Three specific suggestions to reduce bad AI trades

[voxpoptart]

I'm a lot less bothered than many OOTP players by the game's trading AI. I understand how hard it is to program, and I'm quite aggressive about using Commissioner Mode to undo awful trades when I see them happening. But I would say, from that experience, that about 80% of the terrible trades wouldn't happen if these three changes could be programmed in.

1) Teams should remember who they have on the DL. Which I seriously don't think they account for, not competently. Too often I see them trade real value, in the offseason, for a player who deserves to be a regular and *as of January* slots into their lineup, but who will obviously be back on the bench, and wasted, by March.

2) Teams should almost never trade prospects for very marginal/ questionable upgrades (except when in the thick of pennant races). In other words, sure, it's nice to have a 1.5 star batter replace one half of your platoon of 1 star players. But real life teams don't normally deal away a couple of 2nd-round draft picks to do it -- especially not if that 1.5 star batter batted .197 last year. Trade those prospects for a 2.5 star player? Maybe. In OOTP I see a lot of multi-prospect trades for the 1.5 star guy.

3) I'm willing to use Force Trade to undo a trade if I hate it enough, but I prefer not to -- if neither side is interested in taking it back, sometimes I'll decide it's a more believable deal than it seems. But more than half the time, when I'm trying to reverse a trade, one or the other GM will look at the same trade they just made, except offered right back to them in reverse, and say "Yes". I'm not kidding: this is common. So as a basic check: if possible, give the computer the ability to automatically test the same trade backwards. And reject it if it would accept the deal reversed.

I hope these ideas make sense and can help.

[MogulFiend13]

This seems pretty good. I also think some sort of option that chances the amount of dealing when the game nears the trade deadline or winter meetings would be nice for real life immersion.

[jaysdailydose]

Originally Posted by MogulFiend13

This seems pretty good. I also think some sort of option that chances the amount of dealing when the game nears the trade deadline or winter meetings would be nice for real life immersion.

They do this for the winter meetings already, for the record.

[voxpoptart]

Until the AI trading is fixed -- and seriously, the three principles I state above would get rid of almost all of the worst of it -- I dread the Winter Meetings and the increased trading. I end up reversing about 80% of the deals made then.

I don't just turn trading off, because sometimes the bad trades inspire me to create good ones. Which is also something I do on my own: jumping into each team's shoes and helping them tinker (or overhaul) can be one of the most fun parts of the game then. But it's tedious going back and erasing history.

A fourth principle, by the way: rebuilding teams should be far, far less eager to sell off their best players for a bunch of middling prospects in Rookie ball. I'm not saying it never happens IRL, but more often a real rebuilding team wants to get *something* back that will help it on a nearer, and more likely, time frame.

[bwburke94]

"Favor prospects" helps a lot with this.

[RchW]

Originally Posted by MogulFiend13

This seems pretty good. I also think some sort of option that chances the amount of dealing when the game nears the trade deadline or winter meetings would be nice for real life immersion.

That already happens.

[voxpoptart]

Burke: No, "Favors Prospects" would make bad teams even *more* likely to deal their best players for a mess of Rookie Ballers. The issue isn't willingness to trade for prospects. It's a trade engine that does too little to tell apart "prospect who is ready, or almost ready, to contribute" from "prospect who is years away and has a good chance of never approaching his Potential at all" -- and does too little to value them differently.

I'm not saying it's an easy issue to fix, by the way. Or that it's the most important issue for the developers to focus on (it's not). But I'd reverse fewer trades if they're ever able to handle that better, and I think other players would be less frustrated as well.

[marc5477]

Originally Posted by voxpoptart

Burke: No, "Favors Prospects" would make bad teams even *more* likely to deal their best players for a mess of Rookie Ballers. The issue isn't willingness to trade for prospects. It's a trade engine that does too little to tell apart "prospect who is ready, or almost ready, to contribute" from "prospect who is years away and has a good chance of never approaching his Potential at all" -- and does too little to value them differently.

I'm not saying it's an easy issue to fix, by the way. Or that it's the most important issue for the developers to focus on (it's not). But I'd reverse fewer trades if they're ever able to handle that better, and I think other players would be less frustrated as well.

This is not my experience at all. With favor prospects they almost never trade away any high potential prospects of any kind at all. I just went through the logs for the last 2 seasons and no prospect with potential over 60/80 was traded by anyone. There are tons of 40-59 prospects traded but most of those will never see the light of day in the majors anyway and those that will are barely average even if they meet potential. The only exception I saw was a good 56/80 pitcher for a couple of iffy 50s prospects but then I noticed the pitcher was fragile... I also saw a star pitcher traded for 4 prospects but they were all iffy anyway so I figured it was fair if the team was trying to trim salaries.

