Trade Issues

[SirMichaelJordan]

Originally Posted by Orcin

I will try a historical league with the settings at Very Hard and Heavily Favor Prospects. What about using AI evaluation vs. pure ratings... which setting makes the AI more competitive? If AI evaluation, how do you set the ratios? I would guess that AI evaluation with a heavy emphasis on ratings would yield the best results, but that's just a guess. I have tried a few sample trades and it doesn't seem to make much difference.

My AI Evaluation is

0% Ratings
50% Current Year
25% Last Year
25% 2 Year's Ago

I used to have 25% ratings but It seem with the importance of defensive ratings this year, 25% ratings were too strong for my liking.

I've never played a historical save so I don't know if you'll get similar results.

[Honorable_Pawn]

Originally Posted by harrumph

it's not like taking dubious five-for-x trades is the only thing the ai does wrong in trading. It's not an isolated bug, it's a symptom of the ai not doing any long-term planning. The ai just plugs the players it has into the places it thinks are best for them right now; in trades, it simply assigns each player a value and does some adding and subtracting. What it should do is create goals related to players.

bam!!

[Honorable_Pawn]

Originally Posted by harrumph

I appreciate that programming this sort of thing fully must be incredibly hard, but it seems like there are a few simple first steps that would make a world of difference. For one thing, the AI should just put its best young players off-limits sometimes. Completely off-limits, no exceptions. It should make a serious effort to sign them to long deals, clearing veteran payroll to do so if necessary, and it should just not trade them.

Double Bam!! Bam! Bam! In yo face, Disagreeers!!

[Orcin]

Originally Posted by SirMichaelJordan

My AI Evaluation is

0% Ratings
50% Current Year
25% Last Year
25% 2 Year's Ago

I used to have 25% ratings but It seem with the importance of defensive ratings this year, 25% ratings were too strong for my liking.

I've never played a historical save so I don't know if you'll get similar results.

It wouldn't work well for a historical league with reserve rosters, because the prospects have no stats. For stats only with minors, it would be great. I guess there is no one right answer to that.

[SirMichaelJordan]

Originally Posted by Orcin

It wouldn't work well for a historical league with reserve rosters, because the prospects have no stats. For stats only with minors, it would be great. I guess there is no one right answer to that.

Markus mentioned that if a player doesn't have any stats (or not enough stats) then ratings are used by default.

[SirMichaelJordan]

Originally Posted by Honorable_Pawn

Double Bam!! Bam! Bam! In yo face, Disagreeers!!

I feel like it does place priority on young prospects if you set the trading preference to favor them. In my case I have them heavily favored.

Also these are the type of 5 for 1 trades that involved min salary throw aways that am able to pull off on very hard/heavily favored trade settings.

This was from a test league save and this deal was the best offer and it was for an injury prone 32 year old on a long term cheap contract (5 years 10mil)





[












Obviously I wouldn't make that trade as the 5 players are still young enough (26-28 years old) to be Important AAA players and potentially make an impact of the main roster even though they have underachieved most of their playing career and failed to impress during their call ups.

[harrumph]

Originally Posted by SirMichaelJordan

I do play with high injuries so I know that adds to the difficulty level. (And TCR @ 200)

Oh and I should mention trading preference is set to heavily favor prospects which is probably why I can't offer ton of min salary vets for a top prospect or a young all star.

I don't find it easy at all to steal away a young player from a team even if the player I am giving them is an established all star. I always get in reply "You are kidding right?"

So I guess having trading preference set to neutral treats everyone the same if they are close in ratings or if you give them a few players that are rated high enough. Neutral is probably why it seem as if the AI is not planning for the future because it doesn't have much of a preference.

Beside that one thing I would like to see is small market teams signing their young prospects to long term deals as early as possible. I would like to see the AI buy out the arb years.

Well, this is all good food for thought. I'll try switching trades to favor prospects, and I might go to high injuries too. I don't mean to brag about it being easy, by the way—I'd like it to be more challenging, and I suspect that various settings I've chosen are making my league easier than other people's. There are only three minor league levels, for instance, which makes it easier to stack up a bunch of really good coaching staff. Maybe that's helping my player development? If so, the AI isn't getting any help that way; they employ a real mixed bag of nuts as coaches.

Anyway, it really doesn't change the fact that the AI just doesn't make effective use of its most valuable players. Even if you made it impossible for human players to get these guys, they'll still give up their most important young players in bad trades with each other, let them go to free agency, etc. I think we all agree that teams (and I'd say not just small-market teams) should be buying out arbitration years of guys who are clearly hot properties, if nothing else.

Out of curiosity, how do you find the difficulty of signing extensions in your games? This is the other big factor that I think makes my league too easy—not only does the AI rarely sign guys to smart extensions, but I get a lot of sweetheart deals. I find that everything is basically one step more favorable than it should be. The situation I'd expect, roughly: old guys, fringe guys, and guys who just had a really bad year and are looking to rebound should be willing to sign cheap one-year extensions; high-loyalty, low-greed guys who are happy with their place in the organization should be looking for three- or four-year extensions at sub-market rates; greedy and unhappy guys should be asking for outrageous amounts (insisting on free agency, essentially); and the rest, the average guys, should be asking for the kind of deals they'd get in free agency.

