Trading Way Too Easy

[JMDurron]

Originally Posted by BIG17EASY

So let me get this straight ... trading is too easy because humans can outsmart a computer program when making complex trade offers? Unless your name is Garry Kasparov, this should be something that can happen regularly.

I agree with this, and personally have no problem using such exploits once I find them (real GMs traded with Ed Wade, after all...), but when something gets to the point that it is

1) Repeatable
2) Excessive in scope or result
3) Indicative of a disconnect in different areas of the AI logic

it's worth bringing up. I don't think the OP necessarily makes that point, but there is at least one situation I've come across that has all 3 apply.

The AI should not accept the same trade for the same players repeatedly, then put those same players immediately on waivers, and trade for them again, giving up something of at least nominal value every time! I believe there is also a potential bug enabling the situation to begin with, AND it became more common with the most recent patch.

Originally Posted by RchW

I've seen nothing to indicate that trading is way too easy in this thread. Show that trading is too easy by some specific examples please. Then offer possible solutions which several other posters have done that doesn't depend on a completely different logic structure and concept from the sport being discussed.

Working on it. I don't think the AI has issues valuing players as much as, to use lukasberger's phrasing, it has issues dealing with those values within a framework for the team's current situation. That's the ugly situation I'm seeing right now in my league, which is clearly (to me anyway) due to a combination of league-specific settings and underlying AI logic.

I promise I'm not trying to be a tease, this thread just might be a logical home for my findings once I get done with the phase of trading I have been doing.

[RchW]

Originally Posted by PSUColonel

I am not trying to compare the sports, nor say that FM is better. I am saying the system it uses to identify how players are "viewed" in relationship to their roles, and weather teams may or not be willing to part wit them works well. Just because it is a different sport doesn't mean you can't take a few ideas about how the program goes about certain things. I am saying it's a bit easy to pry away players on the block, and that I agree with injury log about how which players should be on the block, why, and what the AI team's expectation should be. If you have never seen this system within FM, then you wouldn't know what I am talking about. Maybe you have...I don't know. I am sure there are other who would agree with me on this though.

Why not just post about that vs a thread about easy trading which just isn't true? If your claim isn't true then why go through the pretense? I have no problem with you advocating FM as a better model but don't make unsupported claims about a deficiency in OOTP that doesn't exist in order to bring up FM as a solution.

I do know how FM works and I cant see any equivalence in evaluating players. One sport is based on preparation. The other on actual game play. One is based on very few but critical accomplishments. The other ignores every day results but looks for long term consistency. One calls for quick replacement for being off-form. The other calls for riding out a down cycle without overreacting. Player evaluation is completely different, rosters are different, player age vs accomplishment is different, the importance of each win and loss is different and so on... ...

[RchW]

Originally Posted by lukasberger

I don't think trading is too easy. So it'll be hard for anyone to show that. In fact trading on the higher difficulties may even be too hard. But trading is somewhat exploitable at times, for whatever reason. No big deal, but still something that it'd be nice to see improved. It's not a critical thing by any means, just something that'll likely be better in OOTP 20 than it is in OOTP 14.

So why use a nonsensical thread on "trading is too easy" to end up arguing that FM player evaluation is better? I think that's a complete waste of time and poor form.

Have a thread on "FM does everything better" and I'll ignore it maybe.

[Lukas Berger]

Originally Posted by RchW

So why use a nonsensical thread on "trading is too easy" to end up arguing that FM player evaluation is better? I think that's a complete waste of time and poor form.

Have a thread on "FM does everything better" and I'll ignore it maybe.

Good point.

I agree the presentation is poor, which understandably annoys people, but I do think that in this particular case, there's been some substantive discussion of some things that could use some tweaking, especially regarding the trade block.

It's just not the same stuff that's indicated by the thread title

[PSUColonel]

Originally Posted by RchW

Why not just post about that vs a thread about easy trading which just isn't true? If your claim isn't true then why go through the pretense? I have no problem with you advocating FM as a better model but don't make unsupported claims about a deficiency in OOTP that doesn't exist in order to bring up FM as a solution.

I do know how FM works and I cant see any equivalence in evaluating players. One sport is based on preparation. The other on actual game play. One is based on very few but critical accomplishments. The other ignores every day results but looks for long term consistency. One calls for quick replacement for being off-form. The other calls for riding out a down cycle without overreacting. Player evaluation is completely different, rosters are different, player age vs accomplishment is different, the importance of each win and loss is different and so on... ...

My original post, is about a situation where I found a very bad or easy trade situation because a player was on the block. As far as the rest goes, are conversations not allowed to evolve?

[HomerOne]

When I first played OOTP, there were some ridiculous trades being made by several GMs including myself. I solved the problem by changing the trade settings to "hard".

