AI to AI trade logic is awful

[Sweed]

Originally Posted by wallewalls

Yes similar for me, my game is in 2080 so I don't really have the general knowledge of all the players in the league at this point that I would have with the real MLB, and I am also just on autopilot pretty much at this point as I have nothing left to accomplish so it's more so just about seeing what my franchise players become for the rest of their career more than anything else, as I just want to get to 2100 and call it there so I cannot say whether all the deals are bad or not, but it seems to be that they are just trading for the sake of trading, which was more so my initial gripe when this issue first started.

I was hoping that with OOTP 24, I would be starting up a new long-term game finally, but I just haven't gotten very immersed in any of my attempts thus far. Maybe I will try again with some more patches,

In regard to the bold... My thoughts based on my observations in my game.

I don't think that they are trading just to trade. I base that first on not seeing the same guys getting traded again and again. IE once on the AI team they are sticking. Teams aren't thinking "I need a 2b, trading a LF for that 2b, and then backtracking to "I need a LF". Also, after the initial large number of trades, the volume of trades settled down to a "normal" level. To me if the AI was "trading just to trade" it would look a bit like the old "sign, release, sign". IE the trading would be on a treadmill, and not showing such a large drop in volume.

To be fair my sample size is small as I play out all of my games, so only March, April, and half of May have gone by on "average" frequency and 25/25/25/25 evaluation. Maybe when I hit the high volume of late June and July I'll see it differently.

[dodgerblue88]

I'm about ready to migrate my OOTP 23 save and I'm concerned about the initial flurry of trades too. I'm wondering at what point in the season did you guys migrate? I'm thinking about importing on the day after the world series, that would give the AI the opportunity to tender/non-tender contracts and contract options, and hopefully avoid the initial "AI rebalance"

[LansdowneSt]

Originally Posted by dodgerblue88

I'm about ready to migrate my OOTP 23 save and I'm concerned about the initial flurry of trades too. I'm wondering at what point in the season did you guys migrate? I'm thinking about importing on the day after the world series, that would give the AI the opportunity to tender/non-tender contracts and contract options, and hopefully avoid the initial "AI rebalance"

I think in the OOTP manual it says the day after the WS before the engine starts churning to update folks. Away from the game atm, but see if there is anything in there about it. Search for Matt and search terms on importing as I feel this came up recently on another thread. Game's pretty due for a patch so no harm waiting for the next as you hear back from folks on the best time to migrate.

[wallewalls]

Originally Posted by dodgerblue88

I'm about ready to migrate my OOTP 23 save and I'm concerned about the initial flurry of trades too. I'm wondering at what point in the season did you guys migrate? I'm thinking about importing on the day after the world series, that would give the AI the opportunity to tender/non-tender contracts and contract options, and hopefully avoid the initial "AI rebalance"

I did it in early February (but past the point where the preseason started), so teams were pretty much locked in financially at that point and there were still a large number of trades.

[Sweed]

Originally Posted by dodgerblue88

I'm about ready to migrate my OOTP 23 save and I'm concerned about the initial flurry of trades too. I'm wondering at what point in the season did you guys migrate? I'm thinking about importing on the day after the world series, that would give the AI the opportunity to tender/non-tender contracts and contract options, and hopefully avoid the initial "AI rebalance"

I did mine the day after the WS, ie the day before the offseason starts.

Not sure what all you've read but a quick recap of my experience is there were about twice as many trades on import in v24 than I had in v23.

Temporarily went to "very low" frequency during ST, but went back up to "average" when Lukas, and others seemed to be seeing the same high number of trades at the beginning followed by a slow down. Since going back to average I've only completed April. Here is a link to my post on those trades. https://forums.ootpdevelopments.com/...&postcount=165

I'm currently on May 22 and, so far there have been three trades. Same type as the others minor leaguers and AAA players that are capable of coming up to provide depth to the MLB team.

[Thundercrack]

Originally Posted by Sweed

I did mine the day after the WS, ie the day before the offseason starts.

Not sure what all you've read but a quick recap of my experience is there were about twice as many trades on import in v24 than I had in v23.

Temporarily went to "very low" frequency during ST, but went back up to "average" when Lukas, and others seemed to be seeing the same high number of trades at the beginning followed by a slow down. Since going back to average I've only completed April. Here is a link to my post on those trades. https://forums.ootpdevelopments.com/...&postcount=165

I'm currently on May 22 and, so far there have been three trades. Same type as the others minor leaguers and AAA players that are capable of coming up to provide depth to the MLB team.

