How to fix trading? - Page 2

eauhomme · 08-17-2024, 03:16 PM

One area that may be worth starting from would be utilizing a system similar to the Baseball Trade Values website (www.baseballtradevalues.com). Their system, while not perfect, throws up some fairly reasonable values on the ultimate value of each player, with a simulator that allows one to rate trades. Long-term high contracts often have negative values, and many prospects have lower values because of their inherent risks, while untouchable players generally have extremely high values.

Utilizing that kind of a system combined with AI figuring the actual needs of the team (don't need four first basemen at the major league level even though they are high-value, don't want to trade the one decent starter for a bat) could make the trade AI work much better.

PSUColonel · 08-20-2024, 03:24 AM

I would also argue that the time of year (Long with many other variables) could also be a factor. Is it early in the season for instance?

dawg1isclear · 08-20-2024, 03:50 AM

Quote:

Originally Posted by eauhomme

One area that may be worth starting from would be utilizing a system similar to the Baseball Trade Values website (www.baseballtradevalues.com). Their system, while not perfect, throws up some fairly reasonable values on the ultimate value of each player, with a simulator that allows one to rate trades. Long-term high contracts often have negative values, and many prospects have lower values because of their inherent risks, while untouchable players generally have extremely high values.

Utilizing that kind of a system combined with AI figuring the actual needs of the team (don't need four first basemen at the major league level even though they are high-value, don't want to trade the one decent starter for a bat) could make the trade AI work much better.

and rotate 180 degrees

OutS|der · 08-21-2024, 01:05 AM

One thing to remember is even with scouting at 100%, displayed ratings are only part of how good the player is. The true ratings in the editor can shed a lot more light on how valuable the player truly is.

Still the AI needs a lot of work when it comes to roster building.

Bill Veeck · 08-22-2024, 11:48 AM

This thread I think shows the level of difficulty involved in trying to program AI in such a way as to replicate the complicated thinking actual GM's must engage in. There are so many factors involved that for the life of me I don't know how you program AI GMs without a lot of sliders -- then each GM is having different circumstances as you progress. Just saying real GMs have so much to do to rate acquisitions and DFAs/waivers/40-man roster constraints...

This is really an area where ootp has already done so much work with a limited staff... It does drive you crazy to see a prospect DFAd. Any weakness in logic is amplified in an ootp season because the immersion is deep up until a really bad move occurs

CanadaCajanek21 · 08-24-2024, 02:52 AM

Quote:

Originally Posted by mytreds

Great, thanks for the info

Next I have to ask what your trade difficulty settings are.

Sorry! Covid got me down for a couple weeks lol. I've manually changed from normal (slider not touched, default I assume) to 3 ticks on the "easier" end. Still the same trades.

CanadaCajanek21 · 08-24-2024, 02:52 AM

Quote:

Originally Posted by PSUColonel

I would also argue that the time of year (Long with many other variables) could also be a factor. Is it early in the season for instance?

Early July. So getting close to the deadline!

Pelican · 08-24-2024, 09:50 AM

Just speculating, but I wonder if the AI trade algorithm has a filter for trades that look "too good to be true", or maybe saving you from yourself, in giving up all those stars. As a human GM, receiving that one-sided offer, I would be suspicious.

As to prospects, agree with those who support restraint on the part of the AI in dealing them away. You have no way of knowing how good they will be. In Philly this is known as the "Ryne Sandburg" problem. Formerly known as the "Ferguson Jenkins Problem". It seemed such a steal to get Ivan Dejesus for an over-the-hill Larry Bowa and a AAA third baseman (at that time). Not so much. HOF. Same with Fergie (also HOF). Hold onto prospects. Forget the "bird in the hand" approach.

Also, may have been mentioned, but can't the trade slider be moved to have the AI GM's favor veterans over prospects, or the reverse? I usually do that, so that I (GM of historically bad teams) can trade vets for prospects to start the long road back to respectability. It works. The AI GM will sometimes trade prospects for vets.

PSUColonel · 08-24-2024, 01:14 PM

Quote:

Originally Posted by Sweed

I'm happy to hear Matt's answer. AI teams should value some prospects to the point they simply will not trade them, this is reality. Nothing they do is ever going to be perfect or make everyone happy, but IMHO, what they have done with the trading module has greatly improved the game. We used to have endless threads every year about "OOTP trading sucks. How come I can trade a bucket of balls for a superstar?". Don't see many of those threads now and that is a good thing.

