very odd computer AI decision (6.5)

[Giants44]

so the other day I noticed in my solo fictional league that SD had lost its mind and benched this guy:
http://www.abctechnical.com/other/Jim_Leiper.htm

for this guy
http://www.abctechnical.com/other/Merlin_Blanco.htm

for those to lazy or otherwise uninterested in click on the links - we are talking a 4 star 7 in contact 10 in power guy with 4 consecutive 30+ homer seasons for a 1 star - 3 in contact, 6 in power career .176 hitter

??????

[Mike D]

Known problem.

[Another Mike D]

It doesn't really fit Markus's description of the bug, but I guess it's the same. I thought the bug Markus had mentioned was when a player was rated sooooo poorly, his numbers wrapped around negatively making him look studly. If 3 in hits and 6 in power counts as being so low it looks great, perhaps the problem is bigger?

[akw4572]

Originally Posted by Giants44

so the other day I noticed in my solo fictional league that SD had lost its mind and benched this guy:
http://www.abctechnical.com/other/Jim_Leiper.htm

for this guy
http://www.abctechnical.com/other/Merlin_Blanco.htm

for those to lazy or otherwise uninterested in click on the links - we are talking a 4 star 7 in contact 10 in power guy with 4 consecutive 30+ homer seasons for a 1 star - 3 in contact, 6 in power career .176 hitter

??????

Are they in a pennant race? I noticed the better guy isn't singed past 1995, which looked like the current year, and the other guy was staying with the club.

[Pete Grassi]

Another reason the star system needs to be shut off. Your 4 star player is a career .250 hitter who seems to be in a power slump.

Also, is the team way ahead in the standings? Maybe resting the better player down the stretch.

[Giants44]

no to all these inquires - there is absoulety no logical reason for this choice by the AI

* the Padres are in second place
* they are set to "win now"
* Leiper might be a career .250 hitter, but he is projected for 28HR/112RBI this year and is coming off 37/129 and 37/122 seasons - He is a big time slugger
* more importantly Blanco should't even be in the big leagues. His contact skill and tallent are both at 3. He is a career .176 hitter - He is 29 - not a future prospect. Hell, he wasn't even on the roster till 9/1 call-ups
* someone mentioned contract and that can't be it either Leiper just signed a 5M x 5 year extension

Bottom line this is an odd bug, hopefully something that can be addressed in 6.51

[spleen1015]

What are your player evaluation settings? If the game is valuing stats more than ratings, I can see how the scrub could get put in over the other guy.

[endgame]

and no fatigue or other factors that may be affected by depth chart settings, evaluations? Or in some cases I've seen it where I've conceived it to be a matter of playing time. But in the latter case it's been a younger scrub in situations PH early in the game and stays at the position, not starting.

[endgame]

Originally Posted by Mike D

Known problem.

Have a link handy? Can't find the words to locate the previous issue through a search. Thanks.

[What!]

Originally Posted by Pete Grassi

Another reason the star system needs to be shut off. Your 4 star player is a career .250 hitter who seems to be in a power slump.

Also, is the team way ahead in the standings? Maybe resting the better player down the stretch.

Not to get too far off subject but is the star system more of how the player is currently playing compared to everyone else at his position?

[Comedian2004]

THE AI must know something you do not, as the other guy is hitting .353 with a slugging percentage of .765.

[Mike D]

Originally Posted by endgame

Have a link handy? Can't find the words to locate the previous issue through a search. Thanks.

http://www.ootpdevelopments.com/boar...d.php?t=100813

Sorry, I thought it was related to this one.

[Giants44]

Originally Posted by Comedian2004

THE AI must know something you do not, as the other guy is hitting .353 with a slugging percentage of .765.

Originally Posted by spleen1015

What are your player evaluation settings? If the game is valuing stats more than ratings, I can see how the scrub could get put in over the other guy.

Wow - are you guys Serious!!! Can I get in an on-line league with you??? (note: since tone is often lost in posts, please understand I am not trying to be an ******* here - I think you guys simply must have not looked very deep into the situation)

Yes Blanco is hitting .356 - IN 17 AT BATS!!!!!!!!!!!

Not to mention the AI chose to start him over Leiper the day he was brought up on 9/1 when he was batting .000 and had a a career BA of .117.

I'm not sure why the AI even brought him up from AAA with his dominant season of .204 in AAA

If the AI is valuing stats over ratings this is an even WORSE example of a bad choice.

The one thing I did notice was Leiper is "tired" - Though usually when a player is "tired" he is still listed as a starter and then rested by the depth chart in game - not placed on the depth chart as a backup.

I am an OOTP fanboy as much as the next guy, but there can be no valid argument made for the AI to choose Blanco over Leiper.

[PSUColonel]

Originally Posted by Giants44

Wow - are you guys Serious!!! Can I get in an on-line league with you??? (note: since tone is often lost in posts, please understand I am not trying to be an ******* here - I think you guys simply must have not looked very deep into the situation)

Yes Blanco is hitting .356 - IN 17 AT BATS!!!!!!!!!!!

Not to mention the AI chose to start him over Leiper the day he was brought up on 9/1 when he was batting .000 and had a a career BA of .117.

I'm not sure why the AI even brought him up from AAA with his dominant season of .204 in AAA

If the AI is valuing stats over ratings this is an even WORSE example of a bad choice.

The one thing I did notice was Leiper is "tired" - Though usually when a player is "tired" he is still listed as a starter and then rested by the depth chart in game - not placed on the depth chart as a backup.

