Spritze/Garlon DB

[greendiver]

Excellent. In that case, I'll hold off starting until the new database is released. I appreciate all your work. Thanks.

[Spritze]

Your welcome.

I hope to send the db out today but real work keeps getting in the way.

[Spritze]

AllSim is just a maybe as I couldn't figure out how to submit something so I e-mailed it.

[Spritze]

This DB was created to work in conjunction with some changes Garlon is requesting to the OOTP9 historical mode. I have used it with the current version successfully but there are some things I see after cursory use I would like to change for the next build of version 2 to make it even better for the current version (OOTP8/2007).

This CSV creation db was put together modularly so it is now extremely easy to change and alter. What that means is that nearly any kind of Lahman style db can be easily created based on nearly anything anyone wishes to base one on. Team best players, only players from Ohio and Indiana, etc. So if anyone wishes a specific Lahman style db just let me know and give me your e-mail address and a weeks time. It only takes an hour or so to create any type of Lahman style CSV files but I only have 2 or 3 free hours in a week to do this. Busy bee, busy bee.

[f.montoya]

Originally Posted by Spritze

AllSim is just a maybe as I couldn't figure out how to submit something so I e-mailed it.

Available here.

[fredmanrique]

Spritze,
Thanks for all your hard work- great database. I downloaded the new one and started running a historical sim. Up to 1954 and noticed one weird thing- Satchel Paige is listed on the batting leaderboard w/ 3556 games played. If you go to his personal stats page, you can't find anywhere that says he played in all of those games- it shows his outstanding pitching career you would expect w/ 327 wins. AND, he's not listed on the pitching leaderboard anywhere, although he should show up in several places. Any idea how I can fix this? I can't seem to find any other players this happens with.
Thank you in advance.

[fredmanrique]

Sorry- I'm an idiot. Paige is on the pitching boards. So it's only the batting leaderboard "problem."

[Spritze]

A requirement of the db was that every player have a min of 200 ab's per season for OOTP to create good batting ratings. This results in pitchers needing a bunch of games played to reach 200 ab's. So in the db pitchers might have in the batting csv 162 games and 200 ab and in the pitching table 36 games and 57 innings. In the next week or so I plan on changing this so GP in the batting table is never more than double the GP in the pitching table or I might just make them match. I haven't decided which way to go yet.

This has no effect on how the db plays, just how it looks in leaderboards.

[No Pepper]

Originally Posted by Spritze

This CSV creation db was put together modularly so it is now extremely easy to change and alter. What that means is that nearly any kind of Lahman style db can be easily created based on nearly anything anyone wishes to base one on.

Great work on the database Spritze. Any chance that you could supply us with a set of instructions on how to do this ourselves, so that we don't flood you with our own specific requests?

[plannine]

So far, in 5 test runs, Rube Foster (Negro League Pitcher) never even comes close to his estimated stats (or close to who the readme said he is molded after - eddie plank). While his career was shortened due to being a manager & owner, he still pulled off a 13-2 record in 1910, but wasnt pitching much in the later years. With him and plank playing on the same team in one test, Foster during the first 5 years of the league went 100-49, 2.19 and Plank 100-73, 2.58 on a consistently bad hitting team. Durning the next 5 years, Foster only played in 15 games in the majors, while Plank continued with another 100 wins. The real life stats for Foster are not even close to Plank's, so its amazing (to me) that he even comes close to Eddie's numbers even for a couple of years, but in all 5 tests his career is about the same for a while. (Rarely does he play in the majors after 1907). A many other Negro League Players fail the same in later years, while most fall way below what was expected of them - Candy Jim Taylor (base on Baker?) on appears in 25 games in the majors, and in the minors, he's 2nd string.

So, it looks like, many of the Negro Leaguers were weakened, so much, that they are at best late season call-ups, or career minor leaguers...

Since I've only played around (on the new version) with 1901-1920, I havent seen how Mr Paige does or many of the others.


But all that said, It is still a very good set of historical data, with most doing as expected barring freak injuries and surprise rating dives. And those events are random, so it doesnt happen every replay.

thanks for a great set...



footnote: havent noticed, but does the base at bats mess up pitchers who played the field a bit during the season (like Otto Hess, for example) and have fielding ratings at other positions during various seasons.

again, thanks for a wonderful set to use in this fantastic game (OOTP 8)

[Spritze]

Plan-Nine and a half:

Instead of "molding after" I used real Negro League stats and Mr. Foster didn't play much in many years as he was busy making money owning things.

Actually no player was molded at all. Some started playing late (ie. Bullet Joe Rogan was in the army until 28 or so), some got a real job and left early and some played forever. In my test runs players like Bill Foster, Oscar Charleston and Josh Gibson finished where one might think, near the top of the leaderboards. The older negro leaguers flatly were not always as good as their reps and those don't generally play much as you noticed. Most have a couple of decent seasons and then wander off.

By real stats I mean play against other negro league teams and white barn-stormers but NOT play against town teams and all that. In many cases while doing this research I found that many negro leaguers were not as good as their oral reps might lead one to believe.

Just today I added to the db the ability to adjust the debut season backwards in a stable way. As my samples I adjusted Luke Easter to start at age 22 and Connie Marrero to start at age 27 as those were the ages they began playing organized amateur ball. Before that it was only disorganized town ball. So now I am going to think about adjusting a few more negro leaguers backwards.

