OOTP23 New Sim Engine Discussion

[Pdubya64]

Hey all.
So I watched the Scott Braun interview vid with Markus and while it didn't reveal anything "new", I was really interested in Markus' points regarding the completely rewritten Simulation Engine.

So my question to everyone to start the discussion is this...
Will you play your Franchise games with default setup for the player AI ratings percentages? (x/x/x/x)

Personally, I always start a new year with defaults for this particular aspect of OOTP, then dig into threads here to get a feel for how others see things playing out. I definitely want to give Markus the benefit of the doubt out of the box this year.
I am curious what you all think.

[pgjocki]

Are you referring to the default settings for AI evaluation? I didn't catch the video so I'm not 100% tracking what was stated.

But to answer you question I always run a few test leagues with every version just to observe what happens before I transfer over my long term save or start a new one.

I'm looking for major outliers - which is still possible no matter how tight the statistical correlations are in the code (I'm thinking 70+ home runs, batting average over .350, etc) as well as looking for how the AI responds to functions such as roster management (signing and releasing players, trades - both number and "logic") and financials. I like to see if the payrolls change over time so I'll sim 10-20 seasons where I record what each teams final payroll was at the end of the season and see if anything changes over time.

[Pdubya64]

Originally Posted by pgjocki

Are you referring to the default settings for AI evaluation? I didn't catch the video so I'm not 100% tracking what was stated.

Yes, I believe that is what it is called (with the 4 categories). Made an edit to my post, thanks.

[Bluenoser]

The sim engine wasn't completely rewritten, it was upgraded.

From the release notes:

In-game simulation engine upgrades! We’ve crunched massive amounts of data and crafted a significantly improved in-game core simulation engine, which will benefit not just the traditional Franchise Mode but also Perfect Team! Plus, all-new ratings balances for historical players AND a much-improved version of Negro League player ratings are in place!

[Anteater]

In the reveal video around the 20:45 mark Markus said "We've rebuilt the core engine from scratch" so it certainly sounds like a pretty big upgrade. Now what rebuilding from scratch really means could range from a complete rewrite of the core engine code to upgrading what is already there. Either way it sounds like fundamental improvements under the hood, which is good. Although, we'll probably see a few patches to fix some unexpected stats once it's in widespread use. I'm hoping Markus and Matt give us some more behind the scenes background once things slow down for them post release.

[Bluenoser]

Originally Posted by Anteater

In the reveal video around the 20:45 mark Markus said "We've rebuilt the core engine from scratch" so it certainly sounds like a pretty big upgrade. Now what rebuilding from scratch really means could range from a complete rewrite of the core engine code to upgrading what is already there. Either way it sounds like fundamental improvements under the hood, which is good. Although, we'll probably see a few patches to fix some unexpected stats once it's in widespread use. I'm hoping Markus and Matt give us some more behind the scenes background once things slow down for them post release.

Hope you're right. Not nice when they say one thing and post another...

[MathBandit]

Originally Posted by Bluenoser

Hope you're right. Not nice when they say one thing and post another...

They've also said in one of these threads that instead of the game engine operating in "One-Pitch" mode behind the scenes (even if you play in Pitch-by-Pitch), the engine will now operate in Pitch-by-Pitch.

[Chimaera]

Originally Posted by MathBandit

They've also said in one of these threads that instead of the game engine operating in "One-Pitch" mode behind the scenes (even if you play in Pitch-by-Pitch), the engine will now operate in Pitch-by-Pitch.

https://forums.ootpdevelopments.com/...4&postcount=31


I don't know how you could say 100 percent that the game will now operate pitch-by-pitch from what Matt has said here. All that he's said is that cheesing the game to two strikes while managing out games might be different. It could be the game still operates in One-Pitch mode but you just can't cheese it. It could be the game now operates pitch-by-pitch for everything.

[Matt Arnold]

Originally Posted by Chimaera

https://forums.ootpdevelopments.com/...4&postcount=31


I don't know how you could say 100 percent that the game will now operate pitch-by-pitch from what Matt has said here. All that he's said is that cheesing the game to two strikes while managing out games might be different. It could be the game still operates in One-Pitch mode but you just can't cheese it. It could be the game now operates pitch-by-pitch for everything.

Yeah, it's not quite pitch by pitch mode, but basically the game will be able to adjust better to the situation at hand. So sort of a mix overall. I'm sure people will be able to find a new way to cheese the way it changed But I assume not for at least a couple days

[Syd Thrift]

Yeah, *I* speculated that that might be different, but that's all that was, speculation. I personally just have the AI make offensive/defensive strategy decisions when I play because I tend to play against myself, and so that in turn means that I'll still be allowing the game to figure all of this out, but it'll be interesting to see how it plays. My guess is that we won't really "grok" the differences in actual gameplay until someone does a 200 year simulation and reports the results or something...

Or, what Matt Arnold said.

[Syd Thrift]

dola,

I feel like in order to really and truly capture pitch by pitch mode you'd need to add some more ratings / characteristics for players...

