19th Century Play

[Spritze]

Fellow 19th Century Players:

It is possible, by adding 19th Century player movement to the new historical transactions database, to re-create all the teams that come and go and get rid of the necessity of releasing players by hand when teams cease the existing part of their lifespan and the players wander over to play for someone else.

This would also allow OOTP to get rid of the fake teams it is packaged with and possibly create much more realistic league setups.

I know there are lots of work-arounds necessary currently and I'd like to have us delineate them for Markus and propose simple solutions like adding the 19th century transactions noted above.

The only hard-coded requirement I know of is the 2 league, even number of teams thing. I'm hoping that might be the only thing that needs programatical changing. Getting actual teams and leagues would entail merely using an unfiddled with teams.csv available from Lahman. It would be nice to add AA and NA league abbreviations to the already functional AL, NL and FL ones currently available. I believe actual schedules that could be packaged with OOTP are available for all these years and leagues.

More?

[David Ball]

My mistake, I see the question I asked is addressed in the original post.

[Biggio509]

I have not tried auto expand in 12 but I assume not much has changed since 11. The only other thing I would add is a push for more playoff options between leagues. That way you could handle the AA and NL better as separate leagues that could have separate rules but be able to schedule a world series that is saved in history.

The only thing that slightly annoys me is before 1883, the winner was determined by most wins and not winning percentage. Not sure how hard that is to implement but I hate when 6-0 Baltimore Marylands win the NA because they have the highest percentage or some other that played less than half the schedule wins on a high percent.

Hmm, wonder if deleting the team in season would eliminate that? I do remember way back I did delete the teams after the last game on the schedule, well those that were going to fold, so other teams could sign their players. Don't remember if the record was retained for that season and counted in the standings.

[silentrob]

I'm not sure if this is what you had in mind when you started the thread, but if you could give each team an ID, would that allow for the inclusion of accurate ballpark factors that update for 19th century teams?

[Spritze]

Originally Posted by silentrob

I'm not sure if this is what you had in mind when you started the thread, but if you could give each team an ID, would that allow for the inclusion of accurate ballpark factors that update for 19th century teams?

Yes, each team has an ID and ballpark factors already. They are currently ignored by OOTP due to the fake team thing Markus and Garlon put into play.

I'm also thinking Roster Sizes will be an issue as at the default 25 there won't be enough players. We will need Markus to use semi-real roster sizes for the 19th century. Currently 15 is the minimum. I'm thinking 15 for the 1870's, 18 for the 1880's, 21 for the 1890's, something along those lines?

[silentrob]

Was also thinking an injury setting specific to the 19th century. But probably there is not much data to work with, so it would just be educated guessing.

[scott1964]

Maybe turn injuries off up until 1900.

Roster size might be just about right. My reserve rosters have around 20 to 30 players with each team in 1898. Of course I let the players play, I do not use retire according to history or anything like that. I have some player get a shot at stardom, like Park Swartzel, in real life he played in one season. In my dynasty thread he has won over 200 games.

[Spritze]

I've made a 19th Century Injury File I use that keeps injuries to 2 week maximum.

[David Ball]

Before the mid 1880's or so, some teams went through an entire season without using fifteen players. Boston in 1877 comes to mind.

[scott1964]

In my dynasty thread , it looks like teams fielded 18 to 22 in the early 1880s to the mid to upper 20s in the later years of 1880's.

[Le Grande Orange]

Originally Posted by Spritze

I'm also thinking Roster Sizes will be an issue as at the default 25 there won't be enough players. We will need Markus to use semi-real roster sizes for the 19th century. Currently 15 is the minimum. I'm thinking 15 for the 1870's, 18 for the 1880's, 21 for the 1890's, something along those lines?

Originally Posted by David Ball

Before the mid 1880's or so, some teams went through an entire season without using fifteen players. Boston in 1877 comes to mind.

Below is a table containing the minimum, maximum, and average number of players to appear on a ML club's roster during the season (that is, the player played in at least one game with the team) over the 1871-1900 time period.

Year  Min  Max    Avg
---------------------
1871   11   18   12.8
1872   10   26   14.2
1873   12   19   13.9
1874   11   23   15.4
1875   11   35   16.8
1876   10   24   15.5
1877   12   21   16.0
1878   10   16   13.3
1879   12   21   15.9
1880   11   23   16.9
1881   11   26   16.5
1882   12   24   17.9
1883   11   29   17.6
1884   15   35   21.4
1885   13   32   22.1
1886   15   37   21.9
1887   16   33   20.8
1888   17   26   21.9
1889   15   29   21.4
1890   17   46   25.1
1891   16   38   24.0
1892   17   37   25.8
1893   18   31   23.8
1894   17   30   24.5
1895   20   40   25.4
1896   20   35   24.6
1897   19   33   23.3
1898   20   37   26.8
1899   23   41   28.8
1900   21   30   24.4

So, for example, in 1888 the lowest number of players to be used by a team was 17; the highest number of players used was 26; the average number used was 21.9. If you were to strip out those players who only played in a small number of games, the maximum and average values would no doubt be lower than shown in the above table.

