Does the AI undervalue defense?

David Watts · 05-10-2012, 11:51 AM

Bob Horner is a 9 out of 20 at third base and is also an 8 at first base. Daryl Sconiers is rated 8 at first base. Meanwhile, Evan Longoria is a 15 out of 20 at third. The AI is starting Horner at 3B and Sconiers at 1B. If you factor in defense, it seems like Longoria is the obvious choice.

The AI basically benched Longoria midway through the last season (59) and went with the Horner/Sconiers combo. Still doing it to start the 60 campaign. Longoria basically pinch hit and spot started for the entire second half of the 59 season, but still took home rookie of the year honors, .289 23 home runs and 60+ RBI.

I'm not whining here, I just wondered what y'all thought about the AI and defense.

Orcin · 05-10-2012, 12:01 PM

Yes, I think the AI does undervalue defense. Either that, or I OVERvalue defense. I'm not sure which.

jonpfl · 05-10-2012, 12:17 PM

Quote:

Originally Posted by David Watts

Bob Horner is a 9 out of 20 at third base and is also an 8 at first base. Daryl Sconiers is rated 8 at first base. Meanwhile, Evan Longoria is a 15 out of 20 at third. The AI is starting Horner at 3B and Sconiers at 1B. If you factor in defense, it seems like Longoria is the obvious choice.

The AI basically benched Longoria midway through the last season (59) and went with the Horner/Sconiers combo. Still doing it to start the 60 campaign. Longoria basically pinch hit and spot started for the entire second half of the 59 season, but still took home rookie of the year honors, .289 23 home runs and 60+ RBI.

I'm not whining here, I just wondered what y'all thought about the AI and defense.

From what I have seen, I think the AI does undervalue defense

Thx
jonpfl

bailey · 05-10-2012, 01:09 PM

Quote:

Originally Posted by Orcin

Yes, I think the AI does undervalue defense. Either that, or I OVERvalue defense. I'm not sure which.

Since Markus programmed the AI, you could say that Markus undervalues defense. Or, since Markus knows the true value of defense as implemented in OOTP, then maybe players overvalue defense.

madJ · 05-10-2012, 02:38 PM

Winning 1-0 is not as exciting as winning 20-19. Well, that's the way a fan would think though.

But yes, the AI undervalues defense. Makes it more important to look at the players you want to acquire and think about the role they should play in your system.

Childress · 05-10-2012, 03:12 PM

Anecdotal, but with the last version I made a save of the game and duplicated it. I had an mediocre team, sub .500, with very undistinguished fieldeing, especially at the key positions; ss, cf, c. I jacked the up the defense ratings via the Editor to Hall of Fame levels. Nothing else was altered. And replayed the season. My guys ran away with the pennant. ERA dropped a full point.

awick · 05-10-2012, 03:53 PM

My experience agrees with the comments here -- the AI seems to undervalue defense. When I take over a weak team, the first thing I do is upgrade the defense -- a lot, if possible. Even if I degrade the offense somewhat in the process and make no changes to the pitching staff, that gets me a marked improvement in the club's record.

RchW · 05-10-2012, 07:50 PM

This is what Markus said in a thread on AI lineup selection. Bold is mine. What is not clearly stated below is that too many players (at least until v12) were generated with capability at multiple positions and the AI uses that to play musical chairs on defense. I haven't checked but can say with some certainty that no MLB team switches 2B to SS and SS to 2B strictly on the handedness of the pitcher. IRL Tampa Bay's Zobrist does 2B/RF. In OOTP this is a regular occurrence and is often worse, involving 3-5 positions with CF playing 3B and 1B playing SS depending on the opposing pitcher.

That being said a quick look at my converted v12 league in v13 suggests that this has been greatly improved. The only thing that still irritates me is SS/2B switches.

Quote:

Originally Posted by Markus Heinsohn

snip

The AI in terms of constructing the depth chart is pretty complex and leads to the best combination of offense & defense playing on the field. In real life, other factors influence the decision of the manager, i.e. personal preferences & stuff like team chemistry, these do not matter for the AI obviously.

