Trading/Free Agent Logic?

[Rosco Peabody]

Is there anything within the trading logic or free agent signing logic that accounts for whether the team already has a good player at a particular position?


I've seen a few times where a team will either trade for or, more frequently, sign a guy to a particular position where they already have a starter. I wonder if it has to do with depth, like maybe the position as a whole is shallow, though they do have a decent starter.





This one is a little different, but there is one team in my league right now (my fictional league's version of the Yankees) who has a high priced, top of the line closer, a very expensive setup guy and another great middle reliever, just lost two starters to free agency, and is making a push to try to sign the top closer in free agency.

[battists]

There is definitely that sort of logic in the game. However, ultimately it’s just an algorithm, and it’s one with a decent degree of variability, so that it isn’t entirely predictable.

If you come across scenarios like this and really feel they are “beyond the pale“, I would post the details in the bug forum, and perhaps Markus will tweak that algorithm slightly.


Sent from my iPhone using Tapatalk

[dbqs]

Sounds like the actual Yankees.

[battists]

This is one of these areas where it must be so incredibly difficult to write a mathematical formula that comes out “realistically“. I can imagine that if you tweak said formula even slightly to correct for an outlier like this, it could cause another outlier to slip out the other side or whatever.

Certainly there have been cases where teams have made moves addressing a position of strength in unexpected ways before. (A related move in a different sport that comes to mind is the Eagles drafting multiple cornerbacks early in the 2002 draft when they already had Pro Bowl caliber cornerbacks in Troy Vincent and Bobby Taylor.)


Sent from my iPhone using Tapatalk

[Rosco Peabody]

Makes sense. Related to this, I can't stand when a computer owner starts one player over another because, on paper, it makes sense for the OVR rating. However, realistically, in real life the second guy wouldn't because the first guy either is an all time player on the back nine or a huge contract. I mean if a guy is a future hall of famer, for instance, even if he is playing out one last year and stinks, the team would still play him out of respect.

Obviously that stuff is impossible to program into an algorithm, which is why I wish there was an option in Commissioner mode to force the computer to start a guy or play him at a particular position or play him in a particular role. I know you can do it for your own team, but I wish that you could do it for computer teams too

[jdw]

It happens a bit too often for comfort. Running a historical replay starting in 1974, this one popped up the last day I played:

Tuesday, November 22nd, 1977
Boston Red Sox: Signed free agent C Ted Simmons to a 5-year contract worth a total of $3,800,000.

Which led to this give away:

Thursday, December 1st, 1977
The Boston Red Sox traded 29-year-old catcher Carlton Fisk to the Oakland Athletics, getting 25-year-old closer Adrian Devine and 23-year-old reliever Don Aase in return.

Fisk is about a year and a half older than Simmons. Ted has been better in each of the four seasons of the replay. But Fisk has been good in all those four seasons, and hasn't sprouted an injury of note in any of them. Catcher isn't a problem for the Sox, who have won the AL East the last two seasons and the ALCS in 1977 before losing in the WS. There are other places they can spend money. Worse, the AI bombed out on getting value for Fisk.

Any of us could pull examples. My favorite example of the worst in this replay was:


Bobby Grich
11/21/1975 Signed a 8-year contract worth a total of $4,928,000 with the Cincinnati Reds organization.

The Reds of course have Morgan under a multi-year deal. They just let Concepcion walk to the Cards, so I thought they might do what really happened in 1976 - the Halos played Grich at SS which he had played a good deal in his first full season for the O's in 1972. Nope, didn't happen. Grich was stuck at DH. He went from being a 6.9 & 6.1 WAR in 1974-75 (3rd in the 1975 MVP) to 1.2 & 1.3, partially due to those not being the best years for him in real life but also due to zero defensive value rather than Gold Glove Second Base value.