Instead, I find that I can sign almost anybody in his twenties to a team-friendly contract that runs into his early thirties and then sign a series of one-year deals for more or less favorable rates after that. Sometimes I get guys who insist, into their thirties, on getting two- or three-year deals; very rarely, I get guys who actually ask for rates close to what they'd get on the open market. Literally once in 46 seasons have I had a guy ask for something totally unreasonable (a starter making around $11m/year who asked for $95m over three years).

This is weird, right? Any idea what's causing it?

[PSUColonel]

Originally Posted by RchW

Hard neutral.

Is this recommendation being made after the bump in difficulty of the average setting?

[Orcin]

I have finished some testing with Very Hard & Heavily Favor Prospects vs. Hard/Neutral in historical leagues. I think VH/HFP is really useful if you want to restrict the human's ability to make ordinary trades. However, it does not prevent exploits.

The AI will just fall in love with one of your players and offer anything, and the difficulty setting does not effect the AI's preference. This doesn't happen often, but it happens enough that you can totally exploit the AI over the course of a season on any difficulty setting. The AI will accept a ridiculous offer regardless of setting in these cases, but VH may tone the offer down. For example, on Hard the AI will give $50 for your $1, but on VH it will only give $25 for your $1. It is still a huge disparity in value.

Conversely, the AI will begin to hate one of its players for no apparent reason, and trade that player for a bag of balls. This is harder for the human to exploit, because you have to stumble on to this hatred before the AI dumps the player to another AI team. There are two reliable ways to find the hated player: (1) shop a good player from your team and see who the AI offers up, and (2) wait for the Ai to offer you a trade and make a counter with a really poor offer frokm your side (the AI will often take it).

In conclusion, the only reliable way to have a competitive game against the AI in a historical league is with house rules. Therefore, you might as well leave the setting at average/neutral because you will have to police yourself anyway.

By the way, not using the "shop player" function is a house rule. Avoiding anything that the game allows, such as subbing players in and out of a deal until you find the best combo of scrubs or piling 5 players on until you get a yes, is a house rule.

[PSUColonel]

Originally Posted by Orcin

Conversely, the AI will begin to hate one of its players for no apparent reason, and trade that player for a bag of balls. This is harder for the human to exploit, because you have to stumble on to this hatred before the AI dumps the player to another AI team. There are two reliable ways to find the hated player: (1) shop a good player from your team and see who the AI offers up, and (2) wait for the Ai to offer you a trade and make a counter with a really poor offer frokm your side (the AI will often take it).

What about the trading block?

[PSUColonel]

I think you can use "shop player" (even though I am not a fan of the feature) on "Hard" as the AI will only ask for your very top prospects. Usually this is a loss for the human. I do not play historical leagues, but I was not aware there is a difference when it comes to the trading AI...If there is, please correct me. To me, (after trying to game the system multiple times...in other words testing) Hard is still a good setting. Average is better than it was, but is still a bit too easy IMO. Hard is probably a little too hard, but it's very tough to really fleece the AI. Even on neutral it wants young prospects more so than veterans, so that part (preference) seems OK. I just feel there should still be a setting between average and hard. If trading difficulty were done on a 1-10 scale (as opposed to 1-5 which it currently is) I'd like a 6.5

5 would equal average, with 7 being Hard.

[Orcin]

Originally Posted by PSUColonel

What about the trading block?

Yes, that's another good way (check who the AI puts on the block). The AI doesn't seem to use the trading block in historical leagues though. At least, I don't see it.

[Orcin]

Originally Posted by PSUColonel

I do not play historical leagues, but I was not aware there is a difference when it comes to the trading AI...If there is, please correct me.

There is a difference in the human's knowledge of the players in question, which provides a greater ability to exploit the AI.

[rpriske]

Originally Posted by PSUColonel

I think you can use "shop player" (even though I am not a fan of the feature)

Out of interest... why are you not fan?

I find it is pretty much the only way to out a trade together that is not incredibly tedious.

[RchW]

Originally Posted by rpriske

Out of interest... why are you not fan?

I find it is pretty much the only way to out a trade together that is not incredibly tedious.

Agreed. To me shop player is like making phone calls to gauge interest. I almost never make a trade from shop player.

[RchW]

Originally Posted by PSUColonel

Is this recommendation being made after the bump in difficulty of the average setting?

Not a recommendation. It's what I use currently.

[PSUColonel]

Originally Posted by rpriske

Out of interest... why are you not fan?

I find it is pretty much the only way to out a trade together that is not incredibly tedious.

I guess I should re-state that. I am not a fan in its current form. I would rather we had a full fledged working trade block...and perhaps in stead of a "make this work now" I'd rather see the AI submit a list of players it may be interested in. Also the AI should be using the "untouchable" feature. Basically all of this was originally set up for online leages, but I don't see any reason why these features shouldn't be used by the AI in leiu of shop player and make this trade work now.

So what I am saying is, putting a player on the block (human or AI) should be the act of shopping a player. It is them up to the other teams to offer deals via email.

[rpriske]

Yeah, I like 'Shop' a lot, but that is partly because the 'Trading Block' is pretty much useless. If it wasn't, then maybe my opinion would change.