{"116551":{"#nid":"116551","#data":{"type":"news","title":"Georgia Tech Professors Make 2012 Final Four Predictions","body":[{"value":"\u003Cp\u003EWant to win the NCAA basketball office pool this year?\u003C\/p\u003E\n\n\u003Cp\u003EThen consider the predictions made by Georgia Tech\u2019s\nLogistic Regression Markov Chain (LRMC) method, a computer ranking system that\nhas historically been more accurate than the NCAA\u2019s own Ratings Percentage\nIndex.\u003C\/p\u003E\n\n\u003Cp\u003ELRMC predicts this year\u2019s NCAA Final Four matchups will most\nlikely be Kentucky vs. Michigan St. and Ohio St. vs. Kansas, with Kentucky\nbeating Ohio St. for the championship.\u003C\/p\u003E\n\n\u003Cp\u003EOther predictions by the system include:\u003C\/p\u003E\n\n\u003Cul\u003E\u003Cli\u003ETexas, Belmont and N.C. State are the underdogs\nmost likely to pull off an upset in the first round. \u003C\/li\u003E\u003Cli\u003ECalifornia, N.C. State, Belmont and Texas could\nbe this year\u2019s \u201cCinderella\u201d teams; they are the most likely double-digit seeds\nto make it to the Sweet 16. \u0026nbsp;\u003C\/li\u003E\u003Cli\u003EMichigan St. will be the No. 1 seed with the\ntoughest second-round matchup.\u003C\/li\u003E\u003Cli\u003EWichita St. vs. Indiana and New Mexico vs. Louisville\nare other intriguing potential second-round matchups.\u003C\/li\u003E\u003Cli\u003EThe West Region, led by Michigan St., is the\ndeepest of the four regions.\u003C\/li\u003E\u003Cli\u003EWichita St. vs. Indiana and New Mexico vs.\nLouisville are \u0022intriguing\u0022 potential second-round matchups because\nLRMC says they will be close games, even though both pit a lesser-known team\nagainst a better-known team.\u003C\/li\u003E\u003C\/ul\u003E\n\n\n\n\n\n\n\n\n\n\n\n\u003Cp\u003E\u201cKentucky is the likely champion because they\u0027ve won almost\nall their games,\u201d said Joel Sokol, operations research professor at Georgia\nTech who developed LRMC along with colleagues. \u201cThey\u0027ve won by convincing\nmargins at home and on the road against very good teams, and they\u0027ve done it\nall against a strong schedule, including Kansas, North Carolina, Indiana and\nFlorida.\u201d\u003C\/p\u003E\n\n\u003Cp\u003ESince the 2003 season, LRMC has correctly predicted the\noutcomes of more NCAA tournament games than competing ranking systems and major\npolls.\u003C\/p\u003E\n\n\u003Cp\u003EIn 2010, for example, LRMC correctly predicted the winners\nof 51 out of 64 NCAA games\u2014beating out more than 50 of the top-ranking sites.\nIn 2008, the system predicted the Final Four, final two and the eventual\nvictor, as well as several upsets in earlier rounds.\u003C\/p\u003E\n\n\u003Cp\u003EGeorgia Tech Operations Research Professors Sokol and George\nNemhauser and Statistics Professor Paul Kvam developed the LRMC method, along with\nMath Professor Mark Brown of the City College of New York.\u0026nbsp;\u003C\/p\u003E\n\n\u003Cp\u003EThe system looks at the results of all the college basketball\ngames played during the season. Specifically, it examines which team wins, which\nteam loses, where the game was played and the team\u2019s margin of victory. The\nresearchers then run that data through several mathematical models\u2014empirical Bayes,\nlogistic regression and Markov Chain\u2014to determine the ranking of teams.\u003C\/p\u003E\n\n\u003Cp\u003EYet even with the best formula, it\u2019s impossible to predict a\nperfect bracket, Sokol said.\u003C\/p\u003E\n\n\u003Cp\u003EAbout one-quarter of all tournament games are affected by\nupsets, injuries or last-second, buzzer-beating baskets. Such was the case last\nyear when only one top seed made it to the regional finals. This human factor\nis where the LRMC predictions can falter.