parsing the pbp logs

uruguru · 08-01-2024, 09:24 AM

I've been toying around with the idea of writing a program that could parse the game's pbp logs and create some sort of scoresheet and a maybe a tool to query the game's pbp data. I did this a long, long time ago for Retrosheet data (real baseball) so I'm thinking maybe I can do the same for OOTP pbp logs (fake baseball).

I was looking through the html of the pbp logs and it seems pretty doable except for one issue... some of the more unusual plays don't have the play scoring.

Obviously, all of the easy G6-3, F8, U3 stuff is covered. That's 99% of the plays. But there are a few I've seen that are missing.

Some examples:

- An 8-5 trying to stretch a double

Quote:

1-2: DOUBLE (Flyball, 8RXD, EV 102.5 MPH) - OUT at third base trying to stretch hit.

- A 1-6-3 DP turned on a sac bunt attempt

Quote:

2-0: Sac Bunt - play at second, runner OUT -> throw to first, DP!

- An 8-5-4-6 rundown.

Quote:

Runner from 2nd tries for 3rd, OUT after rundown.

So, the my question is: are these types of issues considered bugs? I can understand if they are not. After all, the fielding reports are correct. But if so, should I collect them all and report them?

fredbeene · 08-01-2024, 08:38 PM

Quote:

Originally Posted by uruguru

I've been toying around with the idea of writing a program that could parse the game's pbp logs and create some sort of scoresheet and a maybe a tool to query the game's pbp data. I did this a long, long time ago for Retrosheet data (real baseball) so I'm thinking maybe I can do the same for OOTP pbp logs (fake baseball).

I was looking through the html of the pbp logs and it seems pretty doable except for one issue... some of the more unusual plays don't have the play scoring.

Obviously, all of the easy G6-3, F8, U3 stuff is covered. That's 99% of the plays. But there are a few I've seen that are missing.

Some examples:

- An 8-5 trying to stretch a double

- A 1-6-3 DP turned on a sac bunt attempt

- An 8-5-4-6 rundown.

So, the my question is: are these types of issues considered bugs? I can understand if they are not. After all, the fielding reports are correct. But if so, should I collect them all and report them?

according to others the devs are overloaded and can't handle any more work. social media not with standing.
so i think they would like you to do that.

I 'thought' all pbp was fluff. At some point posters were saying the play out come is detrmined before you hit click. pbp is randolmly inserted tokens from english.xml file

NoOne · 08-01-2024, 09:29 PM

the game is not physics based, so where it goes on the field is "fluff" to some extent but should still mimic whatever is normal.

there are grounded into double plays and line drive double plays, not sure if the game distinguishes, but seems to based on how DP is higher than GIDP and both are output by the stat dumps for league total yearly stats. But maybe this is only meant to be DPs from the outfield or other rare occurences? can't be sure

so in some ways it is dictated but not all and only in a general context sense.. won't see too many DP from outfield without a major running blunder but does happen, obviously. so most dp will involve the IF.

About beene's comment - no matter what the resulotion it mostly works at - per pitch or per PA, you still depend on an RNG. if you want to feel it is predetermined, that's a bit of a false perception. It's just not resolved down to per pitch (in all ways), unless something drastic changed last year or 2. That detail has ramifications.

If you get a good "seed" (or think of it as a digital dice roll) to start a PA, chances are you get a good result no matter what you do with the per pitch options beside the obvious ways to sabotage yourself. This is why the "take a pitch until 2-strikes" method works well. you don't ge the major negative effect of a 2-strike count because that is already included in the percentages of what varying values of the seed result in. Kinda prvents them from killing this manipulation/cheat code, too.

Fielding errors are seperate.. certain things you do pre-PA will impact percentage of each potential outcome of a PA - like defensive adjustments etc. you can definitely improve your odds by doing the "right" things.

E.g. the player AB gets a seed right near threshold of being a nearly guaranteed hit (but no power), changing the defensive alignment may shift that threshold further up with pull hitters etc etc... now it's a good chance to be an out with that same seed value because of how you moved the defense to match context. you still can make decisions that matter.

uruguru · 08-02-2024, 10:43 AM

I think from the current pbp logs I can generate a pitcher vs batter dataset that would be relatively straightforward to query against ("rank all of the hitters by HR for July",
"what are the career hitting stats against the Yankee pitchers", etc), but without the play data there's no reliable way to do fielding stats or scoresheets.

I just need an idea of how much I should constrain this project.

08-02-2024, 10:43 AM	#4
uruguru All Star Starter Join Date: May 2022 Posts: 1,364	I think from the current pbp logs I can generate a pitcher vs batter dataset that would be relatively straightforward to query against ("rank all of the hitters by HR for July", "what are the career hitting stats against the Yankee pitchers", etc), but without the play data there's no reliable way to do fielding stats or scoresheets. I just need an idea of how much I should constrain this project. Last edited by uruguru; 08-02-2024 at 10:47 AM.

08-01-2024, 09:29 PM	#3
NoOne Banned Join Date: Apr 2015 Posts: 7,273 Infractions: 0/1 (3)	the game is not physics based, so where it goes on the field is "fluff" to some extent but should still mimic whatever is normal. there are grounded into double plays and line drive double plays, not sure if the game distinguishes, but seems to based on how DP is higher than GIDP and both are output by the stat dumps for league total yearly stats. But maybe this is only meant to be DPs from the outfield or other rare occurences? can't be sure so in some ways it is dictated but not all and only in a general context sense.. won't see too many DP from outfield without a major running blunder but does happen, obviously. so most dp will involve the IF. About beene's comment - no matter what the resulotion it mostly works at - per pitch or per PA, you still depend on an RNG. if you want to feel it is predetermined, that's a bit of a false perception. It's just not resolved down to per pitch (in all ways), unless something drastic changed last year or 2. That detail has ramifications. If you get a good "seed" (or think of it as a digital dice roll) to start a PA, chances are you get a good result no matter what you do with the per pitch options beside the obvious ways to sabotage yourself. This is why the "take a pitch until 2-strikes" method works well. you don't ge the major negative effect of a 2-strike count because that is already included in the percentages of what varying values of the seed result in. Kinda prvents them from killing this manipulation/cheat code, too. Fielding errors are seperate.. certain things you do pre-PA will impact percentage of each potential outcome of a PA - like defensive adjustments etc. you can definitely improve your odds by doing the "right" things. E.g. the player AB gets a seed right near threshold of being a nearly guaranteed hit (but no power), changing the defensive alignment may shift that threshold further up with pull hitters etc etc... now it's a good chance to be an out with that same seed value because of how you moved the defense to match context. you still can make decisions that matter.