THE BOOK--Playing The Percentages In Baseball

I’d like to see “Times facing opponent” split by relievers and starters.

http://www.baseball-reference.com/leagues/split.cgi?t=p&lg=AL&year=2009#times

Tango:
Not a Sean request, but two thoughts related to your times-facing-opponent work:

1) Have you ever tried analyzing times-thru-order effect by comparing outcomes only to prior outcomes in the same games (same pitchers and hitters)?  That is, when looking at outcomes on 3rd time through the order, compare them to the two prior PAs in these same games (rather than all 1st and 2nd time PAs).  Doing this would remove games in which the pitcher got knocked out early, and might indicate an even stronger times-thru-order effect.  It would be especially interesting for looking at the small 4th-time group, where it appears that pitchers suddenly regain the upper hand (but is probably because these pitchers are especially effective that day). 

2) Has anyone ever looked at this in terms of number of pitches seen in game?  Does seeing more pitches in early PAs increase the amount of “learning” that hitters do?  Seems possible that it would.  And if so, would mean that we have to slightly change how we value hitter patience.

Guy
1. Good suggestion.

2. Pizza did that a year or two ago.  Ideally, we’d need to combine his work (number of pitches thrown) with my work (number of times thru order).  But, we’ll have alot of overlap.  It’s hard to get the leadoff hitter the second time up and have pitches thrown to this point as much different than say 25 to 40.

Tango:

Not sure if we’re thinking of the same thing, but BR does have GDP opps/percentages (and also league averages) for each player on their situational hitting pages, if not on the initial page for hitters. They also have productive outs/opps, baserunning advancement, and other stuff.

I was looking at their splits pages:

http://www.baseball-reference.com/players/split.cgi?n1=raineti01&year=Career&t=b#bases

Actually, the runner on 3B, less than 2 outs IS there.  My apologies.  It’s not labelled well, because it shows --3, when it should show something else (--3 means 3B ONLY).

So, I just want to see the 1xx less than 2 outs right below that.

Tango,

We have splits by SF opportunity

http://www.baseball-reference.com/players/split.cgi?n1=morame01&year=Career&t=b#bases

I see now the title may be confusing, so I’ll change that to ..3, lt 2 out rather than --3, lt 2 out.

http://www.baseball-reference.com/players/split.cgi?n1=pettian01&year=2009&t=p#times

For 4+ times we’ve had that for 2009 season, but haven’t rerun the splits for the old seasons yet.

I’ve added the logic for DP opportunity.

Mitch, just added the logic for times through separated as SP and RP.

Sean, cool, we must have cross-posted on the SF.  And, yeah, I thought you told me you already had the 4+, but then I guess I forgot you had done it for 2009 only.  Ok, we’re good.

I’d like to see ballpark splits by batter handedness.  It’d be nice to see how certain parks affect LH/RH hitters.

Actually, a ballpark page would be great in general. It seems so obvious that I think there has to be some reason why Sean hasn’t done it yet.

You can get complete ballpark stats on the league splits pages pages.  It will tell you how many doubles in Wrigley etc. 

A more complete ballpark page has been on my list for a long time.

Thank you all for the suggestions.  We will have a number of them included when we update in a few weeks.

Sean— Who can I talk to look about getting errors corrected on both the Databank and retrosheet datasets?  I have contacted each Yahoo group with no response.

Jeff,

What are the nature of the errors.  The DataBank runs through me and I try fix things when I can, but my time is very short.  Retrosheet generally logs what they can and does make changes.

For databank, the college players went to are half of what is posted at baseball-reference.  I wouldn’t mind updating it, but I figured you got the data from the same sources.  I did some previous work looking at differences of pitchers that came from college and high school and it is useless because of missing data.

Article showing differences:
http://www.beyondtheboxscore.com/2010/2/22/1321392/colleges-that-graduate-the-most

When collecting this data, I found there was no wind blowing in form right field and I thought I had made a mistake somewhere. I searched on the games database for hat wind direction and the most last time it happened was in 2003.  There are no #7 (Park wind direction) in the last 5 years or so.

Looks like they’ve adopted the Guy version of ERA+ now.