Do you know how the AI calculates...

[DetroitStyle]

That debacks lineup looks pretty good to me. Upton as the lead off with guys like goldschmidt in the power role makes sense. Upton has speed and can hit for a high average and power. Jacoby went 30/30 from his lead off spot so it's not that far fetched...

[EAGLESFAN10]

Re: Do you know how the AI calculates...

Dont line-ups go off of contact and power ratings

Like I just started my YANKEE chise and the rays had Pena hitting 8th the 1st game, 2nd game he was batting 3rd or 4th

[aykyle]

Re: Do you know how the AI calculates...

Yea. It seems to me that every team has their best contact swinger or best on base percentage from the previous year, I would assume. It isn't random.. at all. There is a concept behind it. It is just a code we have to break and a pattern we have to find.

Inspector Kyle is on the case.

[wudl83]

Re: Do you know how the AI calculates...

The people who care, control all teams. Good that we have that option is how I see it.

So you want the CPU to not only keep MLB-accurate lineups, you also want it to switch around hot and cold guys? You are asking for the moon IMO.

Perhaps there are people that CARE but have not the time to edit always lineups for 30 teams. Is the demand for realism only legit for those who have more time?

That debacks lineup looks pretty good to me. Upton as the lead off with guys like goldschmidt in the power role makes sense. Upton has speed and can hit for a high average and power. Jacoby went 30/30 from his lead off spot so it's not that far fetched...

Why should you set the best contact and power hitter into slot 1 when having NO other player coming close to his ratings? You set him in 3 or 4.
Do you think the Sox would play Ellsbury in first if they had no Pedroia/Gonzalez/Ortiz?

Yea. It seems to me that every team has their best contact swinger or best on base percentage from the previous year, I would assume. It isn't random.. at all. There is a concept behind it. It is just a code we have to break and a pattern we have to find.

Inspector Kyle is on the case.

I didn't want to say that it is easy to set up those criterias but those guys are paid for doing a good job. They have some time to test things and when they see that there happen things that DON'T happen in real life they shouldn't let it be. Instead they should search for the cause and work on it. But they do NOT work on it and they have NOT worked on it for years.

It is the same with the position alignment in the field, where CFs play RF and RFs play CF although they would have higher ratings at the primary positions.



EDIT:
I am seeing that the thread wents into the wrong direction.
I didn't mean this thread as a complaining thread, I wanted to start a discussion on how exactly the lineups / positions are set up ingame because there are some things I can't find out.

[aykyle]

Re: Do you know how the AI calculates...

I didn't want to say that it is easy to set up those criterias but those guys are paid for doing a good job. They have some time to test things and when they see that there happen things that DON'T happen in real life they shouldn't let it be. Instead they should search for the cause and work on it. But they do NOT work on it and they have NOT worked on it for years.

It is the same with the position alignment in the field, where CFs play RF and RFs play CF although they would have higher ratings at the primary positions.
[/I][/B]

They simply CAN'T mimic real life in a video game. It is impossible. Real life managers switch their lineups around depending on how a player has been performing, if a player has just gotten off an injury and if the current lineup isn't working. It is IMPOSSIBLE to code that into a game. They can't work on it because there is nothing to work on. They can't improve it anymore. Technology won't let them. You can't hack into the mind of a manager and determine what he will be doing with his lineups on a given day. And the people that code the game, don't necessarily TEST the game. They test to see if it works, but there are paid beta testers that know how to fix the bugs they encounter that will do 99% of the testing. So the coders don't have time to fix these problems(which imo, there isn't a problem).

