Actual Lineups and Transactions??? Not Really

[Dodger1]

I usually play from the current season forward, but decided to try a historical season. I picked 1953 and checked both historical transactions and actual lineups. A week or so into the season, I noticed some guys I knew played that season were not in lineups. I checked, and every team had multiple players on the dl, and EVERY ONE of them was for 255 days, essentially the entire season. Since I was playing with reserve rosters all those guys actually played in that season, several over 100 games! How, exactly, is that “actual lineups “?

If that wasn’t enough, I noticed Ernie Banks playing shortstop for the Phillies??? Checking his record I saw on April 12 he was rated the #1 prospect in baseball whhile with the Cubs. The very next day, April 13, the Cubs RELEASED him, and he was signed by the Phillies. Seriously, would ant team release a 22 year old shortstop rated the #1 prospect? What kind of ai is running these historical leagues?

Then I tried 1988, managing LA with historical transactions, but not lineups. In early May Don Sutton went down with a season ending arm injury after 4 starts. Thst seemed wrong, so I checked. Sutton did in fact have a seaon ending arm injury, but in August after 12 starts! Again, how is that “actual historical transactions”. I’m really disappointed by these replays.

[pstrickert]

1 - Some players did not appear in a game until later in the season. Banks is one example. So, in order to keep Banks from being on the Cubs roster the entire season, he has to be assigned a "phantom" injury before Opening Day and placed on the DL. Don't worry about the 255 days. He'll be activated when he needs to be, later in the season. The same is probably true for the other players you noticed. It is an inelegant solution, but the only one we have available.

2 - Banks should not be released. I'll have our transactions database guru take a closer look to see what might be happening.

3 - Sutton should not suffer a season-ending injury after four starts. Did you enable injuries by chance? Injuries should be disabled when historical transactions are used. That said, I noticed the same issue with Sutton being released. Again, I'll ask our transactions database guru to look into it.

[pstrickert]

Update: Sutton was released in August 1988 in real life, so that's no problem. Banks was a Negro League player before signing with the Cubs in September 1953. So, he was technically a free agent. However, if you enabled historical lineups and transactions, you should see no AI roster moves (such as the Phillies signing Banks as a free agent). I contacted Kim (the aforementioned guru). He'll tell us what's what before too long, I'm sure.

[Dodger1]

I understand what you’re saying about Banks, but why aren’t guys who aren’t on the regular 25 until later in the season simply placed on the reserve roster if actual transactions are in place? No need to dl them. And, again, all the players (at least two for every team) on the dl with 255 days listed actually played. Why not just put them all on reserve until they’re activated.

[pstrickert]

The AI has access to the reserve roster. It can shuffle players, especially marginal players, between the active and reserve rosters. It needs access because once in a while even a marginal player will start a game. If Banks were placed on the reserve roster, the AI could and probably would activate him, in order to have a good bat on the bench. Kim can tell you more, but the reality is that this is the best solution at our disposal. IRL, Banks played 10 games for the Cubs in 1953. If the AI has access to him, it'll use him. By season's end, he'll probably have an extra 100 ABs or more than IRL. Then we'll hear (and have heard) complaints that having Banks available the whole season, even as depth, ruins the replay. Damned if you do, damned if you don't.

[M's rule]

Originally Posted by pstrickert

1 - Some players did not appear in a game until later in the season. Banks is one example. So, in order to keep Banks from being on the Cubs roster the entire season, he has to be assigned a "phantom" injury before Opening Day and placed on the DL. Don't worry about the 255 days. He'll be activated when he needs to be, later in the season. The same is probably true for the other players you noticed. It is an inelegant solution, but the only one we have available.

2 - Banks should not be released. I'll have our transactions database guru take a closer look to see what might be happening.

3 - Sutton should not suffer a season-ending injury after four starts. Did you enable injuries by chance? Injuries should be disabled when historical transactions are used. That said, I noticed the same issue with Sutton being released. Again, I'll ask our transactions database guru to look into it.

Guru here. I don't see anything in the transactions file that deviates from what actually happened to Banks and Sutton.


(I started writing this before the latest flurry of posts, so it might be a little out of date by now.)



1 - If you start a league in 1953, the game opens on April 12, one day before Opening Day. Because Banks imports as a rookie during this season, the game puts him on the Cubs roster. But in real life he was signed as an amateur free agent in September. So there's a line in the transactions file that makes him into a free agent (No Team), which is accomplished by releasing him from the Cubs.



