tag:blogger.com,1999:blog-31600912.post116380945282565300..comments2023-05-06T08:59:13.282-07:00Comments on Hockey Numbers: Predicting what games go to OvertimeJavaGeekhttp://www.blogger.com/profile/02529258432033746879noreply@blogger.comBlogger3125tag:blogger.com,1999:blog-31600912.post-1163882177651355562006-11-18T12:36:00.000-08:002006-11-18T12:36:00.000-08:00I create a part II for all those interested.I create a <A HREF="http://hockeynumbers.blogspot.com/2006/11/overtime-going-into-extra-innings.html" REL="nofollow">part II</A> for all those interested.JavaGeekhttps://www.blogger.com/profile/02529258432033746879noreply@blogger.comtag:blogger.com,1999:blog-31600912.post-1163864890466067512006-11-18T07:48:00.000-08:002006-11-18T07:48:00.000-08:00Binary Logistic Regression:Concordant: 61.6Discord...Binary Logistic Regression:<BR/>Concordant: 61.6<BR/>Discordant: 37.6<BR/><BR/>This is a <I>prediction</I> algorithm, so it's using past data to predict future games.JavaGeekhttps://www.blogger.com/profile/02529258432033746879noreply@blogger.comtag:blogger.com,1999:blog-31600912.post-1163834950989748112006-11-17T23:29:00.000-08:002006-11-17T23:29:00.000-08:00A great example happened today:Anaheim lost in a s...A great example happened today:<BR/>Anaheim lost in a shootout to Chicago.<BR/><BR/>Obviously over 15 years you might find something (but the new OT rules were started only recently).<BR/><BR/>It shouldn't a be a shock that a 5 minute OT or a shootout are essentially random. You can show this with the Poisson data or some simple probabilities for the shoot out.<BR/><BR/>Now how bad teams get to JavaGeekhttps://www.blogger.com/profile/02529258432033746879noreply@blogger.com