[jaysdailydose]

Originally Posted by injury log

Nunyer said essentially what I was going to. You can't talk about what the OOTP AI should do with Trade Block players without first talking about which players should be on the Trade Block to begin with, which is why I didn't think the previous discussion made much sense.

Of course there is no official 'trade block' in real life, so there's no real life model to follow. For me, the real question is: what do OOTP users want the trade block to be? I certainly don't want it to be what it is now - a list of really bad grossly overpaid players I would never want to trade for. What's the point of that? Instead I want to see a list of players an AI team is taking offers on - a quick overview of veterans I might be able to pick up from rebuilding teams at the trade deadline (so I don't need to scour every roster in the league), or an arb-eligible star that a rebuilding team a few years from contending has decided to move for a big package of prospects in the offseason. For the Block to work that way, though, AI GMs need to be smarter; they need to understand short and long term goals, and need to be able to evaluate how players fit those goals.

And when good players end up on the Trade Block, AI teams need to compare the value of what a human GM is willing to offer with the value of what other AI teams are willing to offer. One reason trading is easier than it should be in OOTP is that you're never bidding against anyone but yourself.

I disagree on a few different fronts... for one, I think you probably are bidding on players against the CPU. I'm sure the AI "calls up" the other AI GM's to propose trades and the like...

The players are all sets of numbers. Then you bounce them off the scouts numbers, the GM's numbers, and the like, and the AI either approves, counters, or rejects.

Of course the AI can be improved, but in the end, it is just weighing variables against each other and seeing if the variables meet certain criteria. Analysis might be better served thinking along those lines... how can you improve it in the system you have in place?

[jaysdailydose]

Originally Posted by PSUColonel

this is an area where I think FM does a great job. Now obviously you don't transfer list players in baseball, but FM gives you (and the AI) a sense of what each player is worth to each organization. This is don how by players are "listed". Some of those listings include: "hot young prospect", first team regular, back up player, not needed or excess player etc...

Unless you're playing stats only, isn't this what ratings are for? Scouting reports?

I honestly wouldn't care if they got rid of the trade block the way it works now... I think it is almost a "dumbing down" of the sim.

But then again, I don't need a map to make a trade...

[RchW]

Originally Posted by PSUColonel

My original post, is about a situation where I found a very bad or easy trade situation because a player was on the block. As far as the rest goes, are conversations not allowed to evolve?

If it makes sense.

You didn't come close to showing that players on the trading block were easier to trade. When challenged you resorted to "FM does it better" no matter what others may have posted.

I think the focus should be on improving OOTP with respect to how baseball works in real life, not posting flawed or cherry picked instances of weird AI and then suggesting that FM is the answer to these unproven "flaws".

I think we have the same goals. I have a problem with how you present your data and how you suggest solutions.

[jaysdailydose]

OP: Do you still have a clone of this game at this exact point? I have questions.

Is scouting on, for one? If so, what level? Also, would it be possible to see what OSA ranks the Rangers weaknesses as?

What are the owner attitudes for the Rangers, and if you can enable commissioner mode, you could even give us some of the numbers there?

I guess what I am trying to say is that there are far too many variables to just say that this was an "easy" trade by any means. You're giving three players (I won't just dismiss that, though) that may fill needs and at least one decent prospect that you've mentioned for a solid starter and salary relief.

Maybe the CPU's scouts are telling it that a guy is about to be a breakout star, or something of the sort. Plus, they already have Lewis on the block... is he still in the rotation, or has he lost his spot? Is he just pitching badly if he is in the rotation? Is he going to become a free agent and they are just looking to recoup anything of value before he leaves for nothing?

You're looking too much at the names and the many, many variables that may affect why a CPU would accept a trade.

[PSUColonel]

Originally Posted by jaysdailydose

OP: Do you still have a clone of this game at this exact point? I have questions.

Is scouting on, for one? If so, what level? Also, would it be possible to see what OSA ranks the Rangers weaknesses as?

What are the owner attitudes for the Rangers, and if you can enable commissioner mode, you could even give us some of the numbers there?

I guess what I am trying to say is that there are far too many variables to just say that this was an "easy" trade by any means. You're giving three players (I won't just dismiss that, though) that may fill needs and at least one decent prospect that you've mentioned for a solid starter and salary relief.

Maybe the CPU's scouts are telling it that a guy is about to be a breakout star, or something of the sort. Plus, they already have Lewis on the block... is he still in the rotation, or has he lost his spot? Is he just pitching badly if he is in the rotation? Is he going to become a free agent and they are just looking to recoup anything of value before he leaves for nothing?

You're looking too much at the names and the many, many variables that may affect why a CPU would accept a trade.

