Play-By-Play Data

[rluzinski]

Anyone ever play with PBP data, like what you can get at retrosheet.org? I've downloaded an Access database that parses and loads that PBP data, but I'm not very good with Access.

 

Just curious if anyone else has played around with the stuff.

[ILBrewerFan]

I just read them right into sas. I'm not good with access either. I can probably get you a sas dataset if you have access to sas.

[rluzinski]

Well, I had to google "sas" just to find out what it is... so no, I'm not familiar with it http://forum.brewerfan.net/images/smilies/smile.gif

[ILBrewerFan]

Sorry, given your "stat-savvy'ness" I thought you would be familiar. It's just a statistical software package. Any of the other packages should be able to read in .dB files.

 

SPSS

S+

R

 

R & S+ are basically the same thing and R is free, so you can download it and use it for free. The documentation is good and free as well.

[rluzinski]

I hope I don't come across as pretending to be a stat-authority, because I certainly am not. My only school experience with statistics are the classes I took for my engineering degree. I assure you, no software like SAS was used in those classes.

 

I'm learning as I go, really. I use Excel for all my stats stuff... that's it. A REAL stats guy would probably laugh at that, but I'm not a real stats guy so... http://forum.brewerfan.net/images/smilies/smile.gif

[endaround]

A real stats guy will only laugh at that because Excel has had some serious problems doing calculations. Stuff you've done so far shouldn't be two bad. But I've heard when you start getting into huge data sets Excel's math breaks down.

 

SAS really isn't too bad to work with. Now getting hold of a "legal" copy not supplied by school or work isn't cheap.

[rluzinski]

The only trouble I've had with Excel is the limit to the size of sample it can handle for binomial distributions.

[ILBrewerFan]

There are issues with using excel's regression. Some times estimates can be incorrect, but we're talking out to decimal places which for this purpose is probably irrelevant. This problem does get worse for larger samples (they must be short cutting the solution to the NE's or something).

 

Anyway, excel is fine and you can do things in it that would take you days to code in sas. If your looking to handle bigger datasets I would highly suggest 'R'. Like I said it's free and the language itself is easy to learn, plus there are help files and message boards everywhere.

[1992casey]

Here's a link to R.

[rluzinski]

I'll check it out.. thanks.

[Jimbo]

Sorry, given your "stat-savvy'ness" I thought you would be familiar. It's just a statistical software package. Any of the other packages should be able to read in .dB files.

 

SPSS

S+

R

 

What about Minitab?

[rluzinski]

Well, I've finally unlocked the some of the mysteries of PBP data from retrosheet.org. You have to use the bevent.exe file to preparse the data. It's a comamnd line program, so I recommend writing a batch file if you want to spit out the whole league's worth of data.

 

You can read the resulting file right into Excel. Admittantly, I should probably be using Access (or another database prgram) but owell.

 

Event File Page

 

If are just looking for seasonal data, the Lahman database is great. They even have an excel version loaded up and ready to go:

 

Lahman Database, Version 5.3 - spreadsheet version

 

Sure would be nice to have a partner in crime for this stuff! If anyone needs help, shoot me an email:

 

rluzinski@hotmail.com