Maybe the problem is your scout? Get a better scout and see if the numbers for prospects make more sense. I have a legendary minors scout with $1.5m to use every year for just minors. You also need to remember than in this game, prospect development is much more realistic than old versions of the game. Most 60/80 (equivalent to 3.5/5 star) prospects barely become average once they are in the majors if they are lucky. Only those above 60 seem to do better but there are so few of those in my league that it works out well. Most teams barely have 3 or 4 players with rating over 60 and there are very few (maybe 1 player per 3 teams) with ratings over 70. I am actually happy that development was toned down a lot... or maybe this is due to my incredible scout?

[Honorable_Pawn]

Originally Posted by bwburke94

"Favor prospects" helps a lot with this.

I disagree.

This setting makes it waaay to easy to fleece the AI of proven veterans for nothing but prospects. As the old saying goes, "a bird in the bush is worth a lot more than a dumb prospect".

To the OP, and to anyone else that is unsure about this setting, keep it at default. Obviously, what Markus optimized as the best setting to use for AI trading is what he set as the default. Don't mess with it unless you have evidence that it needs to be changed. Unless, of course, you prefer a peculiar setting. Setting the AI to favor prospects is peculiar...not evidence-based.

[Honorable_Pawn]

Originally Posted by voxpoptart

I'm a lot less bothered than many OOTP players by the game's trading AI. I understand how hard it is to program, and I'm quite aggressive about using Commissioner Mode to undo awful trades when I see them happening. But I would say, from that experience, that about 80% of the terrible trades wouldn't happen if these three changes could be programmed in.

1) Teams should remember who they have on the DL. Which I seriously don't think they account for, not competently. Too often I see them trade real value, in the offseason, for a player who deserves to be a regular and *as of January* slots into their lineup, but who will obviously be back on the bench, and wasted, by March.

2) Teams should almost never trade prospects for very marginal/ questionable upgrades (except when in the thick of pennant races). In other words, sure, it's nice to have a 1.5 star batter replace one half of your platoon of 1 star players. But real life teams don't normally deal away a couple of 2nd-round draft picks to do it -- especially not if that 1.5 star batter batted .197 last year. Trade those prospects for a 2.5 star player? Maybe. In OOTP I see a lot of multi-prospect trades for the 1.5 star guy.

3) I'm willing to use Force Trade to undo a trade if I hate it enough, but I prefer not to -- if neither side is interested in taking it back, sometimes I'll decide it's a more believable deal than it seems. But more than half the time, when I'm trying to reverse a trade, one or the other GM will look at the same trade they just made, except offered right back to them in reverse, and say "Yes". I'm not kidding: this is common. So as a basic check: if possible, give the computer the ability to automatically test the same trade backwards. And reject it if it would accept the deal reversed.

I hope these ideas make sense and can help.

Love your bullet points.

I think you are right that the AI "forgets" who is on the DL.

[voxpoptart]

Originally Posted by marc5477

With favor prospects they almost never trade away any high potential prospects of any kind at all. I just went through the logs for the last 2 seasons and no prospect with potential over 60/80 was traded by anyone.

Marc: You've just demonstrated my (fourth) point (the one I added after the original post). My complaint was that rebuilding teams in OOTP are too willing to sell good regulars for mid-level Rookie League prospects who are extremely unlikely to provide value; whereas in real life, good regulars are dealt for either elite prospects or for prospects near to contributing in the major leagues.

When Burke suggested "Set the AI to Favor Prospects", I suggested that would make the problem even worse. Which is exactly what you now tell me happens. So, thank you for the support. You seem to think you're arguing with me, but I welcome your evidence that my guess was correct.

(Also, since you bring it up: My scout is "Outstanding" to "Legendary" at all categories, I budget him well, and scouting in general in my game is High accuracy.)

[olivertheorem]

I have no data to back it up, but I do think that rebuilding teams at times go for quantity of prospects over quality to restock a depleted farm system. Besides, if your rebuilding, sometimes you're thinking 3-4 years down the line (Astros, anyone?) not next year.

