AI leaving pitchers in too long

[slugga27]

Has anyone noticed the AI leaving pitchers in too long? I'm playing an MLB 2013 league, with use relievers at default setting (very high) and I've noticed that the AI has been refusing to pull pitchers until they start getting shelled, instead of managing the bullpen like it should. This leads to skewed offensive stats because it gets to the point where my line-up just tees off on pitchers late in games.

Is there anyone who can shed any insight on this? Unfortunately, I don't have any evidence other than it's a trend I've noticed, but if I can find the time, I will play a month's worth of games or so and chart the results.

[Déjà Bru]

Originally Posted by slugga27

Is there anyone who can shed any insight on this? Unfortunately, I don't have any evidence other than it's a trend I've noticed, but if I can find the time, I will play a month's worth of games or so and chart the results.

That's a key concept. If you have not played even a month's worth of games, you can draw no valid conclusions at this point. Things tend to approach normal (realistic) levels over time in this game, as IRL.

[slugga27]

You're right...and i have played well over a month's worth of games...I just haven't noted exact results.

[SteveP]

If the team strategy slider for hooking SPs is set in the middle, the AI will generally pull an SP when he has given up 5 runs during the game. If the slider is set differently, the run count adjusts accordingly.

If the slider is set in the middle and the run counts are significantly higher (more than 6), something else is going wrong. Figuring out what that might be would be something you would have to do by looking closely at the situation in each game your team plays.

[slugga27]

Originally Posted by SteveP

If the team strategy slider for hooking SPs is set in the middle, the AI will generally pull an SP when he has given up 5 runs during the game. If the slider is set differently, the run count adjusts accordingly.

If the slider is set in the middle and the run counts are significantly higher (more than 6), something else is going wrong. Figuring out what that might be would be something you would have to do by looking closely at the situation in each game your team plays.

That could solve the problem. By default, it's slightly skewed towards a slow hook. Brings me to a question: How do I set it so that the team strategy is universal for all AI-controlled teams? Note: This isn't a problem I have with my own team, because I manage all my games.

[slugga27]

Played out seven games (relatively small sample size, but I wanted to get SOME data out there for now). Here are the results:

#1 H. Kuroda 7.1 IP, 5 ER, 103 pitches Last batter (LB): 2-run HR
#2 C. Sabbathia 2.2 IP, 5 ER, 50 pitches LB: E1, run scores
#3 A. Pettitte 6.2 IP, 3 ER, 108 pitches (Forgot to get LB data; all 3 runs allowed 7th inning).
#4 M. Buehrle 5.1 IP, 5 ER, 78 pitches LB: 2-run HR
#5 B. Lincoln 4.2 IP, 3 ER, 92 pitches LB: BB
#6 R. Dickey 5.1 IP, 4 ER, 106 pitches LB: RBI groundout
#7 M. Gonzalez 5.2 IP, 5 ER, 84 pitches LB: 3-run HR

What I have gathered here is that the AI will typically let a starter start the following inning regardless, and pull him at the first sign of trouble. Lincoln is an outlier--he had walked six in those 4.2 innings. If there is no universal setting I can change, and I hate to use this term, but it's sort of game-breaking...

[SteveP]

Originally Posted by slugga27

If there is no universal setting I can change, and I hate to use this term, but it's sort of game-breaking...

Change to what, exactly? Serious question, not snarky. If you can explain what exactly you would like the AI to do, it might be possible to suggest some tweaks. As it stands, the AI will generally take out an SP when (1) he is shown as "tiring" or (2) when he gives up a certain number of runs (all runs not just earned), or (3) it's the ninth inning and a save situation and the game is set to use closers -- whichever comes first.

Modifying the Hook SPs slider is the best way to get SPs taken out earlier, and that's something to can do for all the teams in your league (if you are commissioner). Put a pitch count limit on your pitchers makes it's possible to have them taken out before they get to "tiring".

I don't believe it is possible to force the AI to take out a guy just because he has pitched himself in a dangerous situation, let's say. It's just asking too much of AI programming to evaluate every such situation. At some point, you have to take control yourself (human thinking beats AI thinking every time ).

[slugga27]

So basically, what you're saying is that the AI won't pull a pitcher based on the situation (say, 0-0 game, no outs, bases loaded--sounds like every team's being managed by Bobby Valentine, if that's the case). I don't recall this ever being an issue in 12 or 13...or maybe it was, but just not as bad and I didn't notice.

Like I said, I don't want to have to go into commissioner mode, because then I'd have to play in commissioner mode for the duration.

I was hoping for realism when it comes to this aspect, as opposed to the pitcher meeting some sort of quota, which appears to be five runs allowed (see my data).

I submitted a ticket to see if tech support can shed some light on this. I'm not trying to be difficult or picky...in my view, something isn't right, here.

[RchW]

None of the scenarios you show are unrealistic. If you were to manage with the intent of taking out a pitcher earlier than the above scenarios to avoid bad things happening you'd fry your bullpen by the end of May.

A sample size of 7 single pitcher lines in a month of 375 games X 2 = 750 single pitcher lines is completely meaningless. You are seeing a pattern where none exists. At least no pattern exists that can be determined by these examples.

YMMV