules and states. In this work,  \n∗ Corresponding author  \nEmail addresses: [jolantasliwa@agh.edu.pl](jolantasliwa@agh.edu.pl) (Jolanta Śliwa), [jadamczy@agh.edu.pl](jadamczy@agh.edu.pl) (Jakub  \nAdamczyk)  \n1 Both authors contributed equally to this work.  \n2 ORCID 0009-0004-1889-5655  \n3 ORCID 0000-0003-4336-4288  \nwe instead focus on using machine learning for game design, an area that has attracted comparatively less interest yet is of high importance to the industry, as many steps of game creation are slow, labor-intensive, and expensive.  \nSpecifically, we study role-playing games (RPGs), and in particular pen & paper RPGs, also known as tabletop RPGs (TTRPGs) . These are distinct from, and predate, computer games and computer RPGs (cRPGs) . In this type of game, participants take on the roles of fictional characters within a story narrated by a Game Master (GM) . Players gather around a table, physically or virtually, to create and develop characters within an evolving story. Gameplay is driven by narration and by the resolution of conflicts through a chosen rule system (or simply “system”), typically using multi-sided dice rolls to introduce randomness. The game design task considered here centers purely on the rules of the game itself, rather than on generating story, levels, or music, which are left to the GM and human creativity. TTRPGs are also frequently adopted as rulesets for computer games, with famous examples such as the Baldur’s Gate series using Dungeons & Dragons (D&D) rules, and Pathfinder: Wrath of the Righteous using the Pathfinder First Edition system.  \nThe RPG ma","cbCaicDCAa0Q08Yz","https://ap.wps.com/l/cbCaicDCAa0Q08Yz","pdf",1329833,1,34,"English","en",105,"# Introduction\n## Game-centered AI vs game design\n## Tabletop RPG and rule-driven design task\n## Monster level estimation problem","[{\"question\":\"What is the main problem addressed in the paper?\",\"answer\":\"Predicting a monster’s ordinal level from its numerical attributes to support tabletop RPG game design and balancing.\"},{\"question\":\"How is the prediction task formulated?\",\"answer\":\"As tabular ordinal regression, where the target is an ordered (ordinal) monster level derived from multiple attributes.\"},{\"question\":\"Which model family performs best according to the results?\",\"answer\":\"Tree-based ensemble methods outperform linear models and neural approaches, achieving near-perfect ordinal ranking and high predictive accuracy.\"}]",1784179151,86,{"code":4,"msg":30,"data":31},"ok",{"site_id":24,"language":23,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":85,"head_meta":87,"extra_data":89,"updated_unix":27},"application-of-machine-learning-to-monster-level-prediction-in-tabletop-rpg-game-design","",{"@graph":35,"@context":84},[36,53,67],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,50],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/application-of-machine-learning-to-monster-level-prediction-in-tabletop-rpg-game-design/82249/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":61,"encodingFormat":60,"isAccessibleForFree":62,"interactionStatistic":63},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-16",true,{"@type":64,"interactionType":65,"userInteractionCount":4},"InteractionCounter",{"@type":66},"ViewAction",{"@type":68,"mainEntity":69},"FAQPage",[70,76,80],{"name":71,"@type":72,"acceptedAnswer":73},"What is the main problem addressed in the paper?","Question",{"text":74,"@type":75},"Predicting a monster’s ordinal level from its numerical attributes to support tabletop RPG game design and balancing.","Answer",{"name":77,"@type":72,"acceptedAnswer":78},"How is the prediction task formulated?",{"text":79,"@type":75},"As tabular ordinal regression, where the target is an ordered (ordinal) monster level derived from multiple attributes.",{"name":81,"@type":72,"acceptedAnswer":82},"Which model family performs best according to the results?",{"text":83,"@type":75},"Tree-based ensemble methods outperform linear models and neural approaches, achieving near-perfect ordinal ranking and high predictive accuracy.","https://schema.org",{"og:url":51,"og:type":86,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":88,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":91},[92,96,100,104,109,114,119,122,127,130,134],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":93,"show_sort_weight":94,"slug":95},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":97,"show_sort_weight":98,"slug":99},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":101,"show_sort_weight":102,"slug":103},"Exam",70,"exam",{"id":105,"doc_module":4,"doc_module_name":45,"category_name":106,"show_sort_weight":107,"slug":108},5,"Comic",60,"comic",{"id":110,"doc_module":4,"doc_module_name":45,"category_name":111,"show_sort_weight":112,"slug":113},6,"Technology",50,"technology",{"id":115,"doc_module":4,"doc_module_name":45,"category_name":116,"show_sort_weight":117,"slug":118},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":120,"slug":121},30,"research-report",{"id":123,"doc_module":4,"doc_module_name":45,"category_name":124,"show_sort_weight":125,"slug":126},9,"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":45,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":45,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":45,"category_name":136,"show_sort_weight":105,"slug":137},19,"General","general"]