[Spritze]

Ok then 15, 18 and 21 on a roster it is. That nicely covers LGO's minimums.

[silentrob]

You could also make the rosters progressively larger, like every 3 or 4 years allow one more player.

[David Ball]

If you play with real transactions, so that some teams won't carry the fifteen-player minimum, is the game going to refuse to go forward until you fill the roster up to fifteen? I know some people do create games with real rosters, so maybe this isn't a problem.

At any rate, the chart LGO gives are totals of players used throughout an entire season. I don't think anybody in the '90's was ever carrying 21 men at any one time. Certainly not early in the decade, when for a little while there was a 13-man player limit. During the late '80's a few of the fat cat teams may have carried 18 --the Giants always seem to have had very large rosters-- but seven regular, three batteries and maybe a fourth pitcher or a utility player was about it for a poor team, and probably a lot of the others. In the '70's thirteen men would have been extravagant. I actually think the fifteen-man limit would work for most teams well into the twentieth century. The largest rosters in the late '80's were probably bigger than any in the '90's, because farming was practiced much more heavily in the '90's and you could control a player without having to pay him to sit on your bench and not get any better.

From a purist's point of view the problem is that, with brief exceptions, there was no player limit until some time in the twentieth century. For about a decade the limit on the number of players you could reserve at the end of a season acted as a kind of soft cap, but even that was removed during most of the 1890's, when clubs were known to grow so careless about the matter as to reserve dead men (well, it happened once that I know of). Today every team carries 25 active players and has a full 40-man roster, but in the 19th century both the number of players in uniform and the number controlled by what would be equivalent to our forty-man roster varied from team to team, so that no strict limits are really appropriate except what club management could afford and wanted to pay for.

I actually don't want to be a purist about it, but I do think that as you give the AI manager more players, it will probably become more inclined to practice aggressive substituting in a way that wasn't possible under the rules before around 1890 and even during the '90's wasn't really fashionable to the extent the AI wants to do it.

[Le Grande Orange]

Originally Posted by David Ball

At any rate, the chart LGO gives are totals of players used throughout an entire season.

That's correct. It is the total number of players to have played at least one game with the club. The figures are only meant as a rough guide to the frequency with which players were used by a club.

I gathered the data by looking at the full team rosters as given at Baseball-Reference. The reason I took all players was because (a) it was easier, and (b) I wasn't sure what a good cut-off level would be for participation. Should players with less than 10% of the number of games played by the club be excluded from the count? Or 20%? Or maybe it should be a set number of games, e.g. 10 games or less and the player isn't counted, or 15 games or less.

If someone wanted to make some recommendations as to a good cut-off level for excluding players with limited playing time, then the figures could be recalculated resulting in something closer to main player usage.

[Spritze]

Since 15 is the minimum number of players the AI needs to not make really weird strategical decisions that is what we need to work with. Keep in mind OOTP does not allow actual 2-Way players and there were a gob and a half of them back in the days we are discussing. So a few extra players will be necessary.

I am hoping to simply ask Markus for as few as possible programming changes to allow for semi-realistic 19th Century play, basically just to automate the work-arounds players are already using. The work-arounds work because OOTP can already do those tasks, it just doesn't. I think that will be a good first step.

[David Ball]

That certainly makes perfect sense as a general policy, but if you have to assign fifteen players to teams that never had that many, how can you do real transactions? Will you fill in with players who were out of the major leagues at the time?

[Spritze]

There are no real transactions per se, only real player movement. The 19th Century had no real contracts either. Players basically played wherever they got the most florins.

I have tested OOTP as is with 8 teams and a 15 player roster works. Barely in 1871 but better as time wanders on.

[David Ball]

Originally Posted by Spritze

There are no real transactions per se, only real player movement...

That I understand, but I must be confused elsewhere. My understanding is that the historical transactions are essentially designed the produce the roster the real team had that season -- in principle, on any date in the season, although the dates are not always easy to specify. I don't think anybody will be buying OOTP just to acquire a register of player transactions; the historical transactions file is simply the means to an end, the end of producing the proper roster, so that the team in question has all and only those players available that played for it in real life.

Now, we know some teams actually used only ten or eleven or twelve players in a season, and we know the OOTP minimum is fifteen. How can the historical transactions for those clubs produce a roster containing all and only those players that did appear for them?

[Spritze]

Originally Posted by David Ball

My understanding is that the historical transactions are essentially designed the produce the roster the real team had that season -- in principle, on any date in the season, although the dates are not always easy to specify.

Nonfortunately that information is mostly not available for the 19th Century. That is why the transactions added to the tranny db will mostly take place on 1/1 of each year and simply move the player to his next destination.