[marc5477]

Originally Posted by voxpoptart

Marc: You've just demonstrated my (fourth) point (the one I added after the original post). My complaint was that rebuilding teams in OOTP are too willing to sell good regulars for mid-level Rookie League prospects who are extremely unlikely to provide value; whereas in real life, good regulars are dealt for either elite prospects or for prospects near to contributing in the major leagues.

When Burke suggested "Set the AI to Favor Prospects", I suggested that would make the problem even worse. Which is exactly what you now tell me happens. So, thank you for the support. You seem to think you're arguing with me, but I welcome your evidence that my guess was correct.

(Also, since you bring it up: My scout is "Outstanding" to "Legendary" at all categories, I budget him well, and scouting in general in my game is High accuracy.)

Uh no because this is only true if YOU exploit it. Your original complaint implied that the AI trades badly between itself. This required you to undo their trades. Ive never seen it happen between AI trades with favor prospects setting.

Ive never seen a superstar traded for a bunch of prospects that are not proven. I have seen them trade a superstar for younger stars plus 2-3 other prospects and that was perfectly fine in my opinion. Ive also seen them trade aging star player for draft picks and prospects which is also ok. I have also seen them trade star players going into free agency for prospects and that is also fine because they would have lost that player anyway. I have also seen variation in trades based on performance over ratings. For instance, i see this all the time:

45/80 Actual: 48/80 Pot .325/.375/.484 (29 y/o old CF with cheap contract)

for

59/80 Actual: 61/80 Pot .260/.360/.415 (31 y/o 1B with 2x the contract of above)
+ 52/80 pot 20 y/o prospect in rookie league

On the surface this seems like a bad deal but look at those stats... the 1st guy has been hitting above or near .300 for 4 seasons! in that time he collected a WAR of about 12. The better rated player has barely hit .260 his entire career and barely earned a WAR of 8 the last 4 seasons on top of having a worse contract... but his ratings are 11/14/13/12/9 while the other guy is only 12/12/6/9/8...

So upon review, I let this one go. The prospect is ok but as we all know the odds of a 20 year old meeting their potential (at default development settings).

What have you seen that was very bad on your settings?

[voxpoptart]

Marc: I've never seen a superstar, exactly, traded for mid-quality Rookie League prospects, but very good players in their prime -- and there is no free agency in my league, so it's not about contracts. Elijah Skuse, best 2B in the Western League in year one, age 25, was traded for five prospects in Rookie League that winter, none of them with potential over 2.5 stars; I allowed that one, since five is a lot of players. Largo Novara, 2nd-best CF in the Eastern League, age 28, was traded that same winter for four modest Rookie League prospects; I'm out of town and can't check, but I think I reversed that deal and reworked it as a more balanced one. (Maybe I let it go through.) I mention the modest star ratings, but their stats in Rookie ball weren't good either.

I've seen, and mostly reversed/ reworked, enough variants since then that I no longer recall most of them. The quantity of prospects can be two or three if the major leaguer is a step below Skuse or Novara. I play for pleasure, not to document my aggravations, even if it puts me at "please trust that I'm reporting as truly as I can".

But! Sometimes they're bad the opposite way: Hee-cheong Ogum, a reliever who was 28 in year one and has never posted an ERA under 5 or stayed a full season in the majors, keeps being traded for two low-minors prospects who are actually interesting; I think I've undone versions of that three separate times. That's what I posted in the original points on this thread: teams shouldn't (usually) trade real prospects for tiny upgrades (except maybe in the thick of a pennant race). Even a small chance of reaching excellence beats the heck out of Hee-cheong Ogum. But they also shouldn't normally trade good *players* for tiny *potential*.

[The Game]

Originally Posted by Honorable_Pawn

Love your bullet points.

I think you are right that the AI "forgets" who is on the DL.

In my EBL league the Glenbard Cougars, basically the best team other then mine in the NL has CF terry Lee on the DL til mid June. A 9 defense guy who hit great for GB after being traded there last season. .270 10 hr 40 RBI approx, although that might be his last season totals. GB then traded for CF/DH Brien Keywell who is a lifetime .318 300 HR approx guy in teh last year of his deal and a team going nowhere. Keywell who is a switch hitter while Lee is a rightie can no longer play CF. Mike Hughes who is leading the league in AVG, HR and 2nfd in RBI in Glenbard's DH. So the Cougars will either use Keywell in CF and suffer through the 10 E he had last year or use him at DH and put Hughes at 3B who he has no rating for anymore. (100% scouting) While Keywell can still hit, the trade made no sense. Normally i would have undone the trade but i have relaxed my stance on my league.