Winning is Cool, but... (Full Version)

All Forums >> [Current Games From Matrix.] >> [Sports] >> PureSim Baseball



Message

KG Erwin -> Winning is Cool, but... (8/28/2009 11:28:34 PM)

...making all these trades and building a team of All Stars seems too easy. I'm gonna stop my current association and try something different in the next build. Sure, my Dodgers dominate the NL, but it isn't my so-called managerial genius that's done it -- it's the problems with the AI GMs. The individual games DO provide a good challenge, though.

Problem 1: the AI plays guys who are already hurt, so the injury list is way too long, especially for pitchers. The mandatory 10-man staffs help somewhat, but not enough.

Problem 2: using financials in the reserve clause era. Perhaps this is my mistake, but it is apparent that the AI GMs still can't handle it.

Problem 3, related to problem 2: the way the AI GMs value players. I wish I had an answer for this, but there's no way I should've been able to acquire guys like Joe DiMaggio so easily. Part of the problem is in the low POT ratings for certain veterans, which drastically lowers their value in the eyes of the AI. Perhaps taking the average career POT values more into account could help alleviate this issue.




puresimmer -> RE: Winning is Cool, but... (8/28/2009 11:32:04 PM)

Let's light this thread up, because I am going to next focus on nothing but making the AI more challenging. Especially in real player associations where human players currently have too much of an advantage.



Wrathchild -> RE: Winning is Cool, but... (8/28/2009 11:41:06 PM)

I have nothing to say about this, as in 8 game by game seasons I have yet to clinch a pennant.



Nukester -> RE: Winning is Cool, but... (8/29/2009 12:40:53 AM)

Ive taken a version of the Garlon/Spritz OOTP database and converted it for use with PureSim. This database cleans up and adds a gazillion things to the vanilla lahman database. Basically its a neutralized stats database with calculated data for missing things like CS, K's, and a huge thing for players like Dimaggio, it adds stats for missed seasons like for the war years. I could try and get ahold of him to see if he would be ok with me releasing this to the PureSim community as well. Of course Ive been the only one using it so testing has been limited. Here is a link to all the changes done. Ive removed the negro leaguers and japanese players for my PureSim leagues. If this isnt cool linking to this forum, please remove and except my apologies. I think its ok though http://www.ootpdevelopments.com/board/rosters-photos-quick-starts/153106-spritze-garlon-db.html

This is how Joe D looks from a 1939 import

[image]local://upfiles/21382/56450B2E650F457F89272CC8028B4A87.jpg[/image]



Nukester -> RE: Winning is Cool, but... (8/29/2009 12:44:55 AM)

Here is what Dimaggio's real life stats look like for the 1939 import. They are different because they are neutralized (which could pose a problem with the new 1.84 feature of career stats importing with a player for leaderboards etc.....hmmm....). Notice the war years have estimated data, so the potential goes way way up compared to the 19 potential Ive seem you guys post for Dimaggio





[image]local://upfiles/21382/FF7DCA941BD648BE8A79ECE8FEB90953.jpg[/image]



rich12545 -> RE: Winning is Cool, but... (8/29/2009 1:02:01 AM)

That link really didn't do anything.  Nukester, it would be nice if you were able to post your modified db.



KG Erwin -> RE: Winning is Cool, but... (8/29/2009 1:36:50 AM)

Nukester, while I appreciate this, the war years were lost to these guys, and the absences should not be ignored. We can't pretend that WWII never happened. That completely invalidates a historical replay that includes the years 1941-1945. Sorry, but that's how I feel.



Nukester -> RE: Winning is Cool, but... (8/29/2009 1:55:01 AM)

What do you mean it didnt do anything? It links to the changes they made to the .CSV files (which Ive converted to a .mdb file in the PureSim format). I could try and get the ok from Spritze and/or Garlon to upload it to PadresFan's site. One thing Id have to rework is that I usually start my leagues in 1903, so to get a few more real players in the league with 50 man rosters, I added in an extra stat line for anyone that stopped playing in 1902, copying their 1902 stats as 1903 stats

KG - Thats cool. I guess you can look at it that way if your league is starting post-war, and with the new 1.84 thing with importing with career stats would make his career numbers to high. If the league starts pre-war, then it makes sense, since Dimaggio isnt going to suddenly stop playing in 1942 in the game then start up again a few years later. The only thing the missing years do in that case is make his potential crappy. I dont think Dimaggio's potential was any less when he was in the war than it was before the war. He just didnt get a chance to show it.



rich12545 -> RE: Winning is Cool, but... (8/29/2009 2:25:31 AM)

That link went to a forum.  Garlon had a link in his posts but it was dead.  Couldn't find a link to the file.  And I wouldn't have a clue on how to convert it anyway.  One thing you could do is offer to email it to those who would send you a pm with an email address.  I think that would be ok.  I don't know how many would request the file or how much time it would take but it's an option.



