Big problems with trade AI in 5.10 historical league

[MannyTrillo]

I just started up a league in 1915. When I click on Rogers Hornsby (18 yo SS, B range, BGBAGG talents, 5-star prospect, #1 prospect in the league), the computer lists about 25 players from my A's organization that it would accept, including 2B Jack Barry (27 yo 2B, D range, 445049, AFAPAB, 1-star player), among other very mediocre players. Scouts/finances are off, so that's not the problem.

If I ask for Miller Huggins, a 36 year old 2 star 2B, along with Hornsby, there are no players on my team that the computer would accept.

[MannyTrillo]

...and it's pretty much the same in my fictional league- the #1 prospect in the league, (636654, GFGBBG) can be had for a 3 star MR prospect.

[kq76]

I think a lot of people were waiting for someone else to first bring this up. I see the same thing.

[holyroller]

Likewise. HUGE flaw. Makes it very hard to use the trade function as I don't really like taking advantage of the AI.

Edit- This is a GForce league, so that makes it 3 for 3. Historical, fictional, modern.

[darkhorse]

Originally posted by holyroller
Likewise. HUGE flaw. Makes it very hard to use the trade function as I don't really like taking advantage of the AI.

Edit- This is a GForce league, so that makes it 3 for 3. Historical, fictional, modern.

and Finnish.

[x94blair3]

I was somewhat put off that the trade AI would take Steve Avery from my Tigers for a 2-4-6 A/A/G (reduced ratings) 4 star pitching prospect that was only 24. Then when I got him, his ratings were reduced to 3 1/2 stars so I'm not sure what that means.

At first I was horrified..then I realized for whatever reason Avery was 4, 5, 6 even at 31 and 1 star...

Regardless I do question the validity of having the trade AI "show you what it'll take". If there are any flaws or questions over it's logic, they're sure to be exposed if the engine will "show it's cards". Seems like instead of tweaking the engine to improve upon what the computer will logically want, Markus took a big risk in an effort to provide some sort of trade block which I didn't realize was the the biggest trade AI beef.

Having said that, perhaps the engine knows something I don't and this guy will completely tank it.

-Nick

[suurimonster]

Check out thread:

http://www.400softwarestudios.com/bo...threadid=31770

Personally I think that the AI is improved, especially with regard to evaluating prospects. But at least it will take some systematic, objective tests to check it out.

[MannyTrillo]

Originally posted by suurimonster
Check out thread:

http://www.400softwarestudios.com/bo...threadid=31770

Personally I think that the AI is improved, especially with regard to evaluating prospects. But at least it will take some systematic, objective tests to check it out.

Try this one: trade mediocre players for top prospects, then package the prospects together for star players. Presto, you can turn mediocre players into stars.

Also, I just noticed that half of the players in my historical league with AVG ratings over 6 have been released to free agent pool.

I've never said this about any new version or patch, but the game as it is now is unplayable.

[JWay]

I've also seen this. I traded 7 million dollar Jeff Cirillo for Marlon Byrd. Straight up. Cirillo is junk and Byrd was a 4.5 prospect

[ctorg]

Don't trust those blue stars. It's very rare for one to actually live up to the hype. Remember, the stars are generalizations. Take a look at the player's ratings, talents, and age. If a guy is 23 and has a 1 in average, he's not likely to become a great average hitter, even if he's brilliant in that category. In other words, he's overrated.

[blmeanie]

Originally posted by MannyTrillo
I've never said this about any new version or patch, but the game as it is now is unplayable.

without expressing an opinion, you of course are only commenting on solo play where the AI is needed, correct? Your comment does not apply to online leagues if I read you right.

[JWay]

Even if it doesn't affect online leagues solo play is a lot of fun for most owner. Anyways Byrd is young and filled out basically. hes going to be a a 4-5 star big leaguer unless he gets injured...

[Deft]

Do any of you have 3 years down the road type evaluations of those trades? I think its real easy to make suppositions about future potentials and performance but maybe doing a study and making several of these trades and then evaluating there creadence at that point might give you a better feel for the AI trading effectiveness.

[blmeanie]

for me too, I just wanted to make the point for some of the newer players so they would understand.

[MannyTrillo]

Right, I mean the computer GM AI. I'm not aware of any big problems with the in-game AI, so on-line leagues shouldn't be affected at all.

Here's an experiment that won't take several seasons of tests: take the 1962 Mets. Before starting the season, trade away mediocre players for 4 and 5 star prospects. Bundle the 4 and 5 star prospects together and trade them for good players. See how many games you win. I will try this out later on if I have time (though I probably won't). I would bet an experienced OOTP player could turn the '62 Mets into a 90 win team right away.

[suurimonster]

[i] I will try this out later on if I have time (though I probably won't). I would bet an experienced OOTP player could turn the '62 Mets into a 90 win team right away. [/B]

This discussion is a little pointless unless you do test it. Also it is misleading to talk about the Mets or any real team, you are just imposing your prejudices. Lets talk numbers - in my fictional league, the trades look reasonable - 5* prospects almost always end up 3* or less, and the new AI seems to recognize that ratings are what counts. When I look at the numbers and ignore the stars, it looks very reasonable. The computer will not give up 9-10 ERA rated SPs for ANYTHING (except another 9-10 ERA SP . ), regardless of whether they are 3,4 or 5 *. He will happily give up 5* prospect SPs, but then they are 24-25, with 1-3 ERA - definite no-hopers.

[Sweed]

Originally posted by MannyTrillo
Try this one: trade mediocre players for top prospects, then package the prospects together for star players. Presto, you can turn mediocre players into stars.

.

Here we go again As Steve said in numerous other threads on numerous occasions trading AI will never be perfect. This is only a problem in solo leagues so I can only ask why would anyone do this? I mean one AI team certainly doesn't do this to another AI team does it?

Of course Markus can tighten up the trading and then as we've seen the next guy comes in and says "I can't make any trades", so here we go again

Would I like it to be perfect? Of course but that's not going to happen. Why not just base your trades in some sort of reality and have fun instead of wasting time figuring out how to rip off the AI? Let's remember we are the human player and are in control, none of this happens unless we do it. So if you can't control yourself enjoy all your championships, I'll be content just trying to get to the top in as realistic a way as possible

Ok flame away I'll be enjoying my solo league while you do. Maybe I'll check back in when I see this thread has reached 10 or 20 pages.

[KCRoyalsFan]

I agree with Sweed. It will never be perfect.

[holyroller]

I agree with the fanboys, the game is perfect. Let's not fix any bugs and just pretend there are no problems with the game until a new game comes along and OOTP dies. We could be just like High Heat.

[ghulten]

Aw, jeez. This deteriorated quickly.