Agreed 100 percent.

OutS|der · 08-24-2024, 02:20 PM

I hope one day we'll have a setting to approve trades or not

CanadaCajanek21 · 08-28-2024, 12:59 AM

Quote:

Originally Posted by OutS|der

I hope one day we'll have a setting to approve trades or not

Feels slimy to use "force trade".

Just want to have a calm talk with the GM's I want to trade with lol.

PSUColonel · 08-30-2024, 11:26 PM

The only thing I will say is that if TCR is now much more volatile (and it sounds it is) then perhaps the default trading AI should start valuing established MLB players (veterans) a bit more now than in the past.

In the past it absolutely made sense to make the AI very protective of prospects since they were not flaming out the way a lot of real life players' actually do. IRL prospects are an absolute crapshoot (within reason...1st rounds picks usually will make it to the show...how good they might be or how long they last is another story) and if the game is going to reflect this now (which isn't a bad thing for realism IMO) then perhaps AI trading also needs to value the old "a bird in the hand" philosophy a bit more.

EDIT: in reading the original post, perhaps that's what is going on here as we speak? Prospects are just that prospects. It really comes more down to an organizations' situation (re-building, making a pennant run etc) that determines how they trade IRL. A team who is ready to make a run will value players that can contribute now much more than a club who is looking to the future. Team needs are also a factor as well as finances.

PSUColonel · 08-30-2024, 11:31 PM

Quote:

Originally Posted by CanadaCajanek21

Feels slimy to use "force trade".

Just want to have a calm talk with the GM's I want to trade with lol.

I get what you are saying, but in the quest for a better AI this is just not really as solution, at least not one that all too people want to settle on.

pgjocki · 08-31-2024, 02:04 PM

Quote:

Originally Posted by jets12

The biggest problem is that the AI in this game is weak in many areas and the trade AI is one of those areas. You can't expect the game AI to "act" like real life, because if you do you will be disappointed. After all it is just a game.

I would hope that in the future the AI will be upgraded, but I wouldn't hold my breath. My opinion of the 25 edition of the game is not good. So many issues that take away from the overall experience. I play this version the least of any version I have bought in the past.

If you've paid attention to the absolute bonehead moves by the Cardinals GM over the last few weeks you'd understand why it is impossible for the AI to act life real life because you'd throw your laptop threw a window if it made the same moves as Mo...

CanadaCajanek21 · 09-01-2024, 03:06 AM

Quote:

Originally Posted by pgjocki

If you've paid attention to the absolute bonehead moves by the Cardinals GM over the last few weeks you'd understand why it is impossible for the AI to act life real life because you'd throw your laptop threw a window if it made the same moves as Mo...

Well I made a new trade offer to Boston, and not even Mo would turn it down. For Kyle Teel, a top catching prospect I offered:

2 5 star prospects, 1 an OF and 1 a SP
3 4 star prospects
2 3.5 star prospects
3 3 star prospects

Guess catcher is really that rare that 10 players for 1 isn't a good trade for a team.

This game LMAO

pgjocki · 09-01-2024, 06:26 PM

Now you have me curious about your scouting level. It appears you are putting a lot of value on your prospects and without knowing how the Catcher is rated not only by OSA but by their scout and how he slots into their plans it's impossible to determine.

Think about Baltimore two years ago if someone had come to them with a set of prospects for Rutschman; they'd of said no thanks. As a GM I'm not giving up a potential Hall of Fame catcher for prospects because it's almost always better to keep what you have than take a chance on complete unknowns.

Even Warren Buffet is famous for saying that one in the hand is better than two in the bush...same thing applies to known MLB proven players over prospects.

OutS|der · 09-01-2024, 10:01 PM

Quote:

Originally Posted by CanadaCajanek21

Well I made a new trade offer to Boston, and not even Mo would turn it down. For Kyle Teel, a top catching prospect I offered:

2 5 star prospects, 1 an OF and 1 a SP
3 4 star prospects
2 3.5 star prospects
3 3 star prospects

Guess catcher is really that rare that 10 players for 1 isn't a good trade for a team.

This game LMAO

I don't understand what you're offering. You give stars for players as if they mean the same thing to you and me.
I don't see anywhere you mentions you have scouting off or accuracy set to 100% so all this mention of stars is pointless.