Watch out for the off-season. I set to very low, had only a few trades in-season, then the AI went nuts in the offseason

[Sweed]

Just finished May, still on average frequency and evaluations at 25/25/25/25.

9 trades in May, v24. For comparison: in my last season in v23 for May there were 11 trades using the same settings.

At first I was adding comments to each trade in my notes within OOTP. These are bolded in the quote, not because of importance, but only to make them easy to see. Stopped adding comments as the trades were pretty much all of a similar type, ie minor leaguers for AAA/MLB borderline depth help, or minor leaguer for minor leaguer.

There's not an every day MLB player in any of these trades, nor highly rated prospects. At best AI teams are trying to improve trading age for youth, adding positions of weakness for positions of strength. Or maybe trading just to trade? The trades are of such little value, in the big picture, I haven't dug deep enough to see. If I untick the "include minor league deals" I doubt many of these even show up in the report.

So far there has been no reason, I can see, that any of this is detrimental to my on-going game.

On to June.

May 15, Trade #2: the Phillies announced today the acquisition of 26-year old minor league 3B Eddie Dominguez (45/45) from the Seattle Mariners in exchange for
20-year old minor league LF Manny Sotilo (20/35) and 23-year old minor league CF Alvaro Morales (35/40).
2 guys with bench potential, the other maybe a cup of coffee.

May 16, Trade #3: the Blue Jays announced today the acquisition of 29-year old minor league LF Danny Campana (40/40) from the New York Yankees
in exchange for 24-year old minor league 3B Ryan Blumenstein (35/45) and 22-year old minor league RHP Joe D'Angelo (25/35).

Campana has played in 120 games and posted a lifetime .293 average. He has 115 hits, 26 doubles, 4 triples and 16 homers. In addition he has driven in 80 runs and scored 51 times.

More minor leaguers for borderline MLB/AAA player.

May 23, Trade #4: the Mariners announced today the acquisition of 28-year old minor league RF Jeremy Manning(40/40) and 23-year old minor league RHP Bobby Duran(25/40) from the Baltimore Orioles in exchange for 24-year old minor league LHP Mike de Jesus(20/25) and 23-year old minor league RF Angel Sandoval (40/45).

More of the same type of trades. Sandoval could end up sticking as a bench player. Manning will stick if his control improves.

May 23, Trade #5: the White Sox swapped 23-year old minor league SS Travis Love to the Angels getting 33-year old minor league RHP Pasqual Estolas in return.
In his lifetime, the 33-year-old Estolas has won 18 and lost 38, piling up 301 strikeouts and 291 walks in 408 innings with a 5.91 ERA. Opponents have hit .267 against him.

SS could make it as a backup someday. RP, has control issues. Almost a meaningless trade.

May 27, Trade #6: Tampa Bay Rays and the San Diego Padres have agreed to a deal that will have 24-year old RHP Mike Green (45/50/) and 24-year old minor league RHP Nate Reich (20/20) heading to San Diego in exchange for 26-year old minor league 2B Matt Calkins (35/40)

To this point in the season he has had no saves. Green has made 16 relief appearances, tossed 20.1 innings and struck out 16 with a record of 1 win and no losses.

May 27, Trade #7: the Rangers announced today the acquisition of 28-year old minor league 3B Ralph Tracey (20/25) from the Cleveland Guardians in exchange for 23-year old minor league 3B Hugo Salazar (30/45) and 20-year old minor league 1B Tony Pagan (20/35)

May 27, Trade #8: The Minnesota Twins and the Seattle Mariners have agreed to a trade that will send 20-year old minor league RHP Jesus Sanchez (20/25) and 19-year old minor league 1B Jeremy Patterson (20/40)to Seattle, while 27-year old minor league CF Jonathan Weilnau (45/45) will be shipped to Minnesota.

May 28, Trade #9: Cardinals announced today the acquisition of 30-year old minor league 3B Chris Christensen (45/45)from the Colorado Rockies
in exchange for 24-year old minor league C Ryan Bullard (40/40) and 23-year old minor league 2B Tim Winters (35/45)

For the season Christensen has hit .182 with no home runs and 1 RBIs. He has played in 14 games and scored 2 times.