Anyway, I do agree that the development of the player defensive ratings could get some work. Right now, the overall defensive rating at a position is determined by experience & the composite defensive ratings. The latter do change over time (i.e. decrease when aging), experience just grows though (until maxed out). Maybe when a player does not play a position at the higher end of the defensive spectrum than his current position, he should lose his experience over time?

David Watts · 05-11-2012, 10:36 AM

Time for me to eat some crow. Ugh! I looked at Horner's 1B rating wrong. He is a 4 out of 20, not an 8. So playing Sconiers at 1B makes a whole lot more sense now. Now, the question is, will the AI trade Horner or Longoria? Horner has put up big power numbers in the first 2 seasons, 28 bombs in an injury shortened 58 season and 36 in 59. The goofy thing is, I was the one the decided to draft Longoria. The AI wanted to take a pitcher in the 2nd round, but I stepped in to void that decision. My fault.

The Wolf · 05-11-2012, 01:07 PM

Don't beat yourself up over it. The pitcher might have been a bust or gotten injured.

05-10-2012, 11:51 AM	#1
David Watts Hall Of Famer Join Date: Apr 2002 Location: Looking for a place called Leehofooks Posts: 9,942 Infractions: 0/1 (1)	Does the AI undervalue defense? Bob Horner is a 9 out of 20 at third base and is also an 8 at first base. Daryl Sconiers is rated 8 at first base. Meanwhile, Evan Longoria is a 15 out of 20 at third. The AI is starting Horner at 3B and Sconiers at 1B. If you factor in defense, it seems like Longoria is the obvious choice. The AI basically benched Longoria midway through the last season (59) and went with the Horner/Sconiers combo. Still doing it to start the 60 campaign. Longoria basically pinch hit and spot started for the entire second half of the 59 season, but still took home rookie of the year honors, .289 23 home runs and 60+ RBI. I'm not whining here, I just wondered what y'all thought about the AI and defense.

05-10-2012, 12:01 PM	#2
Orcin Hall Of Famer Join Date: Aug 2010 Location: Indiana Posts: 9,850	Yes, I think the AI does undervalue defense. Either that, or I OVERvalue defense. I'm not sure which.

05-10-2012, 02:38 PM	#5
madJ Major Leagues Join Date: Mar 2012 Posts: 361	Winning 1-0 is not as exciting as winning 20-19. Well, that's the way a fan would think though. But yes, the AI undervalues defense. Makes it more important to look at the players you want to acquire and think about the role they should play in your system.

05-10-2012, 03:12 PM	#6
Childress Minors (Double A) Join Date: Oct 2007 Posts: 180	Anecdotal, but with the last version I made a save of the game and duplicated it. I had an mediocre team, sub .500, with very undistinguished fieldeing, especially at the key positions; ss, cf, c. I jacked the up the defense ratings via the Editor to Hall of Fame levels. Nothing else was altered. And replayed the season. My guys ran away with the pennant. ERA dropped a full point.

05-10-2012, 03:53 PM	#7
awick Major Leagues Join Date: Dec 2009 Location: Lansing, NY Posts: 313	My experience agrees with the comments here -- the AI seems to undervalue defense. When I take over a weak team, the first thing I do is upgrade the defense -- a lot, if possible. Even if I degrade the offense somewhat in the process and make no changes to the pitching staff, that gets me a marked improvement in the club's record.

05-11-2012, 10:36 AM	#9
David Watts Hall Of Famer Join Date: Apr 2002 Location: Looking for a place called Leehofooks Posts: 9,942 Infractions: 0/1 (1)	Time for me to eat some crow. Ugh! I looked at Horner's 1B rating wrong. He is a 4 out of 20, not an 8. So playing Sconiers at 1B makes a whole lot more sense now. Now, the question is, will the AI trade Horner or Longoria? Horner has put up big power numbers in the first 2 seasons, 28 bombs in an injury shortened 58 season and 36 in 59. The goofy thing is, I was the one the decided to draft Longoria. The AI wanted to take a pitcher in the 2nd round, but I stepped in to void that decision. My fault.