Stars only represent the talent in your league, not how talented the player is overall. So they are to be a guide not the end all factor in determining trades.

Unless you have scouting off or accuracy set to 100% then that will play a factor in how the AI rates your player, same with AI evaulation settings.

jcard · 09-01-2024, 11:22 PM

Quote:

Originally Posted by OutS|der

I don't understand what you're offering. You give stars for players as if they mean the same thing to you and me.
I don't see anywhere you mentions you have scouting off or accuracy set to 100% so all this mention of stars is pointless.

Stars only represent the talent in your league, not how talented the player is overall. So they are to be a guide not the end all factor in determining trades.

Unless you have scouting off or accuracy set to 100% then that will play a factor in how the AI rates your player, same with AI evaulation settings.

His scout sees Teel as a top prospect, making it extremely unlikely Boston could have a significantly higher opinion on him. It follows then that the scouting disconnect would have to be between the two organization’s assessment of the package he is offering. Given the exaggeratedly one-sided nature of the trade from his scout’s perspective, there would have to be a significant difference on the value of that package of player—by the same two organizations that agreed on Teel’s upside.

I’m sure there is some combination of factors that whose felicitous calculus could explain the rejection, but I’m placing long odds on that parlay.

kidd_05_u2 · 09-02-2024, 01:20 PM

Quote:

Originally Posted by CanadaCajanek21

Guess catcher is really that rare that 10 players for 1 isn't a good trade for a team.

This game LMAO

I wonder how you would react if the AI offered ten randoms for your top player or top prospect. These complaints are usually about a human player trying to throw a bunch of stuff together to get the most valuable player in the trade.

Also, how often do we see a killer prospect close to the MLB traded for a bunch of other prospects?

OutS|der · 09-02-2024, 07:22 PM

From 2009 to 2018, a span of ten drafts, 47 percent of the position players selected by all MLB teams in the June draft ended up with a negative Wins Above Replacement (at best) in the majors. One in five position players picked in the first round (2009-18) have played fewer than 50 games in the big leagues.

Selecting pitchers has been almost as much of a crapshoot: in the ten June MLB Drafts from 2009 to 2018, 39 percent of pitching prospects either did not pitch in the big leagues or have produced negative WAR in the majors. A stunning 26 percent of pitchers picked in the first round have never pitched in the majors.

I wouldn't trade 10 unknowns for a known.

08-24-2024, 09:50 AM	#28
Pelican Hall Of Famer Join Date: Mar 2021 Location: Wilmington, Delaware Posts: 2,891	Just speculating, but I wonder if the AI trade algorithm has a filter for trades that look "too good to be true", or maybe saving you from yourself, in giving up all those stars. As a human GM, receiving that one-sided offer, I would be suspicious. As to prospects, agree with those who support restraint on the part of the AI in dealing them away. You have no way of knowing how good they will be. In Philly this is known as the "Ryne Sandburg" problem. Formerly known as the "Ferguson Jenkins Problem". It seemed such a steal to get Ivan Dejesus for an over-the-hill Larry Bowa and a AAA third baseman (at that time). Not so much. HOF. Same with Fergie (also HOF). Hold onto prospects. Forget the "bird in the hand" approach. Also, may have been mentioned, but can't the trade slider be moved to have the AI GM's favor veterans over prospects, or the reverse? I usually do that, so that I (GM of historically bad teams) can trade vets for prospects to start the long road back to respectability. It works. The AI GM will sometimes trade prospects for vets. __________________ Pelican OOTP 2020-? ”Hard to believe, Harry.” Last edited by Pelican; 08-24-2024 at 09:51 AM.

08-30-2024, 11:26 PM	#32
PSUColonel Hall Of Famer Join Date: Feb 2002 Posts: 13,094	The only thing I will say is that if TCR is now much more volatile (and it sounds it is) then perhaps the default trading AI should start valuing established MLB players (veterans) a bit more now than in the past. In the past it absolutely made sense to make the AI very protective of prospects since they were not flaming out the way a lot of real life players' actually do. IRL prospects are an absolute crapshoot (within reason...1st rounds picks usually will make it to the show...how good they might be or how long they last is another story) and if the game is going to reflect this now (which isn't a bad thing for realism IMO) then perhaps AI trading also needs to value the old "a bird in the hand" philosophy a bit more. EDIT: in reading the original post, perhaps that's what is going on here as we speak? Prospects are just that prospects. It really comes more down to an organizations' situation (re-building, making a pennant run etc) that determines how they trade IRL. A team who is ready to make a run will value players that can contribute now much more than a club who is looking to the future. Team needs are also a factor as well as finances. Last edited by PSUColonel; 08-30-2024 at 11:30 PM.