[billyray1984]

Originally Posted by Thundercrack

Watch out for the off-season. I set to very low, had only a few trades in-season, then the AI went nuts in the offseason

I have the newest patch and the first offseason after a live start in May, even with very low trades and the new defaut settings for AI evaluation, the AI was trading to trade and there were some pretty wild stuff in there. For example, the Brewers traded both Burnes and Woodruff after extending them both in season for an unimpressive prospect package that looked like a salary dump and then traded for Clarke Schmidt and Domingo German since they had lost 2/5 of their rotation.

As a player, I did not get offered such sweetheart deals, but the AI to AI trades were nuts.

In season, the trades (number/frequency/players involved) were fine though.

[wallewalls]

Originally Posted by billyray1984

I have the newest patch and the first offseason after a live start in May, even with very low trades and the new defaut settings for AI evaluation, the AI was trading to trade and there were some pretty wild stuff in there. For example, the Brewers traded both Burnes and Woodruff after extending them both in season for an unimpressive prospect package that looked like a salary dump and then traded for Clarke Schmidt and Domingo German since they had lost 2/5 of their rotation.

As a player, I did not get offered such sweetheart deals, but the AI to AI trades were nuts.

In season, the trades (number/frequency/players involved) were fine though.

Saw something similar. The Dodgers extended Urias just to then trade him to the Cardinals.

[md40022]

I did my best to pay pretty close attention to this during the 1st season of my sim and while I don't think the logic of the individual trades was awful, it did seem that teams were trading just for the sake of trading. No rhyme or reason at all.

You'd see the same team trades prospects for a veteran, leading you to believe they were in "win now" mode, but then 24 hours later that trade a veteran for some prospects as you'd do in a rebuild. And this same team would do this over and over again as the trade deadline approached.

There definitely seemed to be WAYYYYYYYYYY more trades in year 1 of my OOTP24 sim than I have ever seen in any version of the game ever before. Given all of the trade changes made in this year's game, it seems pretty clear that was the new feature that kind of got beaten to death and definitely needs some tweaking for next year. That said, I didn't really see bad trades when evaluation each trade individually, but it seemed like none of the teams had any idea what direction they were going in. Making a rebuild trade Monday only to make a "win now" trade Tuesday and this would repeat over and over and over again. If this was an IRL thing, you'd think your team's GM was on drugs.

IMO the number of trades probably could be cut in half easily and the purpose behind trades definitely needs some adjustment too (ie. how teams should operate when in "win now" vs. "rebuild")

[billyray1984]

Originally Posted by md40022

I did my best to pay pretty close attention to this during the 1st season of my sim and while I don't think the logic of the individual trades was awful, it did seem that teams were trading just for the sake of trading. No rhyme or reason at all.

You'd see the same team trades prospects for a veteran, leading you to believe they were in "win now" mode, but then 24 hours later that trade a veteran for some prospects as you'd do in a rebuild. And this same team would do this over and over again as the trade deadline approached.

There definitely seemed to be WAYYYYYYYYYY more trades in year 1 of my OOTP24 sim than I have ever seen in any version of the game ever before. Given all of the trade changes made in this year's game, it seems pretty clear that was the new feature that kind of got beaten to death and definitely needs some tweaking for next year. That said, I didn't really see bad trades when evaluation each trade individually, but it seemed like none of the teams had any idea what direction they were going in. Making a rebuild trade Monday only to make a "win now" trade Tuesday and this would repeat over and over and over again. If this was an IRL thing, you'd think your team's GM was on drugs.

IMO the number of trades probably could be cut in half easily and the purpose behind trades definitely needs some adjustment too (ie. how teams should operate when in "win now" vs. "rebuild")

I'm thinking of starting a normal game, leaving the default settings on and disabling trades until Opening Day of the next year to see if it's a year 1 problem or if it's a player eval problem. My guess is that the AI wants to construct its roster and minor leagues in a different way than in previous years (and in a way that might not be so easily tweaked with the AI eval weights) so it makes moves like crazy in year 1 and then slows down once it's "happy" with its roster and minor league depth in year 2.

[md40022]

Originally Posted by billyray1984

I'm thinking of starting a normal game, leaving the default settings on and disabling trades until Opening Day of the next year to see if it's a year 1 problem or if it's a player eval problem. My guess is that the AI wants to construct its roster and minor leagues in a different way than in previous years (and in a way that might not be so easily tweaked with the AI eval weights) so it makes moves like crazy in year 1 and then slows down once it's "happy" with its roster and minor league depth in year 2.

