tag:blogger.com,1999:blog-7608390901106026172.post342672040698962260..comments2023-04-07T03:37:03.408-07:00Comments on The Statisticator: Morey's Law: How do points scored and points allowed tie to win percentage?Statisticatorhttp://www.blogger.com/profile/05704610338790666912noreply@blogger.comBlogger3125tag:blogger.com,1999:blog-7608390901106026172.post-53839344213813314042016-01-25T20:46:13.893-08:002016-01-25T20:46:13.893-08:00Hi Kumaran, thanks for your comment. Indeed, a cor...Hi Kumaran, thanks for your comment. Indeed, a correlation is definitely expected, but correlations as high as 0.97 rarely attained with actual data. Personally, I find it all the more surprising with basketball data: in many of the analyses I did, variables rarely performed as expected!<br />Also, while you would expect a positive correlation (if you outscore more you win more), but the relationship could be more complex than a straight line. A sigmoid might actually be more realistic (outscoring by an average of 20 shouldn't guarantee many more wins than outscoring by an average of 15). A correlation of 0.97 somewhat contradicts the sigmoid in favor of a straight line.<br />Thanks again for your comment, please share a link to your work, would be interested to read it!Statisticatorhttps://www.blogger.com/profile/05704610338790666912noreply@blogger.comtag:blogger.com,1999:blog-7608390901106026172.post-64353104096190137522016-01-13T23:57:53.233-08:002016-01-13T23:57:53.233-08:00hello, i'm a undergrad in statistics who is do...hello, i'm a undergrad in statistics who is doing a project on basketball and working on pythagorean expectation as well. I would like to know why a strong correlation (0.97) between winning % and scoring differential is seen as a surprise? they are suppose to be strongly correlated given that a larger scoring differential usually does mean you you outscored more opponents and thus won more games. Thanks for reading my comment and hope to hear and learn from you. Great blog post. Thanks.<br />Anonymoushttps://www.blogger.com/profile/08110252960553879308noreply@blogger.comtag:blogger.com,1999:blog-7608390901106026172.post-11462163310197046292016-01-13T23:57:37.389-08:002016-01-13T23:57:37.389-08:00hello, i'm a undergrad in statistics who is do...hello, i'm a undergrad in statistics who is doing a project on basketball and working on pythagorean expectation as well. I would like to know why a strong correlation (0.97) between winning % and scoring differential is seen as a surprise? they are suppose to be strongly correlated given that a larger scoring differential usually does mean you you outscored more opponents and thus won more games. Thanks for reading my comment and hope to hear and learn from you. Great blog post. Thanks.<br />Anonymoushttps://www.blogger.com/profile/08110252960553879308noreply@blogger.com