Simulating Deadball Era

[CrashDavis]

One of the most difficult things to do in PureSim is simulate the deadball era. Using the default settings with game simply produces wild numbers.
Late last year, I decided to undertake an experiment to simulate the seasons 1903-1919 and see what can be done to adjust for the era.
My first stop was to the Baseball Reference website to collect each seasons league totals for hits, doubles, triples, hr, and runs. For pitching
stats I collected league totals for k and walks. This would give me something to compare on a season-by-season basis with what Puresim produced.

Next I had to decide on the league structure. I decided to use:
Puresim v1.71
Lahman55 (w/ patch for PS v1.71)
Begin with 1903 Season using real players
40 Man roster
Finances off
AI Trading off
Injuries on
Early Entry off
Let AI handle draft, roster moves, tendency settings

I also used the default puresim.xml file as I was not willing to modify that. And, no player ratings were modified - the ratings were used as imported
(with the exception of Babe Ruth who I changed from a pitcher to a hitter in last year 1919).

Now, looking over the stats that I gleaned from the Baseball Reference site, I found something kinda interesting. I interpreted what I was looking at to mean
there were two deadball era's between 1903 and 1919. Right in the middle, between 1911-1913, the stats took a decidedly upward jump during this period. I
have no explanation as to why that happened. But stats don't lie...do they? Was there a rule change in baseball? Was it aluminium bats? Steriods?

Next I had to gather all the ballparks that were going to be used during this period. This took some time to do. And some had to be created - Columbia Park,
Robison Park to name two. Many of the parks during this period came from the Classic_Parks zip file from PadresFan site. And a few odd ones can be found
inside PureSim itself.

Now it was time to experiment with the association adjustment sliders on the options screen. I halved all the offensive sliders and left the pitching
sliders alone. I kept experimenting with those until I found what worked. Although visually, they looked goofy at where they were set.

As an experiment, I simmed several seasons. One of the first things I noticed was that HR totals were way too high. I knew it couldn't be caused by the
HR slider adjustment. So, I took a closer look at the ballpark dimensions. The dimensions were set correct according to how they actually appeared.
Problem was, alot of those parks are too small. I swear, my niece could have hit a HR in some of those parks. How grown men didn't hit 40+ HRs is
a mystery. Of course, there were other ballparks where the fence was so deep, you'd need a howitzer to clear it.
So, in the end, to be fair, I decided to mandate a minimum distance in the left/right foul line of no less than 290' and Left/Right field fence of no less
than 325'. Anyone wanting any of the ballpark dimensions used can contact me and I'll get that info together for you. I would post them, but even the
zipped files are too large for my modest webspace.

After some final tweaking I was ready to begin. Between each season I would compare the season stats with those from the Baseball Reference site to see how
close I was.

Here is a link to the Real MLB stats from those seasons (1903 to 1919):
Real MLB Stats

Here is a link to the stats produced by Puresim for same period:
PureSim Stats

Here is a Hitting Comparison Chart showing both stats on a year-by-year basis:
Hitting Chart

Here is a Pitching Comparison Chart showing both stats on a year-by-year basis:
Pitching Chart

HRs, while within reasonable bounds, was a bit erratic from year to year. And pitcher k's I should probably have been more aggressive when backing down the
adjustment slider.

My final adjustment slider settings look like this:

PureSim Sliders

Deadball Era ('03-'10, '14-'19)
Out 52
Hit 48
Walk 48
SO 50
1b 60
2b 25
3b 50
HR 20


Mini-Live Ball Era ('11-'13)
Out 55
Hit 54
Walk 50
SO 67
1b 60
2b 35
3b 55
HR 66


Live Ball Era ('20-Present)
Out 50
Hit 50
Walk 50
SO 50
1b 50
2b 50
3b 50
HR 50

With regard to the HR totals, I do wonder if some that discrepancy isn't because alot of HRs in the MLB at that time were inside-the-park HRs? As far as
I'm aware, Puresim simulates a HR only when the ball clears the fences. Has anyone had an inside-the-park HR ever?


CrashDavis

[KG Erwin]

Thanks for your research work, Crash. I'd considered starting a deadball era association, but apart from the XML settings, the weird parks of that time are the biggest bugaboo.

This is part of why I think the default settings work best with post-WWII (roughly, "the modern era") historical associations.

As for ever seeing an inside the park HR, no, I've never seen it occur in-game. Again, the combination of ballpark dimensions and hard-coding might prohibit it, perhaps for good reason.

[Wrathchild]

CrashDavis,

That's a great spreadsheet. The association I'm playing started in 1900 so this information is very pertinent to me. I've also tried tweaking some of the sliders to see if I can make it reflect more accurately the historical data. I would be interested in checking out your ballpark info, as that is an area that I've had much trouble with, as far as trying to reconcile the real world data with the various entries into Puresim. Part of it is that the data available doesn't necessarily correspond nicely to the Puresim settings, i.e. I'm not sure which real data should go in which entry for the ballparks as PS includes more distances than most websites show for the parks. The park that has really caused me grief in my association is the one used by the Boston Beaneaters. It seems that this park had a rather short right field distance (I currently have the farthest right dimension set to 255'), which allows that team to get what I consider to be a skewed number of home runs.