Probably not even worth mentioning that the game doesn't even have a SS value for Grich when looking at it right now despite the fact that it's where he played in 1976. I thought one of points of having the game check after each season of the replay for adjust the players changing/adding players. I have the uneasy feeling that Robin Yount is going to play his entire career at SS and never get a CF rating. Or if you start at 1974, the game thinks Jim Rice is a CF and pretty decent one, while Fred Lynn is a lesser CF and a better LF... yes, exactly opposite of reality. It is a pain in the ass to "zero out" Rice as a CF so he doesn't play it and adjust Lynn to make sure he's the CF. Was kind of hoping that the game would auto figure this out by reading these stats for 1975:

CF Starts
144 Lynn
0 Rice

LF Starts
90 Rice
0 Lynn

RF
0 Rice
0 Lynn

DH
54 Rice
0 Lynn

You would think that it's not too hard for the program too read those starts as a sign of the positions they played, and the DH aspect as a sign that Rice wasn't a very good LF let along a CF.

I don't even want to get into noticing halfway through 1976 that O's rookie Eddie Murray was starting in CF (with a completely insane +10.9 ZR in 647 IP!!!!).

Anyway, I digress on the game having major positional issues when playing a historical career replay, which you would think it wouldn't given it's checking the real data every year.

So on the trade/free agent issues when the team already had a good-to-great player at the position:

It happens. Much too often. The AI is not too smart. It kind of screws up historical replays, especially when you don't have the time (or desire) to play good over every team and every player to avoid this stuff. I tend to play to have some fun after long days of work. Having to "work" on 23 teams other than mine, and a 700+ players... don't have the time.

[NoOne]

for the positional stuff, submit to proper forum or individual.. whichever it is that handles that. that stuff is being improved each year. it's a %$@-ton of players to go over.

[smiller]

This is why I never make a trade, no matter how good it sounds, before checking the lineup of the team I'm trading to. If they already have a good, productive player in the position of the player I want to trade, I cancel the deal or try to trade someone in a position where the need is greater. And this is with trading set on very hard.

[Rosco Peabody]

See, the AI's logic sometimes is flawed, but I also understand that it's nearly impossible to create a perfect AI. My problem is not the mistakes it makes. My problem is that there aren't any easy ways to fix/prevent it.

That's why I suggested a "Force start" button or something of the like that you could activate even for AI teams in Commissioner mode.

We already have it for the teams that we control, I would like it if I could switch over to control a different team, check a "Force start at X" box for a particular player on their player strategy and overwrite what the AI for that team wants to do.

Obviously this is playing a little bit in god-mode, but that's essentially what Commissioner mode is for. I can force trades however I want (sometimes I will force a trade between two teams that I don't even control because I think it makes sense for both of them in my universe), why wouldn't I be able to control how they use their players in a particular scenario?

[smiller]

I agree. For example, in my fictional league the manager of the team my team was playing pinch hit for the league leader in home runs and RBIs during a game where the player had already hit one home run. The pinch hitter struck out.

[Silfir]

I don't think putting in ways to more effectively police the AI is the right way to go about it - once you're forcing it to start a player at a certain position, are you going to periodically check to make sure that player hasn't imploded? Where does it end?

I often sign players in free agency whose positions are already filled, if I can get them on a really good contract. That gives me pieces I can use to trade for players in positions I do need. When it comes to pitchers, losing two starters might result in some relievers with borderline starting abilities moving up in the rotation, and therefore in a shortage of relievers.

All I'm saying is: The criteria for judging that the AI should no longer be allowed the agency of its own decisions should be very strict, if only for sanity purposes. Personally, if I judged that to be the case, I'd wait for OOTP20 to fix the issues; I'm really not interested in playing against myself. (Or other humans for that matter. It's not the humans I mind, just having to stick to schedules like a proper organized person.)

EDIT: My powers of rationalization are very well-honed, admittedly. In smiller's example:

For example, in my fictional league the manager of the team my team was playing pinch hit for the league leader in home runs and RBIs during a game where the player had already hit one home run. The pinch hitter struck out.