I am an OOTP fanboy as much as the next guy, but there can be no valid argument made for the AI to choose Blanco over Leiper.

I must agree. In another post I disagreeed with SkyDog's reccomendation of having the Ai evaluate ratings at only 65%, I think that is opening a can of worms. In some ways I wonder if 100% ratings evaluation is the road to go. If not, no less than 85%. I'd like to see what people thin. Also- what do y'all think about injury frequency settings? Is normal too few?

[Chappy]

Originally Posted by Giants44

Yes Blanco is hitting .356 - IN 17 AT BATS!!!!!!!!!!!

Not to mention the AI chose to start him over Leiper the day he was brought up on 9/1 when he was batting .000 and had a a career BA of .117.

Yes, but maybe the AI evaluation was set very high on the current year's stats, and the starter was tired when Blanco was brought up on 9/1. He started, went 3-for-3 and then was batting 1.000 on 9/2. So, he started again and went 2-for-4...

I would be curious to see what the settings were, and if he would stay in the lineup if he hit .100 over the next 5 games...

[Dagrims]

Originally Posted by Chappy

Yes, but maybe the AI evaluation was set very high on the current year's stats, and the starter was tired when Blanco was brought up on 9/1. He started, went 3-for-3 and then was batting 1.000 on 9/2. So, he started again and went 2-for-4...

Plus he had a homerun in each of the first two starts. Leiper went 0-8 with five strikeouts in his final two starts before Blanco was called up. Blanco might have been given a spot start on 9/1, and then the AI went with the hot hand.

I do agree that this isn't something that would happen in real life. No way Lieper should be sitting five straight games. Is there any chance Leiper suffered a day-to-day injury, which might cause the AI to start a benchwarmer but still PH Leiper?

It's definitely odd - we're just trying to come up with possible reasons but it doesn't seem right, particularly if you notice this happening on more than one occasion.

[Dan Theman]

Originally Posted by PSUColonel

I must agree. In another post I disagreeed with SkyDog's reccomendation of having the Ai evaluate ratings at only 65%, I think that is opening a can of worms. In some ways I wonder if 100% ratings evaluation is the road to go. If not, no less than 85%. I'd like to see what people thin. Also- what do y'all think about injury frequency settings? Is normal too few?

Not to hijack the thread, but I feel that if you play with scouts then having at least some percentage attributed to stats is a way to account for the level of error in player assessment. If scouts don't exist in your league, then assiging a percentage to stats only accommodates a false sense of realism. If that's important to you, go for it.

[Eckstein 4 Prez]

I've found a couple cases in my solo league where the AI has started far, far worse players over far, far better ones. These aren't close decisions - it's like starting Tony Womack in left over Barry Bonds... that type of thing.

[jbmagic]

i tought this was suppose to be fix in 6.5. with the improve AI?

[endgame]

I'd probably fall short of calling this my opinion, but it's at least my overall first impression after reading and considering the thread and its merits. Some observations:

The decision is a valid argueable point. Why would such a decision be made? The debate and discussion it inspires is not unlike those taking place in the media and audiences of nearly any sport, perhaps more so in gaming environments that involve artificial intelligence.

It is no coincidence that after a relatively grand reception of a new version/addition to a familiar game that time ensues and affords the keener minds the luxury of closer examination and potential shortcomings are brought to light.

It is also no surprise that after the resulting discovery of a few minor glitches the momentum intensifies and even keener investigative eyes bring to light yet another issue such as player ID and retirement along with its unsettling ramifications.

The impetus naturally continues, and rightly so, with fruitful scrutiny of all the various aspects of the game from statistical integrity to the merits of computer-generated decisions such as the one originating this thread. Again a natural, just, and expected course for the inquistive and, by now, slightly skeptical clientle.

The fact that this subject has generated in its life, albeit a relatively short one, 18 responses comes as a bit of relief to me as I find it additionally very odd that this concern about AI lineup decisions arrives on the heels of nearly an unprecedented response to a contribution breathing life into the AI model for management styles. That is, we want on the one hand with seemingly extreme enthusiasm to embrace an AI that will vary its style and tendencies and with the other raise our palm to deter unusual signs of individuality in its form. My point is, or rather wonderment, if it is not an unusual, pervasive, and proven harmful trend, effects AI rosters and lineups and not our own, then why can it not be presumed to entail the similar oddities and argueable moves by "other" managers which, whether AI or not, we cannot control? It is ours to evaluate and critique, but to no greater end than the morning news anchor might question the sudden replacement of last night's starting pitcher without evident cause.

Lastly, not to support these thoughts or unduly infer that the author is any fashion positioned to speak to our particular situation, I am providing an excerpt from Tim McCarver's Baseball for Brain Surgeons and Other Fans to illustrate yet another assessment of causal agents for such occurences:

p.145 with occasional snips - "...Managers juggle their lineups from game to game....often the manager just wants to give one guy a rest or a bench player a chance to start and get some at-bats. ... the manager may move a slumping player toward the front of the order to give him more at-bats, make him feel more involved ... may take the opportunity to insert a young player into the lineup rather than ..waste away on the bench. ..

He speaks to the issue of some shortstops or catchers or those, at least from the manager's point of view who "possess the intriguing combination of talent mixed with sheer will. ... They are winners who are needed in the lineup (that manager's lineup at least for that day and for as long as he deems fit - my notes) even if that means sitting down players with better overall talent."