\u003C\/p\u003E\n\n\u003Cp\u003EStill, LRMC\u2019s odds aren\u2019t bad.\u003C\/p\u003E\n\n\u003Cp\u003EAccording to a study of historical data just completed by\nthe research team, LRMC is significantly better at predicting NCAA Tournament\ngames than almost all of the other ranking systems, such as Sagarin\u2019s\npredictor, Pomeroy\u2019s ranking, Las Vegas Favorite and the NCAA\u2019s RPI.\u003C\/p\u003E\n\n\u003Cp\u003ESokol, Nemhauser and Kvam are professors in the H. Milton\nStewart School of Industrial \u0026amp; Systems Engineering in Georgia Tech\u2019s\nCollege of Engineering.\u003C\/p\u003E\n\n\u003Cp\u003E\u0026nbsp;\u003C\/p\u003E","summary":null,"format":"limited_html"}],"field_subtitle":[{"value":"LRMC picks Kentucky beating Ohio St. for the championship."}],"field_summary":[{"value":"\u003Cp\u003EGeorgia Tech\u2019s Logistic Regression Markov Chain (LRMC)\nmethod is a computer ranking system that has historically been more accurate than\nthe NCAA\u2019s own Ratings Percentage Index.\u0026nbsp;LRMC predicts this year\u2019s NCAA Final Four matchups will most\nlikely be Kentucky vs. Michigan St. and Ohio St. vs. Kansas, with Kentucky\nbeating Ohio St. for the championship.\u003C\/p\u003E","format":"limited_html"}],"field_summary_sentence":[{"value":"Georgia Tech\u2019s Logistic Regression Markov Chain (LRMC) method is a computer ranking system that has historically been more accurate than the NCAA\u2019s own Ratings Percentage Index."}],"uid":"27462","created_gmt":"2012-03-13 11:38:33","changed_gmt":"2016-10-08 03:11:52","author":"Liz Klipp","boilerplate_text":"","field_publication":"","field_article_url":"","dateline":{"date":"2012-03-13T00:00:00-04:00","iso_date":"2012-03-13T00:00:00-04:00","tz":"America\/New_York"},"extras":[],"hg_media":{"107631":{"id":"107631","type":"image","title":"Joel Sokol Makes NCAA Tournament Predictions","body":null,"created":"1449178188","gmt_created":"2015-12-03 21:29:48","changed":"1475894725","gmt_changed":"2016-10-08 02:45:25","alt":"Joel Sokol Makes NCAA Tournament Predictions","file":{"fid":"194004","name":"12c3046-p1-003.jpg","image_path":"\/sites\/default\/files\/images\/12c3046-p1-003_0.jpg","image_full_path":"http:\/\/hg.gatech.edu\/\/sites\/default\/files\/images\/12c3046-p1-003_0.jpg","mime":"image\/jpeg","size":2221546,"path_740":"http:\/\/hg.gatech.edu\/sites\/default\/files\/styles\/740xx_scale\/public\/images\/12c3046-p1-003_0.jpg?itok=ZzDr6AD4"}}},"media_ids":["107631"],"related_links":[{"url":"http:\/\/www.lrmc.gatech.edu\/","title":"Background on LRMC system"},{"url":"http:\/\/www2.isye.gatech.edu\/~jsokol\/profspicks\/profspicks12.htm","title":"LRMC 2012 \u0022Profs Picks\u0022 bracket"}],"groups":[{"id":"1214","name":"News Room"}],"categories":[{"id":"135","name":"Research"}],"keywords":[{"id":"27021","name":"College of Engineering; H. Milton Stewart School of Industrial \u0026 Systems Engineering; LRMC; Joel Sokol; NCAA Men\u0027s Basketball Tournament"}],"core_research_areas":[],"news_room_topics":[],"event_categories":[],"invited_audience":[],"affiliations":[],"classification":[],"areas_of_expertise":[],"news_and_recent_appearances":[],"phone":[],"contact":[{"value":"\u003Cp\u003E\u003Cstrong\u003EGeorgia Tech Media Relations\u003C\/strong\u003E\u003Cbr \/\u003ELaura Diamond\u003Cbr \/\u003E\u003Ca href=\u0022mailto:laura.diamond@comm.gatech.edu\u0022\u003Elaura.diamond@comm.gatech.edu\u003C\/a\u003E\u003Cbr \/\u003E404-894-6016\u003Cbr \/\u003EJason Maderer\u003Cbr \/\u003E\u003Ca href=\u0022mailto:maderer@gatech.edu\u0022\u003Emaderer@gatech.edu\u003C\/a\u003E\u003Cbr \/\u003E404-660-2926\u003C\/p\u003E","format":"limited_html"}],"email":["klipp@gatech.edu"],"slides":[],"orientation":[],"userdata":""}}}