Sounds like you're more into probability modelling and machine learning than statistics in the traditional hypothesis testing sense. Besides ESL, a book I'd recommend is Bishop's Pattern Recognition and Machine Learning. It starts from the beginning of probability theory applied to computer science problems, and covers every modern topic.
videolectures.net is filled with lectures on CS-flavored probability modelling and machine learning topics. The best bet is the multi-hour "tutorial" lecture series and minicourses; it may take a while to choose the right starting point.
For serious stats and probability without the CS flavoring (not useful for the quick-road-to-hacking-power agenda):
For classical deep stats theory, everyone I know begins with Cassela and Berger's Statistical Inference. Don't expect algorithms in this though.
On the probability side: Feller's An Introduction to Probability Theory and Its Applications. Deep, readable, sometimes funny, full of "whoa" insights. Would be hard to actually grok every chapter in both volumes, but you read it for insight into the power of probability and then use it as a reference.
videolectures.net is filled with lectures on CS-flavored probability modelling and machine learning topics. The best bet is the multi-hour "tutorial" lecture series and minicourses; it may take a while to choose the right starting point.
For serious stats and probability without the CS flavoring (not useful for the quick-road-to-hacking-power agenda):
For classical deep stats theory, everyone I know begins with Cassela and Berger's Statistical Inference. Don't expect algorithms in this though.
On the probability side: Feller's An Introduction to Probability Theory and Its Applications. Deep, readable, sometimes funny, full of "whoa" insights. Would be hard to actually grok every chapter in both volumes, but you read it for insight into the power of probability and then use it as a reference.
(grad student in stats, among other things)