Possible AI Tweaks for 12
 

Recently restarted a historical league. Previously I would just sim out innings or half innings, but this time I decided to play out every at bat. I'm noticing some things that I missed by simming that might be tweaked in the next version.

1) ERA calculation / RBIs awarded
There was a long discussion about this when 11 came out, so I won't rehash old news. Suffice to say that some sort of 'shadow' half-inning that correctly calculates the stats after the last out would be a big improvement.

2) Runners on third not attempting to score on fly ball outs
I'm seeing quite a few occurances where there is a fly ball out listed as 'deep' in the log and a runner at third does not attempt to tag up and score. It is especially frustrating when the play by play does something like 'a diving catch' or 'runs into the wall and holds on' or similar. I realize that the PbP is probably just random, but in the real world the runner on third could probably walk home on his hands and still score if those events actually happened. Some previous discussion mentioned that there might be a 'fudge factor' so that sacrifice fly numbers are more in line with real world stats. If this is the case, a PbP line of 'the runner stumbles' or 'boneheaded base running' would be in order. My best case fix would be that you get the 'Advance? <yes> <no>' dialog box any time there is a fly ball out with a runner on third and less that two outs.

3) Runners not being aggressive when there are two outs
By this I mean that you often have a runner on base with two outs only advancing one base on a single to the outfield or perhaps two bases on a double. With two outs, runners should be going at the crack of the bat and have a much better chance of taking the extra base because they don't have much to lose by running aggressively. Obviously, this isn't a 100% case but it should happen a lot more than it does.

4) Runners moving on a 3-2 pitch and not taking the extra base on a hit
Maybe it is just a PbP artifact, but I often see a comment about a runner on first 'moving with the pitch' on a 3-2 count and only ending up at second base if the batter gets a single. This should be treated more as a hit and run situation than it is.

Thanks in advance to any devs reading this and/or feedback from others.

Good post. Thank you.

I thought of one more that I wanted to mention but had a brain fade this morning. :p

5) Hit and run is on for the entire at bat
When selecting 'hit and run' in one-pitch mode (and I would assume 'run and hit') it is on for the entire at bat. When I noticed quite a few swinging strikes with runner thrown out I started looking at the logs. The sequence would usually go: 1st pitch is fouled off, 2nd pitch is a pitch-out with a swinging strike and the runner a dead duck. The logic already exists during a sacrifice bunt attempt where a foul ball allows the manager to keep the play on or change his mind. I would like to see the hit and run (and run and hit) sequence go through this same logic. As it stands now, if I'm serious about a hit and run attempt I have to change to pitch-by-pitch mode for that at bat.

Thanks again to the devs for reading this.

Noticed another AI oddity.

I just started another league (in the MMO world I am what is known as an altoholic :p) with the concept of the team being defense, pitching, speed and small ball. What I am seeing is:

6) Trailing runner does not always advance during suicide squeeze bunt play

Say that you have runners on 2nd and 3rd base and you call for a suicide squeeze bunt. About half of the time the runner on 2nd stays at 2nd. I can't logically imagine why, other than missing the sign, the runner wouldn't advance to 3rd base. The logic should be very nearly, if not exactly, the same as if you have runner on 2nd and a batter sacrifice bunts. There might need to be some tweaking on the number of plays at 3rd base, although theoretically a play at 3rd could still happen.

I'm sure there is some statistical underpinning I don't know about, but the implementation doesn't seem very real world.

Thanks again to any dev that takes notice of this.

+1, I see this too in some cases. Would be great if this can get fixed in a future version.