{"id":41777,"date":"2026-07-20T15:27:07","date_gmt":"2026-07-20T13:27:07","guid":{"rendered":"https:\/\/www.graviton.at\/letterswaplibrary\/self-promotion-live-multi-sport-clv-dataset-with-de-vigged-fair-probabilities-and-results\/"},"modified":"2026-07-20T15:27:07","modified_gmt":"2026-07-20T13:27:07","slug":"self-promotion-live-multi-sport-clv-dataset-with-de-vigged-fair-probabilities-and-results","status":"publish","type":"post","link":"https:\/\/www.graviton.at\/letterswaplibrary\/self-promotion-live-multi-sport-clv-dataset-with-de-vigged-fair-probabilities-and-results\/","title":{"rendered":"[Self-Promotion] Live Multi-sport CLV Dataset With De-vigged Fair Probabilities And Results"},"content":{"rendered":"<p><!-- SC_OFF --><\/p>\n<div class=\"md\">\n<p>Disclosure up front: this is my project, so I&#8217;m tagging it as self-promotion per the subreddit rules.<\/p>\n<p>A few weeks ago I was trying to build a historical closing line value dataset and realized I couldn&#8217;t find one that combined de-vigged fair probabilities, closing values, and game outcomes across multiple sports. Everything I found was either limited to one sport, only included raw odds, or couldn&#8217;t be redistributed because of licensing.<\/p>\n<p>So I started building my own derived dataset.<\/p>\n<p>Each row contains things like:<\/p>\n<ul>\n<li>Sport<\/li>\n<li>Market<\/li>\n<li>Selection<\/li>\n<li>Event time<\/li>\n<li>Opening fair probability<\/li>\n<li>Closing fair probability<\/li>\n<li>De-vig method<\/li>\n<li>Whether the fair line came from a sharp reference (Pinnacle) or a consensus fallback<\/li>\n<li>Edge at entry<\/li>\n<li>Closing line value (CLV)<\/li>\n<li>Beat close (yes\/no)<\/li>\n<li>Final result<\/li>\n<\/ul>\n<p>A couple notes:<\/p>\n<ul>\n<li>This does <strong>not<\/strong> include raw sportsbook odds or sportsbook names since my data provider doesn&#8217;t allow redistributing that data.<\/li>\n<li>The dataset only contains derived metrics like fair probabilities and CLV.<\/li>\n<li>&#8220;Beat the close&#8221; is meant as a research metric, not proof that a bet was good or profitable.<\/li>\n<\/ul>\n<p>One thing I think is useful is the <strong>anchor<\/strong> field. You can separate observations that were generated from a sharp market reference from ones that used a consensus fallback instead of treating them as the same thing.<\/p>\n<p>I&#8217;m planning to keep expanding this over time as more sports and markets are added.<\/p>\n<p>Project: <a href=\"https:\/\/edgedesksports.com\/\">https:\/\/edgedesksports.com<\/a><\/p>\n<p>I&#8217;d really appreciate feedback from people who work with betting or forecasting datasets.<\/p>\n<ul>\n<li>Are there any columns you&#8217;d want added?<\/li>\n<li>Is there another derived metric that would make this more useful for research?<\/li>\n<li>Has anyone found a comparable open dataset that covers multiple sports?<\/li>\n<\/ul><\/div>\n<p><!-- SC_ON -->   submitted by   <a href=\"https:\/\/www.reddit.com\/user\/Clean_Reference_9927\"> \/u\/Clean_Reference_9927 <\/a> <br \/> <span><a href=\"https:\/\/www.reddit.com\/r\/datasets\/comments\/1v1k4a6\/selfpromotion_live_multisport_clv_dataset_with\/\">[link]<\/a><\/span>   <span><a href=\"https:\/\/www.reddit.com\/r\/datasets\/comments\/1v1k4a6\/selfpromotion_live_multisport_clv_dataset_with\/\">[comments]<\/a><\/span><\/p><div class='watch-action'><div class='watch-position align-right'><div class='action-like'><a class='lbg-style1 like-41777 jlk' href='javascript:void(0)' data-task='like' data-post_id='41777' data-nonce='ee32349510' rel='nofollow'><img class='wti-pixel' src='https:\/\/www.graviton.at\/letterswaplibrary\/wp-content\/plugins\/wti-like-post\/images\/pixel.gif' title='Like' \/><span class='lc-41777 lc'>0<\/span><\/a><\/div><\/div> <div class='status-41777 status align-right'><\/div><\/div><div class='wti-clear'><\/div>","protected":false},"excerpt":{"rendered":"<p>Disclosure up front: this is my project, so I&#8217;m tagging it as self-promotion per the subreddit rules&#8230;.<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[85],"tags":[],"class_list":["post-41777","post","type-post","status-publish","format-standard","hentry","category-datatards","wpcat-85-id"],"_links":{"self":[{"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/posts\/41777","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/comments?post=41777"}],"version-history":[{"count":0,"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/posts\/41777\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/media?parent=41777"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/categories?post=41777"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.graviton.at\/letterswaplibrary\/wp-json\/wp\/v2\/tags?post=41777"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}