The challenge about doing this according to Garlon is that too much of this fiddling and faddling will mess up OOTP's ability to assign the correct historical yearly totals to the correct players. If Josh Gibson hits 50 HR's say then maybe Joe DiMaggio only hits 42 instead of 44 as the overall HR totals will not change. That is why there is only a smattering of added players.

Base ab's doesn't seem to have messed up anything in my tests but be sure to let me know what you think.

[Spritze]

No Pepper:

I plan on releasing the entire db at some vague future point so others can add to it and improve on it. It will require knowledge of and a copy of Ms-Access or an SQL based db program that can read Access files. It outputs Lahman style CSV files. There will be no instructions as if you don't already know how you won't be able to change things without hosing the whole magilla. I only used standard SQL queries and made sure I didn't use any weird code to accomplish things. It makes the db a little bit less efficient but since it produces Lahman style files in a few seconds the K.I.S.S. method seemed preferable to forcing the next feller to figure stuff out. And before you ask there will not be a Kanji version either.

[plannine]

Originally Posted by Spritze

Plan-Nine and a half:

Instead of "molding after" I used real Negro League stats and Mr. Foster didn't play much in many years as he was busy making money owning things.

I know by John Holway's research Rube was 28-15 on stats over 5 or 6 years, but i've seen variations that put him at 58-39 (both Negro League & Cuban League Games). That doesn't make him Cy Young, but he should be better the Mark 'the bird' Fidrych. But I can modify the players if I think they should have different career settings.

Bullet Joe needs a 8 year pitching career, with a overlaping 8 year batter career since the game does not handle 2-way players or players over 40 very well.

Overall, Home Runs need to be reduced greatly on N/L players. I think they never should be over 40 for any of the players. Keeping the batting averages near the Major League leaders (since these are the best of their league) would keep them on Major League teams.

I know I don't want them all making the top of the leader boards, or overall dominance, but I do want them to be 'in the news' in their time in the game.

[Spritze]

Martin DiHigo is the only negro player given a two-way career in this db.

[Spritze]

Plan Nine and 7/8ths Guy:

Upon further review the ruling on the field has been overturned and Negro League pitchers have been adjusted more better. IP were too low in many years because the db made that adjustment before negro league pitchers were added and it only affected less than 30 IP fellers anyway and OOTP wasn't using negro pitchers correctly in that instance. Does now.

[plannine]

Originally Posted by Spritze

Plan Nine and 7/8ths Guy:

Upon further review the ruling on the field has been overturned and Negro League pitchers have been adjusted more better. IP were too low in many years because the db made that adjustment before negro league pitchers were added and it only affected less than 30 IP fellers anyway and OOTP wasn't using negro pitchers correctly in that instance. Does now.

Thanks.....can't wait.....

[plannine]

"Plan Nine and 7/8ths Guy"

Wow, next thing I'll be a 10!!!!

[fredmanrique]

Spritze,
If I could impose- I feel like the previous Garlon versions actually dealt very well with the added Negro Leaguers. I have used 2006 DB with much success and really like the idea that this handful of players, relatively, IS near the top of the leaderboards b/c the anecdotal evidence suggests these were all if not great then very, very good players. I can see the argument that if you added 100 or 150 Negro Leaguers then you could have an argument on your hands, but it's only 50 players and it's a hell of a bunch.
Is there any way to get a DB where the added player stats are based on comparisons to other major leaguers and not based on their NeL stats? This seems to me a better way to do it- had Josh Gibson played in the majors, there are many reasons to argue he would have put up stats similar to Jimmy Foxx. I guess I just think if the point of this DB is to neutralize the environment and you can argue Gibson was equal as a power hitter to players like Foxx then by using their Negro League stats the added players are not in fact on the same playing field, so to speak?
If I should just shut the hell up and edit the players myself, then just tell me. Thanks again for all your hard work with this.

[Spritze]

An Example: The reason not to use Mr. Foxx's stats is you just get another Foxx (and you have already got one) which seems awfully boring, redundant, silly and just plain wrong.

I also believe Mr. Gibson and the rest of the negro leaguers can stand on their own merits. Besides which using Mr. Gibson's actual negro league records in this db he winds up with 531 HR's 1989 RBI's and a .297 BA in my main test league, which seems awfully Foxx like to me anyway.

I expect there may be a bit of fine tuning needed as this db progresses with certain players as the Negro and Japan MLE formulas will most likely need some tweaking. Especially pre-1920. These formulas basically work similar to the esteemed Mr. Garlons 750 run adjustments that make up the main portion of the db with some changes necessary due to uneven schedule lengths and some non-existant negro league information.

On the whole I think I'll just leave things as they are for now unless you can convince me that a whole league of Foxxes is preferable to a whole league of varigated hounds!

Please feel free to make any changes to the db you like of course. Edit away!

[fredmanrique]

OK, I'll edit away. I completely agree with you about Gibson. He's coming out just fine in my tests as well. Sorry, bad example.
A better example would be Pop Lloyd, and I'm probably just missing something, but if you look at his real life stats tab, he's a career .202/.272/.236 hitter. This is a guy who Connie Mack compared to Honus Wagner. It's probably too fine a distinction, but in that instance I would personally rather have two Honus Wagners (and I don't mean exactly, just somewhere close), b/c in my opinion that would have been closer to what Lloyd would have done than a career .508 OPS. With any sort of recalc on, Lloyd never really has a chance with those real life stats (well, a pretty small chance.) This is also true of other guys like Pete Hill but you did say that especially some of the pre-1920 guys still need to be tweaked, so you're probably already on this.
So I'll just shut the hell up and edit.
Thanks again.