- BABIP I guess can be what it says it is, but really and truly you'd want something like line drive rate, which would interact with speed and GB/FB tendency

- Avoid Ks and Eye would need to be split out into a few new categories, like Pitch Judgement (ability to recognize a pitch and its eventual location), Selectivity (related to the previous but some batters don't swing at all until they see something or at least think they see something in one particular part of the zone), Strike Zone Coverage (related to the previous; how much and what parts of the zone will a batter swing at), Wrist Strength (the ability to check a swing when a batter realizes they made a mistake in swinging at something), Contact (as in, the pure ability to get wood on a ball once the batter does decide to swing - there are definitely guys with massive strike zones who nevertheless excel at fouling everything off), and so on.

To make matters worse for historical players, all of this could probably be gleaned from modern Statcast stats but would need to be figured out based on limited stats for pre-2010 leagues. I do think that eventually the game is going to want to go in a physics-based direction, and this would be part of it, but.. buckle your seatbelts when that happens...

It's probably entirely possible to do a full, robus pitch-by-pitch mode with the current ratings. It's just... I think the game will begin to show its limits when that happens.

[turtle4499]

Originally Posted by Syd Thrift

dola,
- BABIP I guess can be what it says it is, but really and truly you'd want something like line drive rate, which would interact with speed and GB/FB tendency

I don't want to get too into the details because I've done a fairly heavy deconstructing of the ingame calculations for 22 and it is like too revealing for a general comment. But the game does have a linedrive rate its driven by GAP power and it isn't subject to the same rules as normal babip or GB/FB rate. The unrealistic part is actually the GB to FB rate handling.

[Pelican]

Originally Posted by Bluenoser

The sim engine wasn't completely rewritten, it was upgraded.

From the release notes:

Seriously for me the upgrade to the Negro league player ratings, if done fairly, would be more than worth the price for OOTP 23.

[Syd Thrift]

Originally Posted by turtle4499

I don't want to get too into the details because I've done a fairly heavy deconstructing of the ingame calculations for 22 and it is like too revealing for a general comment. But the game does have a linedrive rate its driven by GAP power and it isn't subject to the same rules as normal babip or GB/FB rate. The unrealistic part is actually the GB to FB rate handling.

I'd really enjoy seeing those findings in its own post, if not as an aside here. Do you have a post on reddit?

[turtle4499]

Originally Posted by Syd Thrift

Do you have a post on reddit?

https://www.reddit.com/r/OOTP/commen...ing_mechanics/

I have another one on pitching mechanics pinned to my profile. Once this game releases I will make more in depth posts covering both and post them to reddit. I can post them here as well if people are interested in it. The big tool I need to spend some time setting up is carving out is a play-by-play file converter so I can do more in depth analysis on situational effects on outcomes. Situations dictate outcomes to such a high degree that I need dramatic sample sizes to measure effects.

[BusterKing]

Originally Posted by turtle4499

https://www.reddit.com/r/OOTP/commen...ing_mechanics/

I have another one on pitching mechanics pinned to my profile. Once this game releases I will make more in depth posts covering both and post them to reddit. I can post them here as well if people are interested in it. The big tool I need to spend some time setting up is carving out is a play-by-play file converter so I can do more in depth analysis on situational effects on outcomes. Situations dictate outcomes to such a high degree that I need dramatic sample sizes to measure effects.

No thanks, knowing all that kinda ruins the game for me.

[turtle4499]

Originally Posted by BusterKing

No thanks, knowing all that kinda ruins the game for me.

Yea that was sorta my concern and main reason for not posting it here. This stuff gets too in depth and too revealing.

[Sweed]

Originally Posted by BusterKing

No thanks, knowing all that kinda ruins the game for me.

Totally agree. The point of OOTP, for me, is to have a simulation of real baseball and all that entails. I don't want a large math problem that I can solve so I can win.

Originally Posted by turtle4499

Yea that was sorta my concern and main reason for not posting it here. This stuff gets too in depth and too revealing.

First I appreciate you not posting findings here. Like BK, this kind of thing ruins the game, IE the illusion that OOTP is presenting to me.

I am curious, and feel free to not answer, as to why you dig under the hood?
Natural curiosity? Play online or PT? Other?

[turtle4499]

Originally Posted by Sweed

I am curious, and feel free to not answer, as to why you dig under the hood?
Natural curiosity? Play online or PT? Other?

So I am not digging under the hood directly. I am not that skilled at decompiling to be able to do anything useful. Everything is done via testing and regressing. It is frankly something I enjoy doing as a software dev it is interesting to me how a game is handling things. This game is also a pretty amazing example of why statistical independence, outside of quantum mechanics, is nonsensical.

Even knowing exactly how batting events are calculated it is still the byproduct of dozens of variables that produces the actual result. It doesn't tell you: how does run variance change win-loss outcome groups over 162 game samples? How does era and ballpark effect this players value? How does this change come playoffs? The game still has tons of nuance going on.

[Sweed]

Originally Posted by turtle4499

So I am not digging under the hood directly. I am not that skilled at decompiling to be able to do anything useful. Everything is done via testing and regressing. It is frankly something I enjoy doing as a software dev it is interesting to me how a game is handling things. This game is also a pretty amazing example of why statistical independence, outside of quantum mechanics, is nonsensical.

Even knowing exactly how batting events are calculated it is still the byproduct of dozens of variables that produces the actual result. It doesn't tell you: how does run variance change win-loss outcome groups over 162 game samples? How does era and ballpark effect this players value? How does this change come playoffs? The game still has tons of nuance going on.

Thanks, I appreciate the answer.