Re: [baseball-databank] Re: Play by play data (Retrosheet): How would you analyze it?
- On Tue, Sep 29, 2009 at 10:05 AM, Nick Miceli wrote:
>This is definitely moving off-topic for this group.
> I want to get better levels of predictive validity for fielding that are presently possible with available summary measures.
What I'm going to suggest is that you put together a summary of the
information that you would LIKE to have in order to come up with a
"better model", post it on the Retrosheet list, and then those of us
on that list (which does include many people on this list but which
has little to do with the purpose of this list) who have long
experience with using the event files can help you figure out whether
and to what extent you can get that information out of the PBP data.
Specific fielding (and fielder) information tends to be fairly
complete back to 1989, but much less so prior to that - sometimes all
you know is that a batter hit a single or made an out, but you don't
always know who fielded the ball or made the play, let alone what type
of ball it was.
"Confidence isn't knowledge; in fact, confidence can *prevent*
knowledge and innovation from happening, because an unquestioned
belief means you never measure, never test, never look at
-- Esther Schindler