And in the case of CF's playing RF is because if you look at their fielding attributes, the player that is playing RF but typically plays CF has a RF fielding rating that is decent enough for the computer to say: Hey, this guy can play this position so we can squeeze this guy into CF to make the best possible team. Players in the MLB aren't constricted to ONE position. Most MLB players can play multiple positions if they are asked to. You see MLB Catchers switch to infielders all the time. You see 1B players switch to 3B(Cabrera and Fielder). Outfielders can play RF, LF or CF. It isn't hard. So, the game actually putting some players at RF when IRL they play CF is realistic enough. The computer does that because if the CF guy CAN play RF if the team doesn't have anyone else in the lineup that CAN play RF, they put them there so another guy who is better at playing CF than RF could play.
If you don't understand- A CF player can play both CF and RF. But the team has no RF to put in the lineup, but they have an extra CF player. They will move the guy that can play both CF and RF to RF and the backup CF player to CF. Get it? I do understand that the CPU should trade more, but they can't do that without the CPU trading too much or too little. In this case, they trade too little. I think that when technology and knowledge gets there, they should make it so if a team does not have a RF, they go out and trade for a better one. in 2K12, the CPU will sometimes sign a free agent RF to fill the spot if there is no one else that can play. So this is not a glitch, bug or un-real feature. It is perfectly realistic and if you do not like it, then don't look at it.

[wudl83]

Re: Do you know how the AI calculates...

No, no, no, they don't do it so it IS the best possible team.

Look at the ratings of Young and Upton and think again. When the AI switches Young and Upton the AI does this NOT because AFTER that the Dbacks have the best defensive lineup. Look at the ratings, this is simply not true.

I know that the most players can play multiple positions but when you have a RF like Upton which NEVER has played CF why should you switch him with your really good defensive CF, which has nearly never played RF?

Your argumentation isn't right, it simply isn't right.

Example:
The AI has
- Upton a primary RF, secondary CF
- Young a primary CF, secondary RF
The AI makes it
- Upton CF
- Young RF
Where is the sense? In the nirvana.

And this is by far not the only example. They switch the Rockies around in the same way. They put Gonzalez in CF and Fowler at RF and Cuddyer at LF. Which is the uggliest thing I've seen til now, because Gonzalez is LF, Fowler CF and Cuddyer RF. So no position is right.

There needs no new player to come in. They switch around and switch around.

The Giants do it also with Cabrera and Pagan.

When there are injuries etc. I do not care when they switch around. But if the best players are available they should play at their primary positions.

The Fielder/Cabrera-Example has nothing to do with it.

[Yeats]

Re: Do you know how the AI calculates...

Your argumentation isn't right, it simply isn't right.

Some of the 2K defending here is definitely silly no doubt. We all know this game has AI issues that they've never addressed in seven years, and the boneheaded lineup logic is one of those issues.

Still, I think you're missing the point. While they could easily fix the logic so that the starting fielders in the lineup played in their correct positions, they can't be expected to adjust lineups according to how players are doing game-to-game or as the season progresses. That's unreasonable IMO.

[wudl83]

Re: Do you know how the AI calculates...

Some of the 2K defending here is definitely silly no doubt. We all know this game has AI issues that they've never addressed in seven years, and the boneheaded lineup logic is one of those issues.

Still, I think you're missing the point. While they could easily fix the logic so that the starting fielders in the lineup played in their correct positions, they can't be expected to adjust lineups according to how players are doing game-to-game or as the season progresses. That's unreasonable IMO.

To the first paragraph:
Yes, I agree.

To the second paragraph:
You got me wrong. I do not expect 100% realistic or accurate lineups, I simply wanted to know if somebody found out what patterns where used for the AI setting up the lineups.

[elimack]

Re: Do you know how the AI calculates...

Code = Criteria. If you set the criterias right then you won't have such big flaws.

I do not expect a 100% accurate game. But the status quo is nearly unacceptable to me.

And controlling 30 teams isn't simple. You have to look for injuries and cold and hot streaks when you want realism.

Sometimes I get the feeling that many people simply don't care.

There are so big problems. When a CF and RF are available to play their positions, I don't see any reason in reality why they should switch positions.

Regardless of what the criterias look like. These criterias must be big crap.



I do that too but that does not help for problems like the Gonzalez/Fowler or Cabrera/Pagan one.