09-01-2024, 06:26 PM	#36
pgjocki All Star Starter Join Date: Jun 2013 Location: Maryland - just outside DC Posts: 1,585	Now you have me curious about your scouting level. It appears you are putting a lot of value on your prospects and without knowing how the Catcher is rated not only by OSA but by their scout and how he slots into their plans it's impossible to determine. Think about Baltimore two years ago if someone had come to them with a set of prospects for Rutschman; they'd of said no thanks. As a GM I'm not giving up a potential Hall of Fame catcher for prospects because it's almost always better to keep what you have than take a chance on complete unknowns. Even Warren Buffet is famous for saying that one in the hand is better than two in the bush...same thing applies to known MLB proven players over prospects. __________________ - - - World Series championships: 1926, 1931, 1934, 1942, 1944, 1946, 1964, 1967, 1982, 2006, 2011

08-17-2024, 03:16 PM	#21
eauhomme Minors (Double A) Join Date: Jan 2022 Posts: 120	One area that may be worth starting from would be utilizing a system similar to the Baseball Trade Values website (www.baseballtradevalues.com). Their system, while not perfect, throws up some fairly reasonable values on the ultimate value of each player, with a simulator that allows one to rate trades. Long-term high contracts often have negative values, and many prospects have lower values because of their inherent risks, while untouchable players generally have extremely high values. Utilizing that kind of a system combined with AI figuring the actual needs of the team (don't need four first basemen at the major league level even though they are high-value, don't want to trade the one decent starter for a bat) could make the trade AI work much better.

08-20-2024, 03:24 AM	#22
PSUColonel Hall Of Famer Join Date: Feb 2002 Posts: 13,094	I would also argue that the time of year (Long with many other variables) could also be a factor. Is it early in the season for instance?

08-21-2024, 01:05 AM	#24
OutS\|der Hall Of Famer Join Date: Apr 2010 Location: In A Van Down By The River Posts: 2,674 Infractions: 0/1 (1)	One thing to remember is even with scouting at 100%, displayed ratings are only part of how good the player is. The true ratings in the editor can shed a lot more light on how valuable the player truly is. Still the AI needs a lot of work when it comes to roster building.

08-22-2024, 11:48 AM	#25
Bill Veeck Minors (Single A) Join Date: Aug 2014 Posts: 52	This thread I think shows the level of difficulty involved in trying to program AI in such a way as to replicate the complicated thinking actual GM's must engage in. There are so many factors involved that for the life of me I don't know how you program AI GMs without a lot of sliders -- then each GM is having different circumstances as you progress. Just saying real GMs have so much to do to rate acquisitions and DFAs/waivers/40-man roster constraints... This is really an area where ootp has already done so much work with a limited staff... It does drive you crazy to see a prospect DFAd. Any weakness in logic is amplified in an ootp season because the immersion is deep up until a really bad move occurs

08-24-2024, 02:20 PM	#30
OutS\|der Hall Of Famer Join Date: Apr 2010 Location: In A Van Down By The River Posts: 2,674 Infractions: 0/1 (1)	I hope one day we'll have a setting to approve trades or not

09-02-2024, 07:22 PM	#40
OutS\|der Hall Of Famer Join Date: Apr 2010 Location: In A Van Down By The River Posts: 2,674 Infractions: 0/1 (1)	From 2009 to 2018, a span of ten drafts, 47 percent of the position players selected by all MLB teams in the June draft ended up with a negative Wins Above Replacement (at best) in the majors. One in five position players picked in the first round (2009-18) have played fewer than 50 games in the big leagues. Selecting pitchers has been almost as much of a crapshoot: in the ten June MLB Drafts from 2009 to 2018, 39 percent of pitching prospects either did not pitch in the big leagues or have produced negative WAR in the majors. A stunning 26 percent of pitchers picked in the first round have never pitched in the majors. I wouldn't trade 10 unknowns for a known.