In a league that uses real transactions "out of the box," without any options that might not play well with that feature, a free agent will not be acquired by any team other than the one that actually signed him. I can't imagine how Ernie ended up being signed by the Phillies, unless there's something else going on under the hood.


2 - Ernie's story for 1953 ends when he joins the Cubs roster on September 17 and starts the last 10 games of the season at SS. He was actually signed on September 8, but the transactions database advances the date to the 17th so that he doesn't appear in a game before the day of his actual debut.



3 - It looks now like Sutton was not actually a problem for Paul, but in Dodger1's league be suffered a season-ending injury in May. I don't see how this could happen, unless (for instance) the league also has injuries (of the fictional sort) enabled. In real life Sutton didn't pitch after June 28 until he came back for one final start on August 9, and he was released by the Dodgers on the following day.


What Paul says about the reserve roster is all too true. The AI has control over who's on the active roster and who's on the reserve roster. In the extreme cases, such as Banks, historical accuracy suffers if these guys are allowed to be available all year. If Ernie starts 1953 on the Cubs roster, the chances are excellent (for obvious reasons) that the AI will find a place for him on the active roster. And, conversely, someone who should have been on the active roster instead of him will start the season on the reserve roster.



He still won't start a game until September 17, because that's what the lineups file dictates. But between April 13 and September 16, he will get a whole lot of opportunities to enter a game after it's under way. That's undesirable, because (a) it takes plate appearances away from players who should have gotten them and (b) it never happened.


Any more questions?

[pstrickert]

Excellent explanation, Kim. Thank you.

SOAPBOX: It might help to remind people that OOTP is not limited to season replays (unlike many other baseball sims). The game allows the user to play many different ways and do all kinds of things. But it also means the development team has to parcel out its limited time and resources. Much of what gets done is due to unpaid volunteers. Take the transactions database, for instance, which contains hundreds of thousands of entries, if I'm not mistaken. Markus had to program the game to use the database. But creating, correcting, and maintaining the database (a painstaking and extremely time-consuming effort) is just about all Kim Mohan. A volunteer. He deserves a ton of credit for this gargantuan achievement.

[Ben75]

I posted a thread close to this topic, in the general discussion forum but it didn't receive a reply. Based on what I'm reading here, it sounds like when you enable historical transactions and lineups, this doesn't mean that the 40-man and 25-man rosters would necessarily be correct on opening day...right?

I was trying to figure out a way to make this happen by manipulating the settings but kept noticing discrepancies. For example, when using reserve rosters I was seeing players who should have been in one team's minor league system (for the whole year) start out in the free agent pool, but when enabling the minor leagues they showed up where they should, but then I saw other players who never played in the majors that year on the major league roster for certain teams.

I assume this is simply because we don't have complete records as to 40-man or day-to-day roster moves so the AI is directing this...or is there a way to start the season with accurate 25-man and/or 40-man rosters?

[pstrickert]

You're correct. We don't have historical 25-man or 40-man rosters. No minor league transactions. I can't imagine trying to obtain reliable records, if they even exist, and incorporate them into the database. I'll bet someone could spends months scouring online sources, hoping to capture a single season's worth of roster moves, down to the last detail. But who wants to try it?

[M's rule]

Paul, that pretty well sums it up. Thanks.


Ben75, your experience seems to be evidence to support the "fact" that using historical transactions and actual minor leagues at the same time produces unrealistic results.



I just had a similar experience in a 1953 league I started, using historical transactions and lineups plus minor leagues and Negro leagues. I didn't change any of the defaults for a historical league, so "missed seasons according to history" was enabled. But when I looked over the replay results, I found out that Jim Fridley appeared in 48 games for the Browns. In real life he didn't play in the majors during 1953; he actually spent the year in the Browns farm system. Draw your own conclusions. . . .

[Dodger1]

Thanks. I appreciate the quick responses. I’m still kind of confused how Banks came to be released the day after he was listed as the #1 prospect. Actual transactions aside, what kind of ai would do that?

[drowssap]

So much this!!

I am playing with the terrible 1985 Pirates. So many discrepancies. Examples:

George Hendrick, Tim Foli, Johnny LeMaster, Jerry Dybzinski, Jim Morrison, Marvel Wynn: 0 games played, in minors.

Chris Green, Ray Krawczyk, Bobby Miscik, Mike Diaz, Barry Bonds: all in MLB even though they were either never on the roster in '85 (Bonds, I'm looking at you) or played very limited amount of games.