Cuddyer in reality can play 4-5 positions. In Colorado they have a hole at RF so they play them there. Open the game, make a franchise and they play him at third base.

I DID edit a bunch of ratings and it didn't help really. For some problems it helped, but for the majority it did not help.

When you make changes to fielding abilities, what do you use to determine appropriate ratings?

Offense plays as big a role as defense in the computers calculation of "best" possible lineup combinations. I had this issue with Maicer and Aybar out in ANA, but with proper defensive and offensive rating (all based on real life performance), it was solved.

I'll begin running test on 12 and see if I find any patterns

[wudl83]

Re: Do you know how the AI calculates...

When you make changes to fielding abilities, what do you use to determine appropriate ratings?

Offense plays as big a role as defense in the computers calculation of "best" possible lineup combinations. I had this issue with Maicer and Aybar out in ANA, but with proper defensive and offensive rating (all based on real life performance), it was solved.

I'll begin running test on 12 and see if I find any patterns

I changed the "fielding abilities" for the different positions first. Then arm strength etc.
To determine appropriate ratings I use the fielding statistic e.g. from fangraphs.

It was really annoying. When I lowered e.g. Fowler's RF and LF ability while hoping that the CPU must put him into CF since he would be too bad at other positions, they called up Blackmon and benched Fowler. The outfield looked like this after that: LF Cuddyer (primary RF) - CF Gonzalez (primary LF) - RF Blackmon (primary LF). And Fowler benched.

[Yeats]

Re: Do you know how the AI calculates...

Offense plays as big a role as defense in the computers calculation of "best" possible lineup combinations.

Offensive ratings are the biggest part by far. Offensive ratings comprise probably 75% of a player's overall rating. The game is way out of balance in that regard and always has been.

[aykyle]

Re: Do you know how the AI calculates...

No wudl, my argument is right. You are just not getting it. Upton is at CF because he can play CF, his fielding stat for CF is 80. Yes Young is a better fielder at Center than Upton, but Upton has a far better Arm Strength than Young does. THAT is why the CPU has moved him to CF. CF HAS to have a good arm. So YOUR argument is wrong. Same with the Rockies. Gonzalez has better ARM STRENGTH than Fowler(who is at LF) does. So again, MY argument is right. And YOUR argument is wrong.
Another example is on the Giants(note: in my Franchise, Giants signed Vladimir Guerrero to their team[he plays LF]). The Giants have Melky Cabrera at CF(arm strength of 75) and Pagan(arm strength of 69) on the bench. Guerrero(arm strength of 84, primary position is RF) plays LF and Schierholtz(arm strength of 98 and primary position is RF) at RF.

As for the lineup, EVERY TEAM has a FAST guy at lead off. So the pattern for every team is that their speed guy is at lead off. And if their lead off isn't their FASTEST guy, it is whoever has the most contact and speed combination. Except for the D-backs. Where Upton has the best Speed and Contact then their lead off. But Upton has far better Power so the CPU moved him to 3rd which is good. And also with the Red Sox, where Crawford is lead off and Ellsbury has better speed and contact. But Ellsbury Has one of the best Power to Contact ratio. So they put him at 4th. Again, this is right.

So THAT is how the pattern is going. For players that are switched from CF to RF/LF, it is because they have decent enough CF fielding attribute and better Arm Strength than the guy that is suppose to play CF. And for the lineups, the best Contact to Speed ratio is their lead off man. Everything else I couldn't tell you. Except there are some patterns for 4th which it is the best Contact to Power ratio. But not all teams.
You can check it if you want. My roster might be different than yours. But that is the pattern I saw. Don't argue with it because I know I am right for my roster.

[wudl83]

Re: Do you know how the AI calculates...

