OOTP10 Suggestions: Artificial Intelligence

[lt2cents]

Suggestion--

Greatly improve AI GM player moves throughout the minors. Make the minor league rosters 95% + more stable.


Reason--

50 players are not reassigned every couple days in RL. If you are trying to play a manager career in the minors of a major league organization it is crazy to see these players flying around and your rotation and lineup constantly whacked. Page after page of silly minor league transactions is disgusting.

This improvement should be bundled with the idea of the minor league non-playing list which would help to implement it.


Priority-- High.

This should be very high priority-- not only is the current situation in the game extremely bad modeling, it is singularly the most aggravating issue in trying to take up your suggestion to try managing in the minor leagues.

[Killebrew]

Suggestion: Improve the Depth Chart AI Decisions. All of the following issues are not rare cases - they are the default for the game and very easy to replicate.
1. Do not allow the AI to place the same player as the #1 substitute for every position while having his playing time set to "every 2 games" at every position (in other words, an impossible AI request that the game has to ignore when simming).
2. Do not allow the AI to have a 1st position substitute AND the 2nd positional backup both listed as subbing "every 2 games" (I assume the 2nd sub is just ignored in those cases but who knows).
3. No more defensive subs for the DH position (no impact for the user, just makes the game AI look dumb).
4. Great defensive players that are not in the lineup are usually (or at least frequently) skipped in the defensive replacement column.
5. The AI appears to ignore the batters handedness when doling out out substitute playing time versus LHP and RHP. If there is lefty-righty logic here it definitely needs to be weighted more heavily to encourage smarter playing times splits.
6. [Related issue] All of the CTD reports I get (including after the OOTP9 9.2.11 release) occur when the user is working in the pitching line up screen or the lineups/depth chart screens, either with dragging/dropping or selecting a position from the drop down box beside a player. It seems highly likely there is less AI logic at work here now than in past versions of OOTP, yet since OOTP 2007 the depth charts (& pitching setup) screens have become very unstable.

Reason: The current AI depth chart selection choices are not strong and this is a step backwards, since the OOTP5 depth chart AI was better at auto-filling these spots. If OOTP fans want to sell a new/improved version of the game it can't follow the 1 step forward, 2 steps back method when it comes to existing game features.

Priority: Really only important for any solo players of the game.

[koohead]

Suggestion

- something needs to be added to the AI so that a team under AI control will have at least have a reasonable number of players at each position. Meaning no team should have less than 5 people who would be considered SPs and probably should never have less than 10 pitchers, have 2 catchers, and then at least 1 player who can play each IF position and at least 4 OFs.

Reason

- Just something to make sure a team doesn't end up with 20 MR, 1 catcher, 3 IF, and 6 OFs.


Priority

-Medium.

[koohead]

Suggestion

Have the ability to set the AI to evaluate prospects differently than players with 3+ years of statistics, or in other words, two sets of evaluation settings. One set could be used for players with 2 or 3+ years of major league stats to look at, while another set would be used to evaluate prospects who do not have alot of years of stats (where ratings would come into play more)

Reason

This is more along the lines of how real people would evaluate players. With major leaguers, most would look at stats and actual performance more than their "potential" since after 3 years they should have enough ABs or IPs to get a gauge on whether or not they'll live up to that potential. For prospects and younger guys, we would look at what their potential is and what their ceiling is vs how they are actually performing in the minors. Sure, stats still matter in the minors, but maybe not as much as the 3 years worth in the majors. And along the same thought, potential ratings may matter for a 30 year old, but not as much as the past 3 years of stats. At least that's the way I think...and it would be nice to be able to have the AI think this way too.


Priority

-Medium.

[ovccsteve]

Suggestion: Stop giving out extensions to "Overpriced Veterans" and unused old guys.


Reason: One day, I'm doing trade research and for every player I shop, a team offers me the same old, overpriced player. Few weeks later, I'm looking at that team and they've signed an extension with the same guy they were so desperate to get rid of. Also, I'm finding guys spending four years in their mid-40's getting about fifty AB's a season and a string of one-year ~$500K extensions.


Priority: High. A 54 AB season at age 42 is the end of a career, not the start of a "new phase".

[lt2cents]

Suggestion:

Provide options to allow Spot Starter use to be more realistic.


Reason:

Provide for inserting the Spot Starter to make a start when the rotation is tired and needs to slip back a day rather than replacing one starter without solving the problem. Automatically slide the rotation back--- and coordinate correctly with the 7 day lineup page (covered in another suggestion also). There could also be additional rules that could be used for setting Spot Starter usage, such as when actually skipping a rotation starter is wanted.


Priority:

Medium. Especially when playing a schedule with doubleheaders, where it is particularly useful, but even otherwise, this option is needed.

[koohead]

Suggestion:

Have some kind of Spring Training AI settings so we can tell the AI how to manage playing time during ST games. Maybe something like "Favor prospects to give the youngsters more playing time or Even to evenly distribute playing time for vets and prospects. have separate settings for batters/pitchers.

Reason:

Its tough (if not impossible) to program AI to handle playing time the way it happens in real life, but the way it does it right now is too rigid. From what I gather the starter will get 3-4 innings, and then the player listed as the 1st backup gets the playing time. Players not listed on the depth chart rarely get playing time. I honestly couldn't tell you how pitchers are handled other than the SP will be limited innings.

Priority:

Would only be a medium if Spring Training in general is overhauled by allowing players not on the 40 man roster to attend. Otherwise it would be a low.

[aewin]

Suggestion

Improve how AI deal with waivers on minors players who still have option years left. I find it frustrating that playing solo I can build up a great minors organisation in a couple of months just claiming off the waiver wire. The AI should better protect some of these high potential minor leaguers who still have option years left.

Reason

Realism - makes solo play too easy.

Priority

High for solo play. Doesn't matter for online leagues.

[OldFatGuy]

Suggestion

Add an option in the Strategy section to set a minimum number of pitcher's on the AI's roster.

Reason

The AI sometimes doesn't do a good job of building it's rosters for the era being played, and it's especially noticeable when strategy settings change.

Priority

Medium. If better overall roster management could be developed, this would be unnecessary, but in the interim I think it might help the AI be a little more competitive, or at least seem so to us, the player, if we could tell it "Hey dummy, pitcher endurance is normal now, you need more than three relief pitchers!"

[Curtis]

SUGGESTION: Try to get the AI manager to adapt to how the player manager manages.

REASON: I get annoyed when my 'designated bunter' comes to bat for the fifth time in a game (twelfth in the series), and everyone in the stadium knows this guy is going to bunt… except the AI. My DB always gets one free shot before the AI reacts defensively. Likewise, I'm now following {someone's} advice about 'taking' a lot of pitches to run up the AI pitcher's count and tire him out. It seems to be working, but it shouldn't. After the first five batters the AI should take a trip to the mound and mutter, "Throw strikes", while roughing up the ball. There are other examples. If the AI sees that I usually try to steal second or hit-and-run in a particular situation, it should react accordingly.

PRIORITY: Low, mainly because I suspect the amount of programming effort this would require would be prohibitive. But, it would go a long way toward swinging play balance toward the AI.