THE BOOK--Playing The Percentages In Baseball

I can designate ten moderators (via email) and ten readers (once they login and stay logged in) as “always allow posts” (unmoderated posting privileges).  Anyone else can post, but their posts are moderated, and it’s just a matter of one of the ten moderators to bring their posts live.

As of this morning, “phillies world series parade” was the #5 Google Hot Trend, meaning it was experiencing the 5th biggest unexpected growth in searches out of any query.  When your city has booed Santa Claus, you’d think you would worry a little bit about karma, and jinxing things

Looking forward to seeing this live chat tonight.

I’ll only be able to come online long enough to make sure the liveblog is up and running, and then in the late innings.

***

I’d like to try something, which I’ve been meaning to do for years, and this is probably the best way to try it out.  For every BIP, give it a rating as to how easy the play was, from 0 to 100, with the number representing the percentage of times you expect an out to result, from all players, if he positioned himself (at the start of the PA) where the fielder was.

So, on a routine play, you’d say 90% or 95% or 100%.  And so that it’s easier for me to pick out, preface your number with FANS.  And make that the only thing on your line.  As an example, on the Longoria play at the end of game 3, put:
FANS 5
That tells me that you think that that play would have resulted in an out an average of 5% of the time.  Maybe you think Beltre would make that play 20% of the time, and ARod 2%, and Rolen 15%, etc, etc.  On average, you expect to see an out on that play 5 times for every 100 times you see it.

Then, we can see how often routine plays occur, and those can be weeded out, as they are noise in something like UZR.  And, then we can see how well say UZR can pick up plays that are tougher than routine.

Consider this a trial run, using a game that lots of fans are going to be watching.

Funny, I’ve always thought that was the absolute best way to rate fielding (without using speed, trajectory, and fielder position data - and even then, the “human” rating could still be better), bar none, and I always wanted to do that.

I also always wondered why a company like BIS (or STATS) or some team did not do that.  I would suspect that some teams, like Boston or Oakland, may be doing something like that.

What you would want was a lot of knowledgeable people and then you would want to constantly check the integrity of the data, to see if all your 80% plays are actually being fielded 80% of the time, etc. (if that can even be done - I am not sure how you would do it).  You’d also want to use that “checking procedure” to weed out “bad stringers.”

Great idea though.  A little like when we compared the Gameday pitch locations to what we saw on TV.

It would be nice if official scorers actually did something like that in their heads rather than whatever goofy thing they do (or don’t do) when they decide between a hit or an error.  Surely they are capable of coming up with a number - the % chance that all fielders would have made an out.  If it is more than 75% or 80% (I have no idea what the cutoff should be), then score it an error.  It shouldn’t matter whether he touched it, didn’t touch it, took a bad hop, no bad hop, etc. The only thing that matters is how often that same play normally gets fielded and turned into an out.

Agreed, a great way to weed out bad stringers.  Indeed, this is how I knock out the 2% bad voters in the Fans Scouting Report.  It’s really easy to pick out people who just put b.s. responses in the scouting report, and it will be easy to see who has a much different opinion on how often a ball should be an out or not.  If someone is “unique”, given 10 or 15 hardcore fans, that unique person is probably wrong.

I’m getting an error saying: Error Type: load XML // Description: Incorrect XML