1. yes, scouting on...average, Rangers has weaknesses at catheter, LF, RF, Rotation and bullpen

[injury log]

Originally Posted by jaysdailydose

I disagree on a few different fronts... for one, I think you probably are bidding on players against the CPU. I'm sure the AI "calls up" the other AI GM's to propose trades and the like...

No, the AI doesn't do that.

Originally Posted by RchW

You didn't come close to showing that players on the trading block were easier to trade.

It's funny - you think that Trade Block players are not easier to acquire, but should be, and I think Trade Block players are easier to acquire, but shouldn't be.

I've never formally tested it, but many times I've seen the AI trade valuable players who are on the Trade Block for next to nothing. Now, it's hard to find good players on the Trade Block to begin with; the AI generally only puts players on the Trade Block who are 'Angry'. For a player to get angry, usually a few things need to be true - the player needs to be playing badly, or his team does, and the player needs to be upset about his role. Position players and SPs are never unhappy about their role unless they suck; if they're good, they'll be in the starting lineup or rotation. But there is one type of player who can both be good and unhappy about his role: good relievers who want to be closers, but who aren't. It seems any decent reliever who was at one time a closer (even if just because of injury) believes he should remain a closer in perpetuity. So it seems, if you have Morale on, every OOTP league has twice as many relievers who think they should be closers as there are closer jobs available. I've often been looking for a good reliever at the trade deadline, and if I just try to trade for some random good setup guy, I usually need to give up a very good prospect in return. If instead I find a similar guy on the Trade Block, I can get him for almost nothing.

It's possible this was improved in OOTP14 (I haven't really been looking for the issue, and I've pointed it out to Markus several times, so maybe he did something to fix it) but this was an obvious exploit in OOTP12 and 13.

[Nunyer]

Originally Posted by jaysdailydose

OP: Do you still have a clone of this game at this exact point? I have questions.

Is scouting on, for one? If so, what level? Also, would it be possible to see what OSA ranks the Rangers weaknesses as?

What are the owner attitudes for the Rangers, and if you can enable commissioner mode, you could even give us some of the numbers there?

I guess what I am trying to say is that there are far too many variables to just say that this was an "easy" trade by any means. You're giving three players (I won't just dismiss that, though) that may fill needs and at least one decent prospect that you've mentioned for a solid starter and salary relief.

Maybe the CPU's scouts are telling it that a guy is about to be a breakout star, or something of the sort. Plus, they already have Lewis on the block... is he still in the rotation, or has he lost his spot? Is he just pitching badly if he is in the rotation? Is he going to become a free agent and they are just looking to recoup anything of value before he leaves for nothing?

You're looking too much at the names and the many, many variables that may affect why a CPU would accept a trade.

I agree that there are many variables... But I do understand why PSUColonel would think that this trade appears wacky. The Rangers are throwing in $5m to get rid of a $2m contract... and getting two scrub vets and a marginal prospect in return. Now, if only the scrub players were involved, then maybe that would make more sense... But it appears to me that the Rangers are not properly valuing that cash. It's not a 'salary relief' situation when their budget room drops by $4.9m.

As far as CPU scouting, that's something I can't comment on with v14... As I've had scouting off in the only fictional league I've spent any time with... But in past versions, it seemed to me that the CPU didn't use scouting in the same way it was presented to human players. If you ever wanted to know which one of your 3/4/5 star prospects was legit, just dangle them in trades and see which ones the CPU teams always want. Those are the ones that are real... and the leftovers are the guys that your scout is wrong about. So there is the possibility that the prospect that PSU is offering is much better then his scout is leading him to the believe... That's the only thing that might explain the Rangers giving up $5m in budget space to get him.

[RchW]

Originally Posted by PSUColonel

1. yes, scouting on...average, Rangers has weaknesses at catheter, LF, RF, Rotation and bullpen

Sometimes posting in these forums is like getting a catheter inserted.

[David Watts]

Originally Posted by Nunyer

I agree that there are many variables... But I do understand why PSUColonel would think that this trade appears wacky. The Rangers are throwing in $5m to get rid of a $2m contract... and getting two scrub vets and a marginal prospect in return. Now, if only the scrub players were involved, then maybe that would make more sense... But it appears to me that the Rangers are not properly valuing that cash. It's not a 'salary relief' situation when their budget room drops by $4.9m.

As far as CPU scouting, that's something I can't comment on with v14... As I've had scouting off in the only fictional league I've spent any time with... But in past versions, it seemed to me that the CPU didn't use scouting in the same way it was presented to human players. If you ever wanted to know which one of your 3/4/5 star prospects was legit, just dangle them in trades and see which ones the CPU teams always want. Those are the ones that are real... and the leftovers are the guys that your scout is wrong about. So there is the possibility that the prospect that PSU is offering is much better then his scout is leading him to the believe... That's the only thing that might explain the Rangers giving up $5m in budget space to get him.

