One thing that bugs me is that there seems to be so little model checking done in statistics. Data-based model checking is a powerful tool for overcoming bias, and it’s frustrating to see this tool used so rarely. As I wrote
g, I think my post reflects awareness of that possibility. My request to Andrew in my September 22, 2007 at 01:27 PM post stands. Readers (Andrew included) can make their own judgments as to the merit of my request.
HA, you might be seeking status without consciously seeking status. I think Eliezer's question meant "are you sure you aren't fooling yourself?" rather than "are you sure you aren't lying to us?".
Eliezer,haha.My internal sense is that I'm seeking persistence, and that rationally status-seeking behavior on my part should be completely subordinate to that goal.
Sounds like a putdown to me, HA. Sure you're not seeking status?
Andrew, the sense I know of model checking is the theorem-proving one, but that's obviously not what you mean here. Googling "data-based model checking" didn't turn up much. The impression I get is that you want the authors to write a simulation which uses the model fitted to the data to output new data, and then graph the new data and the real data side-by-side. Is this correct?
Great post, Andrew. I'd like to see more of this from you, and less moralizing or positive self-positioning (particularly of the variety that's not carefully derived from empiricism).
I don't think your post in the comments such as "I don't hold meetings to signal my status" add value the way your OP here does.
g, I think my post reflects awareness of that possibility. My request to Andrew in my September 22, 2007 at 01:27 PM post stands. Readers (Andrew included) can make their own judgments as to the merit of my request.
HA, you might be seeking status without consciously seeking status. I think Eliezer's question meant "are you sure you aren't fooling yourself?" rather than "are you sure you aren't lying to us?".
Eliezer,haha.My internal sense is that I'm seeking persistence, and that rationally status-seeking behavior on my part should be completely subordinate to that goal.
Eliezer,
Yes, that's what I'm talking about.
Sounds like a putdown to me, HA. Sure you're not seeking status?
Andrew, the sense I know of model checking is the theorem-proving one, but that's obviously not what you mean here. Googling "data-based model checking" didn't turn up much. The impression I get is that you want the authors to write a simulation which uses the model fitted to the data to output new data, and then graph the new data and the real data side-by-side. Is this correct?
Great post, Andrew. I'd like to see more of this from you, and less moralizing or positive self-positioning (particularly of the variety that's not carefully derived from empiricism).
I don't think your post in the comments such as "I don't hold meetings to signal my status" add value the way your OP here does.