Idea for AI Generated Depth Charts

GoldenGlove · 04-23-2004, 01:34 AM

1) Create a formula for the offensive ratings. A hidden system with weights on each rating.
2) Order the players based on this system for each team.
3) Place the top player at his best fielding position.
4) Place the next player at his best position. If it is taken by previous players, place this player at his best secondary position.
5) Continue down the list. If all of a player's defensive positions are filled, put him at DH or go on to the next player if there is no DH.
6) After the lineup is filled, compare the DH to each player at his fielding positions. If he has a better fielding ability, swap the players.

Let me know what you guys think. I'd really like to see this aspect improved.

swampdragon · 04-23-2004, 09:31 AM

I think you just benched Derek Jeter. Not that this is necessarily bad....

I agree it needs to be addressed. The irony is that OOTP5 did this better simply by paying attention to primary positions.
I would go something like this:

1. Make a provisional lineup based on the primary positions of each player.
2. Check the bench players to see if any of them are better than the regulars at another position.

crackpott · 04-27-2004, 06:28 PM

It's pretty good, but for fielding you have to take into account importance of fielding position.

For example, you have these three players (who just for ease, are your best three offensive players
player one: A range at third, B at second and shortstop
player two: C range at all of the above positions
player three: B range at third, B at shortstop, D at second
Under your system, player one would play third, player 2 would play shortstop, player 3 would play second, whereas the best situation would be player one at second, player two at third, player 3 at shortstop.

GoldenGlove · 04-27-2004, 06:42 PM

Another comparison could be made after the lineup has been completed. It could compare each player to everyone in the lineup, looking for the best combination. The positions, like the ratings, would also have to be weighted. This would be more of a judgement call though as a B SS/C 2B (IMO) would be better than a C SS/B 2B. Using statistical history, TC could be the basis for weighing positions.

crackpott · 04-27-2004, 06:47 PM

Well, that's basically what I do, although I'm not sure how easily that could be coded into AI. I shove my best batters into the lineup, then assign them positions based on where overall they generate the most value for the team. I'll then haul people out of the lineup who can't play anywhere (or are blocked by someone better) and replace them with the positions i need, and do one final check to make sure it's the ideal defensive alignment.

04-23-2004, 01:34 AM	#1
GoldenGlove Major Leagues Join Date: Feb 2004 Posts: 339	Idea for AI Generated Depth Charts 1) Create a formula for the offensive ratings. A hidden system with weights on each rating. 2) Order the players based on this system for each team. 3) Place the top player at his best fielding position. 4) Place the next player at his best position. If it is taken by previous players, place this player at his best secondary position. 5) Continue down the list. If all of a player's defensive positions are filled, put him at DH or go on to the next player if there is no DH. 6) After the lineup is filled, compare the DH to each player at his fielding positions. If he has a better fielding ability, swap the players. Let me know what you guys think. I'd really like to see this aspect improved. __________________ ISBL - New York Yankees B.I.G. - Oakland Athletics TheMLB - Cleveland Indians

04-27-2004, 06:28 PM	#3
crackpott Hall Of Famer Join Date: Dec 2001 Location: Toronto Posts: 2,961	It's pretty good, but for fielding you have to take into account importance of fielding position. For example, you have these three players (who just for ease, are your best three offensive players player one: A range at third, B at second and shortstop player two: C range at all of the above positions player three: B range at third, B at shortstop, D at second Under your system, player one would play third, player 2 would play shortstop, player 3 would play second, whereas the best situation would be player one at second, player two at third, player 3 at shortstop. __________________ Abandoned Planet - Check it out. The photos don't suck anymore.

04-27-2004, 06:42 PM	#4
GoldenGlove Major Leagues Join Date: Feb 2004 Posts: 339	Another comparison could be made after the lineup has been completed. It could compare each player to everyone in the lineup, looking for the best combination. The positions, like the ratings, would also have to be weighted. This would be more of a judgement call though as a B SS/C 2B (IMO) would be better than a C SS/B 2B. Using statistical history, TC could be the basis for weighing positions. __________________ ISBL - New York Yankees B.I.G. - Oakland Athletics TheMLB - Cleveland Indians

04-27-2004, 06:47 PM	#5
crackpott Hall Of Famer Join Date: Dec 2001 Location: Toronto Posts: 2,961	Well, that's basically what I do, although I'm not sure how easily that could be coded into AI. I shove my best batters into the lineup, then assign them positions based on where overall they generate the most value for the team. I'll then haul people out of the lineup who can't play anywhere (or are blocked by someone better) and replace them with the positions i need, and do one final check to make sure it's the ideal defensive alignment. __________________ Abandoned Planet - Check it out. The photos don't suck anymore.

04-23-2004, 09:31 AM	#2
swampdragon Hall Of Famer Join Date: May 2002 Location: The Lonely Mountain Posts: 2,506	I think you just benched Derek Jeter. Not that this is necessarily bad.... I agree it needs to be addressed. The irony is that OOTP5 did this better simply by paying attention to primary positions. I would go something like this: 1. Make a provisional lineup based on the primary positions of each player. 2. Check the bench players to see if any of them are better than the regulars at another position.