|
||||
|
|
Suggestions for Future OOTP Versions Post suggestions for the next version of Out of the Park Baseball here! |
![]() |
|
Thread Tools
![]() |
![]() |
#1 | |
Minors (Rookie Ball)
Join Date: Jul 2011
Posts: 27
|
Retrosheet Style Logging
Creating game logs following the style used by Retrosheet would go a long way towards resolving internal inconsistencies between OOTP's data (a few of which I have recently noted in the Bug Reports forum) as well as making it much easier for data nerds to perform statistical analysis.
Retrosheet describes their event files here. Typically, one file records a single game's events. These files are mostly human-readable, especially once you learn the codes. For example, this single line: Quote:
8: This event occurred in the eighth inning. 0: This event occurred in the top of the inning. bellg001: This event occurred with George Bell at the plate. 21: This event occurred on a 2-1 count. CBBX: The pitch sequence for this plate appearance was called strike, ball, ball, ball put into play by batter H: This event is a home run. L7D: This batted ball was a line drive to deep left field. All sabermetric research that is done these days using what is called "Retrosheet data" is using data aggregated from Retrosheet's event files. The data in these files can be parsed to build a database that can be used for statistical reporting and analysis. Personally, I want to generate more interesting and (hopefully) more useful defensive statistical analysis but am frustrated in my attempt to do so by OOTP's current method of logging. It is taking a lot of effort to parse the files that are generated by OOTP in a way that catches all of OOTP's glitches and inconsistencies while also aggregating the data that I want. Also, by recording events in this manner, we only need to store this event file per game, rather than storing a game log and a box score and updating players_at_bat_batting_stats and hoping it's all done consistently (which it's not). Then, plain English game logs and box scores can be generated from the event files on demand (for example, Retrosheet provides a file to generate box scores from their event files). Because all statistics trace back to the event files, we can better guarantee internal consistency by generating our statistical database from one source, instead of what appears to be OOTP's current method of updating multiple fields in multiple tables and hoping that we don't miss any (sometimes we do). It is a source of frustration for me that a franchise that prides itself on the detail and accuracy of its statistical simulation is unable to record and present statistics in a consistent fashion and that it is nearly impossible to explore other concepts like run expectancy tables, batter spray charts, and more historical splits than just L/R (OOTP doesn't keep historical home/road data, so if you want to calculate park factors for a season other than the current one, I hope you still have your old CSV exports hanging around!). Logging events in the tried-and-true method that Retrosheet has perfected would go a long way towards addressing these shortcomings and frustrations. |
|
![]() |
![]() |
![]() |
#2 |
Hall Of Famer
Join Date: Nov 2005
Posts: 3,067
Infractions: 1/1 (1)
|
+1
|
![]() |
![]() |
![]() |
#3 |
Minors (Rookie Ball)
Join Date: Jul 2011
Posts: 27
|
And if that's too much trouble, I didn't even realize that Retrosheet was now releasing CSV-formatted files that are already parsed and organized. Just follow this structure!
|
![]() |
![]() |
![]() |
#4 |
Minors (Double A)
Join Date: May 2002
Location: Granby, CT
Posts: 127
|
I've been working on a python script to scrape the box score and game log files to do this, but it sure would be nice if the game just did it itself. There are so many useful tools out there for dealing with the retrosheet files, it would be awesome for the game to utilize that format as well. It would open up a whole range of analysis which could help make the game engine even more accurate by identifying areas where it differs significant from real life baseball.
Last edited by FleetWalker; 11-08-2024 at 10:55 PM. |
![]() |
![]() |
![]() |
Bookmarks |
|
|