But we still don't know if PSU could get the Rangers to make a deal for a single player or 2 players. In fact, his screenshot shows the Rangers turning down a 2 for one offer. He then adds the the magic 3rd player to the mix and the AI bites.

[SirMichaelJordan]

Originally Posted by PSUColonel

this is an area where I think FM does a great job. Now obviously you don't transfer list players in baseball, but FM gives you (and the AI) a sense of what each player is worth to each organization. This is don how by players are "listed". Some of those listings include: "hot young prospect", first team regular, back up player, not needed or excess player etc...

OOTP could def use some improvements but like others have said, football and baseball is comparing apples to oranges. In american sports it's all about the what have you done for me lately model. If a prospect is not playing to his ability in football then that guy is loaned to a lower league team to gain some first team experience but in baseball, this same prospect is only given a certain amount of time before a team is ready to move forward without him and package him in a trade. Teams trade prospects even if the player is good as they would have to give him up in order to acquire a veteran player who can help them out during the pennant race.

Yes it would be useful to perhaps have a label on players to make things easier to see how the AI values them but the game already have something like this when you go into the team's positional rankings (forgot the name of the screen) where you can sort how teams value players In their organization by using current ratings or potential which can also be sorted to only view prospects. This is only using ratings however and this will ultimately come down to user settings.

Football is all about developing homegrown talent. Not so much in baseball. The draft is not even a sure thing enough for baseball GM's to even label a prospect as their next big thing. If they do then those players are immediately placed on their 40 man roster like Harper and a few others.

In FM, its not all about the ratings because you have other factors that the AI will use to label how valuable a player is to the organization and these things are like position/ability, nationality, league rules, reputation and etc.

As for the trade block and the trade presented in the OP, there isn't really enough information to even discuss this topic. No one here are actually testing things. We are currently having a discussions on past experiences from older games and one scenario in the OP. Where is the data?

Are we saying that the trade block is only filled with scrub players? Is that all year round? I've seen teams in rebuild mode place certain players on the block.

Im sure trading for players who are on the block may be easier but the game have to many options, setting and variables to say its too easy without adjusting anything and testing different scenarios. Maybe your user settings are making it easy? After all you are giving up 3 players for a player that the AI doesn't want anymore. Check and see how the AI team used those players and where they rank in the AI's organization. Maybe the AI changed their position and got a rating jump somewhere. On the other hand, maybe this trade is rejected if you favored prospects more while at the same time devaluing scrub vets...

[RchW]

Originally Posted by injury log

It's funny - you think that Trade Block players are not easier to acquire, but should be, and I think Trade Block players are easier to acquire, but shouldn't be.

Yes and it may have been confusing. I don't want them easy to get in the example PSU used. I agree that is a flaw, but not one I'm seeing in my game. Perhaps game type has an effect on these issues

I want them to be easier to get based on their value to their current team. I understand the Upton case you cited earlier but right or not the current trading block in OOTP is generally filled with overpaid duds. That status equates to reduced value and the AI should adjust accordingly. The AI should parse offers that reduce the cost of releasing the player because in the rank of values, releasing a player is the next step after the trading block.

Any offer of any player that reduces the cost of releasing the player by half or more should be almost automatically accepted. Any offer that includes a player of the same position or one that can play the position should have the AI opening the cash drawer even more. This applies to any combination of cash or contract exchanges. The bigger the contract the more the AI should be willing to give in cash.

Please note the AI already does this in the shop player exchange. Many of the offers are for decent performers not on the block that have bigger contracts than they should. It seems that the AI is saying that it has a cheaper replacement in mind.

These comments also apply to players on waivers. The AI consistently and wrongly lets players be claimed off waivers in spite of trade offers because it won't reduce the expected value. It's already made the decision to lose the player, why it would refuse to consider a asset in return escapes me.

My impression is that money, payroll, roster placement and player value seem to exist as separate entities. It's the only way to explain why a player on waivers and a player on the block get treated like they are core assets.

[Lukas Berger]

Originally Posted by RchW

My impression is that money, payroll, roster placement and player value seem to exist as separate entities. It's the only way to explain why a player on waivers and a player on the block get treated like they are core assets.

Agreed.

[jaysdailydose]

Originally Posted by injury log

No, the AI doesn't do that.

Obviously I didn't mean that it works EXACTLY like that. What I mean is that the CPU must have a mechanism to trade between AI teams or else there would be no trade movement.

[PSUColonel]

I think a lot of you are missing the point. My settings are at a very challenging level I assure you. The issue is the cash. Why was it willing to give 5 million in cash in addition to a very nice mid-rotation starter. It's the cash part that has me saying "hey something is not right here"