I just started year 2, so we shall see how it plays out. I will say, the offseason between year 1 and year 2 was super aggressive with trades, but to your point maybe now it will slow down as this was the 1st offseason where the CPU could maneuver. Right now there's just wayyyy too many trades for my liking and teams don't seem to have any real vision as to whether they are trying to contend or rebuild... I'm hoping you're right.

[PSUColonel]

You need to give it some time (a season or so) for things to work themselves out. At the start of a save, there is likely going to be a lot of transaction activity (including trades) until the AI sort f figures out what it wants to do on each particular team. It should calm down after this initial flurry of activity.

[sdw1000]

For as much as people on the forums like to bitch about this game, and trust me I've had/have my issues, I am having fun in my league! I just wanna see more Negro League teams besides from NNL. Hillsdale please!

[Sweed]

Just finished June in my first season in v24. Average frequency, 4- 25's evaluation.


With the settings unchanged from v23 to v24..
V23 had 12 trades.
v24 had 13 trades.
This is with the "include minor league deals" checked.

I did not list all of the trades like previous posts as the patch was released at mid June in my game. So the v24 trades include prepatch and post patch. Still a lot of minor leaguers for minor leaguers. No MLB stars, or MilB stars for garbage. No players traded more than once. Value for both sides of trades appears to be pretty equal.

[PepeSylvia]

Originally Posted by Sweed

Just finished June in my first season in v24. Average frequency, 4- 25's evaluation.


With the settings unchanged from v23 to v24..
V23 had 12 trades.
v24 had 13 trades.
This is with the "include minor league deals" checked.

I did not list all of the trades like previous posts as the patch was released at mid June in my game. So the v24 trades include prepatch and post patch. Still a lot of minor leaguers for minor leaguers. No MLB stars, or MilB stars for garbage. No players traded more than once. Value for both sides of trades appears to be pretty equal.

I think the off-season is when it goes haywire.

[Cactusguy21]

The trade frequency is still insane, we are a month and a half from the deadline and already several trades per day in my 16 team league. This league had an inaugural draft too, so it's not like the AI GMs got stuck with players they don't like!

Absolutely haywire.

[Sweed]

Originally Posted by PepeSylvia

I think the off-season is when it goes haywire.

Maybe, maybe not. Some have posted they've seen way more trades in year one of v24 than they've ever seen in any other version. Some are posting the same vets are being traded more than once, something I have not seen at all.

The best way to proceed is to document data to get a factual picture. Then one can compare it to previous versions with the same settings and time frame. The key word is "document". It's not hard to go to the trade log and count how many trades there were in a month. Then for an imported game the log is still in tact so one can go back to the previous year from the previous version to compare.

I play out all game so getting to offseason number 2 (imported day before offseason from v23 to v24) will take a bit of time. What I saw in v23 from off season 1 to offseason 2 was an almost halving of trades. I will be interested to see how that goes with v24.

This is a link to my initial post in this thread with some numbers. https://forums.ootpdevelopments.com/...2&postcount=60 In there I say I changed to "low" frequency but that only lasted about one game week. At that time I went back to my v23 settings of "average" trade frequency while my player evaluation settings never varied from the "4-25's".

[Marcus75844]

The trade frequency still needs to be toned down. I left everything on default, and I was getting a lot of trades in June/July and even more trades in the off-season.

Also, many of these trades were for random minor leaguers, and they didn't really serve a point. It was almost like the CPU teams were just trading to be trading. If the number of trades and quality of trades can be improved in the next patch, that would be greatly appreciated.

[billyray1984]

Originally Posted by Cactusguy21

The trade frequency is still insane, we are a month and a half from the deadline and already several trades per day in my 16 team league. This league had an inaugural draft too, so it's not like the AI GMs got stuck with players they don't like!

Absolutely haywire.

Do teams have money ? Teams having no money is a recipe for a million head scratching moves since the AI will see that it has no money and that it needs to dump salary (even after an inaugrual draft). After the trade, it'll see that it has a hole that needs to be filled and trade to fill that hole, but sometimes, it fills that hole with another guy making a lot of money, perpetuating the cycle.

The human player knows how to dump salary the right way : send the guy to some other team with a prospect attached and some retained salary, and then plug in a guy making the minimum in his place. The AI will often dump salary with a lot of retention on its part (80%+), without attaching a good player, and replace the player it dumped with another player making a lot of money. Not always, but it happens.