No wudl, my argument is right. You are just not getting it. Upton is at CF because he can play CF, his fielding stat for CF is 80. Yes Young is a better fielder at Center than Upton, but Upton has a far better Arm Strength than Young does. THAT is why the CPU has moved him to CF. CF HAS to have a good arm. So YOUR argument is wrong. Same with the Rockies. Gonzalez has better ARM STRENGTH than Fowler(who is at LF) does. So again, MY argument is right. And YOUR argument is wrong.
Another example is on the Giants(note: in my Franchise, Giants signed Vladimir Guerrero to their team[he plays LF]). The Giants have Melky Cabrera at CF(arm strength of 75) and Pagan(arm strength of 69) on the bench. Guerrero(arm strength of 84, primary position is RF) plays LF and Schierholtz(arm strength of 98 and primary position is RF) at RF.

As for the lineup, EVERY TEAM has a FAST guy at lead off. So the pattern for every team is that their speed guy is at lead off. And if their lead off isn't their FASTEST guy, it is whoever has the most contact and speed combination. Except for the D-backs. Where Upton has the best Speed and Contact then their lead off. But Upton has far better Power so the CPU moved him to 3rd which is good. And also with the Red Sox, where Crawford is lead off and Ellsbury has better speed and contact. But Ellsbury Has one of the best Power to Contact ratio. So they put him at 4th. Again, this is right.

So THAT is how the pattern is going. For players that are switched from CF to RF/LF, it is because they have decent enough CF fielding attribute and better Arm Strength than the guy that is suppose to play CF. And for the lineups, the best Contact to Speed ratio is their lead off man. Everything else I couldn't tell you. Except there are some patterns for 4th which it is the best Contact to Power ratio. But not all teams.
You can check it if you want. My roster might be different than yours. But that is the pattern I saw. Don't argue with it because I know I am right for my roster.

Now tell me where my argument is wrong. The only thing you stated is that 2k makes BS.
CF are never only chosen by arm strength. Other than that all the coaches out there that don't handle it this way must be idiots.

And in the game CF aren't only chosen by arm strengths, too. I upped Pagan's arm strength to 1 point higher than Melky's and they still played Melky @CF.

They didn't play Melky at CF until I lowered his CF ability to 60 or so.

Same with Gonzalez, Upton, and so on.

Not only the arm strength is determining.

[aykyle]

Re: Do you know how the AI calculates...

Now tell me where my argument is wrong. The only thing you stated is that 2k makes BS.
CF are never only chosen by arm strength. Other than that all the coaches out there that don't handle it this way must be idiots.

And in the game CF aren't only chosen by arm strengths, too. I upped Pagan's arm strength to 1 point higher than Melky's and they still played Melky @CF.

They didn't play Melky at CF until I lowered his CF ability to 60 or so.

Same with Gonzalez, Upton, and so on.

Not only the arm strength is determining.

Well.. for one you used the word "argumentation".
But the CPU has edited them before hand. Not after you have edited them. I have NEVER said that REAL LIFE CF are chosen by arm strength. I said I saw a PATTERN in the GAME that the CF are chosen by it. THE GAME. NOT real life. First you make up words and now you can't read. And I wasn't just talking about your argument(if you even call it that, you had no point at all in your post). All I said was I saw the pattern, I never said anything about Real Life. This is a thread on how the CPU determines their lineup, not real life managers. YOU changed this thread into "why is the CPU making unrealistic changes to the lineup". I know MLB Managers do not choose based on arm strength. If you moved the CPU lineups to their correct ones after you edited peoples arm strength, the CPU might not change it back. Or they could change them back to their incorrect format. You gave absolutely no input to this discussion. All you proved was that the CPU makes wrong position changes, You asked for someone to determine the pattern on how the AI chooses position players and lineups, I gave it to you. But you clearly don't understand anything if after I gave you a pattern, you still reject the concept. Please read before you post.

[wudl83]

Re: Do you know how the AI calculates...

The problem is that you think you gave me the pattern, but you obviously gave me only a piece of it. That is why I told you that e.g. they switched Pagan into corner outfield for Melky ALTHOUGH I gave Pagan better arms than Cabrera, do you now understand what I mean?