I have played the '85 team in other seasons of OOTP and it was never this bad. Can't change it because I am in Challenge Mode.

[pstrickert]

@drowssap

Did you enable historical minors? Historical minors is not recommended in conjunction with historical lineups and transactions.

With historical lineups enabled, the AI is obliged to use those lineups. But "historical lineups" is not the same as historical 25-man and 40-man rosters (as mentioned above). So, the AI can move bench players and relief pitchers to and from the reserve roster. This is normally not a problem. Most "scrubs" did not get much playing time in real life and they won't get much playing time in OOTP.

Problems occur, however, when "historical minors" is enabled. Now, the AI has access to all minor league levels and all the players in the organization at those levels (even if they didn't appear at the major league level that season). So, the AI may put Bonds on the major league roster as a bench player, even if he never appeared in a game in 1985. He won't start, if historical lineups are used. But he could pinch hit or appear as a defensive replacement.

This situation can confuse the user who does not expect to see these players on the major league roster. For this reason, it is recommended NOT to enable historical minors when using historical lineups and transactions. I hope that helps.

By the way, I simmed the 1985 season (without historical minors) and everything looks fine.

[pstrickert]

Originally Posted by Dodger1

Thanks. I appreciate the quick responses. I’m still kind of confused how Banks came to be released the day after he was listed as the #1 prospect. Actual transactions aside, what kind of ai would do that?

The AI did not actually release Banks. He was released because the historical transactions file triggered his release, making him a free agent. The transactions file will then trigger his signing as a free agent with the Cubs later in the year. The default settings for historical leagues should NOT allow any other team to sign Banks while he's a free agent. So, both Kim and I don't understand how the Phillies were able to sign Banks in your replay.

[M's rule]

For the record, I agree with everything Paul says above.


Dodger1, I'd be interested to find out what your settings were in the league where Ernie got signed by the Phillies.



Did you enable a feature that is disabled by default, or did you disable a feature that is enabled by default? I think one of those two things must be happening.



And if you played through the 1953 season, what happened to Ernie on September 17, when he is scheduled to join the Cubs?

[drowssap]

I just looked and I do not have historical minors enabled.

Originally Posted by pstrickert

@drowssap

Did you enable historical minors? Historical minors is not recommended in conjunction with historical lineups and transactions.

With historical lineups enabled, the AI is obliged to use those lineups. But "historical lineups" is not the same as historical 25-man and 40-man rosters (as mentioned above). So, the AI can move bench players and relief pitchers to and from the reserve roster. This is normally not a problem. Most "scrubs" did not get much playing time in real life and they won't get much playing time in OOTP.

Problems occur, however, when "historical minors" is enabled. Now, the AI has access to all minor league levels and all the players in the organization at those levels (even if they didn't appear at the major league level that season). So, the AI may put Bonds on the major league roster as a bench player, even if he never appeared in a game in 1985. He won't start, if historical lineups are used. But he could pinch hit or appear as a defensive replacement.

This situation can confuse the user who does not expect to see these players on the major league roster. For this reason, it is recommended NOT to enable historical minors when using historical lineups and transactions. I hope that helps.

By the way, I simmed the 1985 season (without historical minors) and everything looks fine.

[thehef]

Originally Posted by pstrickert

We don't have ... minor league transactions. I can't imagine trying to obtain reliable records, if they even exist, and incorporate them into the database. I'll bet someone could spends months scouring online sources, hoping to capture a single season's worth of roster moves, down to the last detail. But who wants to try it?

Originally Posted by M's rule

Ben75, your experience seems to be evidence to support the "fact" that using historical transactions and actual minor leagues at the same time produces unrealistic results.

Actually, this would not be as gargantuan an effort as it would seem. There was some great discussion in the beta forum on this a few years ago, with the both the late, great Spritze and BigRod seeming to indicate that a minor league transactions file would be very do-able.

