Critical info for Baseball Model

[Regul8er]

Before I develop a model, I am seeking some advice as to what information is most critical to own?
Past lines and results obviously, but how in depth in terms of box score statistics?

Any answers are welcomed and appreciated. Thanks!

[suicidekings]

It really depends on what your focus is. Are you looking to research for situational betting angles? If you're more interested in handicapping individual games for sides/totals then quality projected player statistics are essential (either produced yourself or taken from BP/Fanstistics/etc).

If you're looking into past seasons data to run linear regressions on trends, etc, then having is much data as possible is best. http://www.retrosheet.org/ has very detailed records available from games that will give you everything you need, minus the betting lines.

[mebaran]

I agree with suicidekings. I just set up a database with play-by-play data from retrosheet, and there is a crazy amount of data on each game.

I just don't know how to go about using it yet. I'm using MySQL with MySQL Workbench to access and change it, but I'm having trouble because I don't know how to use MySQL that well yet. Is there another kind of database I can build with all my parsed data? Another program to use that would be more user-friendly?

[boondoggle]

linux and open sauce apps are not meant to be friendly...but powerful. Mysql 5 or sql server 2008 or oracle all require intimate knowledge of SQL.

[LLXC]

Play by Play.

[the_mathman]

It really depends on what your focus is. Are you looking to research for situational betting angles? If you're more interested in handicapping individual games for sides/totals then quality projected player statistics are essential (either produced yourself or taken from BP/Fanstistics/etc).

If you're looking into past seasons data to run linear regressions on trends, etc, then having is much data as possible is best. http://www.retrosheet.org/ has very detailed records available from games that will give you everything you need, minus the betting lines.

wow! from 1950!
greath site!
thank you for this link!

[d_mad]

Has anyone setup a useable header row in Excel for retrosheets?

Data is useless if you can not translate it.

[caracalla]

It really depends on what your focus is. Are you looking to research for situational betting angles? If you're more interested in handicapping individual games for sides/totals then quality projected player statistics are essential (either produced yourself or taken from BP/Fanstistics/etc).

If you're looking into past seasons data to run linear regressions on trends, etc, then having is much data as possible is best. http://www.retrosheet.org/ has very detailed records available from games that will give you everything you need, minus the betting lines.

great site, but useless for betting purpose ("minus the betting lines" )!!

[Indecent]

Play by play or pitch by pitch?

[suicidekings]

great site, but useless for betting purpose ("minus the betting lines" )!!

Depends entirely on what you're doing. If I pick up on a historical trend that appears to have merit in recent seasons why do I care how it was priced in past games? I only care if the current pricing available represents value or not.

[CHUBNUT]

after years of matching up every conceivable combination of stats ive come to the conclusion that the pitchers biorythms are the best pointer to the game, especially totals.