I would assume my manager either had some concerns about L/R matchups or just had one of those funny hunches that lead managers to make questionable decisions - like, maybe he looked at that player's face as he was going to the on-deck circle and just knew he wasn't up for it. At worst, I probably would have a grand old time imagining the meeting in which the idiot gets the sack, and replace him with a manager with a different tactical profile. (I do the same with sacrifice bunts, after all. I think they're stupid, so I try not to hire managers who use them - unless they've just been too successful with it. After all, if it's stupid but works, it can't be stupid.)

To conclude with certainty that this decision is a bug takes knowledge that's just not accessible to the players - it should be very rare to find a decision that's 100% indefensible. If there's a singular line of reasoning, even if it's just "real baseball people also make dumb decisions for dumb reasons", I'll take it.

[Rosco Peabody]

I agree that trying to police in-game decisions is way too much, but I've gotten frustrated in my fictional league at this very scenario that has happened several times. Here are a few examples:

SCENARIO 1: A small market team that had a bunch of talent coming up in the minors, including at first base, signs a big money, aging veteran free agent first baseman that really doesn't fit what they're doing. Before the season even starts, I shop a AAA 20 out of 20 player and they offer this big name guy they just signed because they view him as a behemoth contract they can't afford

SCENARIO 2: A team in desperate need of a Catcher signs a big name free agent Catcher for big money. A short time later, I am shopping my own big money Catcher (one who isn't as good though) to try to create space in my budget. They offer me a good deal and I take it, forgetting they already have a good one. They take my Catcher and immediately bench him, and now have a 16 million dollar backup sitting on their bench.

SCENARIO 3: (and the one I think is most easily preventable) The AI has a hall of fame level player on their team who has been with the team for years, let's say a LF. This LF dips one or two OVR points below another guy on the roster, let's say to a 44 OVR while the other LF is a 46 OVR and a second year player. The AI benches the Hall of Famer because, on paper, the younger guy is technically at this point "better". But when would this EVER happen?? They would ALWAYS defer to the legend

[Silfir]

In scenario 3, I'd love to know a) the batting stats at the time of it happening and b) was the hall of famer in a 44 overall / 80 potential kind of situation? That one signals an ongoing decline, and it's quite possible that the AI's scouting director is already projecting him to be in a much worse state than you're seeing. Either way, if players can bench hallowed individuals without significant repercussions, it would be unfair to force the AI not to.

[Rosco Peabody]

Originally Posted by Silfir

That one signals an ongoing decline, and it's quite possible that the AI's scouting director is already projecting him to be in a much worse state than you're seeing.

This is exactly what I'm talking about. What I'm seeking in my league is realism. Even if the Hall of fame player is in a state of decline, in real life, they would still be started out of respect. Or if they were a starting pitcher, they wouldn't be made into a middle reliever

[jdw]

I agree with Rosco's point.

I had the Pirates sitting Stargell in 1977 for Mitchell Page. I get that - Pops was bad defensively, Page had good ratings, makes sense.

Where the problem came in is that the Pirates didn't DH Pops. Instead they went with Lee May, who frankly they never should have traded for in the first place given a backlog of players in the majors and minors at 1B. Pops had far better ratings than May. The AI wasn't even smart enough to do a platoon out of the two.

I shouldn't care, but Pops chasing 500 HR is he stayed healthier in 1974-77, and DH'd his way into the line up more often in 1978-79, was one of the things that interested me in the replay. 4 HR in 1977 killed that dead.

[Rosco Peabody]

And if a guy was that close to 500 HRs with his home team, then they would definitely start him


Along with that point, sometimes the "popular" numbers bug me. If a guy gets to legendary status, his popularity should have a floor that it can get to. Even if he stinks in the current season, if he has put up legendary numbers, he should never dip below "popular" or "very popular" again. Or when he retires, there should be a separate value for the career popularity. It shouldn't be listed as how popular he was at retirement