The keys to it are accepting that 1) determining the nature of each transactions (trade, release, signing, draft, etc.) is not necessary - it would simply be a Minor League Player Movement file, 2) precise dates - while useful when they are known - are also not required, and 3) coming up with a semi-plausible history for - as an example - a low-minor-leaguer who merely had cups of coffee with multiple teams simply isn't a high priority (these guys are going to barely play anyway, due to the adjust and weaken settings)

Basically, the file would make ensure that every player is in the correct organization. When you factor in that the overwhelming majority of player-seasons are comprised of players who spent the entirety of a season with a single organization, then you have a very simple mechanism for creating txns when those players change organization. For example, if Johnny Rockstar played his first three seasons, 1960-1962, with the Dodgers org, then 1963 with the Red Sox org, and 1964-1966 with the Tigers org, all you'd need are these three txn entries in the file for him:

1/1/1960 - signed by Dodgers
1/1/1963 - signed by Red Sox
1/1/1964 - signed by Tigers
... and then if Johnny's baseball career ended after the 1966 season, you could have...
1/1/1967 - released by Tigers

Where it then can get tricky is where you have a players who played for more than 1 team in a single season. However, due to stints (the order - 1, 2, 3... - in which a player played on multiple teams in a season) being known for most players, and with games played per team being mostly known, reasonable estimates as to the dates of the actual transaction can be made.

What also reduces the volume of minor league txns needed is that for those players who made the major leagues, their minor league txns are mostly known. For example, even though Reggie Smith didn't make his major league debut until 1966, the existing historical txns file "knows" that we was originally signed by the Twins on 6/21/1963, and was then purchased by the Red Sox on 12/2/1963.

So what we might end up with is a file as such:

a) For those who made the major leagues, we'd have what we have now - a very accurate major- and minor-league transaction history.
b) For those who never made the major leagues, we'd have a very accurate minor-league history for the majority of player-seasons (where players played full seasons with a single organization).
c) Where the accuracy would be reasonable but somewhat less than very accurate would be where players played for multiple organizations in a season, and where their stints and game played for each team are known.
d) Where the accuracy would be somewhat of a crapshoot would be where a player only played a few games in a season (we wouldn't know when he was signed & released), or where a player's stints were not known (we don't know which team he played for first, then second, etc., in a given season).

That would be a pretty dang good file, though, if you ask me.

How do-able might this be without Spritze? That, I do not know. Perhaps BigRod might be able to speak to this. But it IS do-able as long as the goal is to get reasonably accurate minor league player movement. Given that precisely accurate MiLB player movement is not attainable, reasonably accurate would seem to be a reasonable goal, given the alternative, which is what we have now: no minor league transactions.

And the thing is, SO much effort was put into introducing minor leagues in to OOTP. Yet, it appears that few gamers use them. I'm guessing part of the reason for that is that OOTP is so rigid in how they can be used. For example, if you want historical minor-league players, you MUST accept OOTP's version of historical minor league structure, but nothing else can be historically accurate. Time to fix that

[thehef]

Originally Posted by Ben75

I posted a thread close to this topic, in the general discussion forum but it didn't receive a reply. Based on what I'm reading here, it sounds like when you enable historical transactions and lineups, this doesn't mean that the 40-man and 25-man rosters would necessarily be correct on opening day...right?

I was trying to figure out a way to make this happen by manipulating the settings but kept noticing discrepancies. For example, when using reserve rosters I was seeing players who should have been in one team's minor league system (for the whole year) start out in the free agent pool, but when enabling the minor leagues they showed up where they should, but then I saw other players who never played in the majors that year on the major league roster for certain teams.

I assume this is simply because we don't have complete records as to 40-man or day-to-day roster moves so the AI is directing this...or is there a way to start the season with accurate 25-man and/or 40-man rosters?

I can't speak to what it would take to have accurate 25- and 40-man rosters, other than to guess it would be very difficult. However, regarding the part in bold green, that is true, but that's not the way it has to be.

Right now the historical txns database includes phantom transactions to release, for example, a player who played for the big league club one year, but spent the following year in the minors. I understand why this transaction exists in the database, but it is also one of the hurdles to being able to play with both historical transactions and historical minor leagues. The solution is a simple one.

Not very long ago, historical injuries were added to the historical transactions file (I believe this is an ongoing effort by Kim that has come a long way recently). For a short time, it was an all-or-none proposition: You want to play with historical transactions, you had to accept historical injuries, too. However, when it became apparent that many gamers wanted the transactions that moved players from team to team, but did not want the historical injury transactions, Markus included a checkbox to include or exclude the historical injury items, which was perfect. So the solution here would be very similar. Basically, one of the following:

a) A checkbox to include or exclude the phantom transactions, or...
b) Logic that ignores these phantom transactions if historical minors are enabled

Simple. And IMO anything that makes historical - and, specifically, historical minors - more-flexible should be considered.