Changes

Thank you for the correction in my calculation of percentages in [[Talk:Main_Page#Did_George_Soros_convince_someone_here_that_prediction_markets_actually_matter.3F|this section]]. I would prefer someone politely tell me when I've made an error, as I try to do when I notice errors in other people's work. That being said, I hope you understand that my point still stands. If a hypothesis predicts one event with a certain probability of occurrence and a complementary event with a different, ''smaller'' probability of occurrence, the occurrence of the event bearing a smaller likelihood absolutely does ''not'' invalidate that hypothesis. This is why statisticians use hypothesis testing: to test the conditions under which such an invalidation would occur. Simply put, you can calculate the likelihood of a specific event happening, ''given your hypothesis as to its probability'' -which is not the same as the originally predicted probability- and analyze the situation from there. If you search basic statistical methodology for hypothesis testing, you will find this. Does that make sense? Let me know if you need anything else clarified; thank you! [[User:KevinDavis|Kevin Davis]] ^{[[User talk:KevinDavis|Talk]]} 20:33, 25 September 2011 (EDT)

Thank you for the correction in my calculation of percentages in [[Talk:Main_Page#Did_George_Soros_convince_someone_here_that_prediction_markets_actually_matter.3F|this section]]. I would prefer someone politely tell me when I've made an error, as I try to do when I notice errors in other people's work. That being said, I hope you understand that my point still stands. If a hypothesis predicts one event with a certain probability of occurrence and a complementary event with a different, ''smaller'' probability of occurrence, the occurrence of the event bearing a smaller likelihood absolutely does ''not'' invalidate that hypothesis. This is why statisticians use hypothesis testing: to test the conditions under which such an invalidation would occur. Simply put, you can calculate the likelihood of a specific event happening, ''given your hypothesis as to its probability'' -which is not the same as the originally predicted probability- and analyze the situation from there. If you search basic statistical methodology for hypothesis testing, you will find this. Does that make sense? Let me know if you need anything else clarified; thank you! [[User:KevinDavis|Kevin Davis]] ^{[[User talk:KevinDavis|Talk]]} 20:33, 25 September 2011 (EDT)