DonBraswell -> RE: Winning is Cool, but... (8/29/2009 2:40:08 AM)

KG said,"...making all these trades and building a team of All Stars seems too easy. I'm gonna stop my current association and try something different in the next build. Sure, my Dodgers dominate the NL, but it isn't my so-called managerial genius that's done it -- it's the problems with the AI GMs. The individual games DO provide a good challenge, though.

Problem 1: the AI plays guys who are already hurt, so the injury list is way too long, especially for pitchers. The mandatory 10-man staffs help somewhat, but not enough.

Problem 2: using financials in the reserve clause era. Perhaps this is my mistake, but it is apparent that the AI GMs still can't handle it.

Problem 3, related to problem 2: the way the AI GMs value players. I wish I had an answer for this, but there's no way I should've been able to acquire guys like Joe DiMaggio so easily. Part of the problem is in the low POT ratings for certain veterans, which drastically lowers their value in the eyes of the AI. Perhaps taking the average career POT values more into account could help alleviate this issue."

_______________________________________________________________________________________________


I agree the AI needs to be stronger. I don't use finnancials for historical leagues, it just doesn't work for me. The AI makes poor draft choices. Good to fair real major leaguers are left undrafted and scrubs are picked instead. During the war years players are demoted or released. Maybe they should be, but it just doesn't seem right to me. Good veterans are demoted because the AI just drafted or signed a hot new rookie three or four years before he really played in the majors. The handling of the rosters and line ups needs work also. If you only have one catcher on the ML roster you need to be fired. [:-] If you pinch hit for that catcher and then sub a short stop at catcher you need to be shot! [X(] [:'(] But really, I see this alot from the AI. Lopsided rosters are common from what I've seen. Four 1st basemen, five short stops. one catcher, one 2nd baseman and two outfielders (sometimes from the same position). In my leagues most of the time there are better choices available as free agents after the first year. Old players, who can't compete any more and scrubs are kept in the minors. Keeping other players from being signed. The AI should release them and look for fresh talent.

I really enjoy this game, really love playing. But, I often kill a league because it can become to easy to take advantage of the computer AI. But, then again with a stronger AI, I would like to be able to win every once and awhile. [:D] [8|] I find it a challenge to take a very poor team (Browns, White Sox, A's, Senators or Pirates) and build that team into a winning team through drafts and trades. Most of the time you have very little any team will trade for. But it's great fun to try and build a strong team.



Nukester -> RE: Winning is Cool, but... (8/29/2009 2:51:25 AM)

Email wouldnt work I dont think. The file is 124 MB zipped up. Also its modified specifically for my usual league setup start date of 1903. I use the .CSV version for my OOTP historical league, so any players starting before 1903 have had their record removed from the master table because that game can import real history up to your start year, and I only wanted league history for players that are actually in the league in 1903. So instead of Cap Anson leading in hits and whatnot, I have Ed Delahanty as the historical league hits leader for a 1903 import. I believe this is what the new 1.84 import will do

Anyway the point of my post wasnt to say that I had this file, but it was to show what filling in those missing years would do to Dimaggio's potential. I only did Dimaggio because Ive seem him mentioned a bunch of times about how easy it is to trade for him because of a low potential. I think this would also solve that issue. Im not that far into my historical league to see how it plays out yet



KG Erwin -> RE: Winning is Cool, but... (8/29/2009 2:52:05 AM)

Ok, one AI/GM trend which bugs me is its decision to sometimes put its best hitter in the leadoff spot. I disagree with this. The best hitter should always be in the 3 or 4 slot in the batting order.

Pitcher injuries -- this is a huge issue. Too many pitchers end up on the DL, and the AI doesn't seem to want to sign a FA to fill the slot if a competent minor-leaguer isn't available.

Trades -- this is a tough one. There are occasions when no one is willing to make a decent counteroffer. At other times, a steal can be easily made. Perhaps an in-season "trade window" should be established, with the dates going from June 1 to July 31. This limitation might encourage the AI to be more aggressive in making/taking deals.



KG Erwin -> RE: Winning is Cool, but... (8/29/2009 4:59:39 AM)

Too many PC games seem to treat the AI as an afterthought. Call me a masochist, but I WANT my games to be challenging. PS is good, but it can be better, and I'm glad that Shaun and Steve take the AI aspect seriously.

I love baseball, and I love this simulation of it. Making the AI "smarter", without resorting to cheats, is indeed the hardest thing to do. Most of us understand this, so we will stand with you and give you our unconditional support.

Speaking personally, I'd love to get an opportunity to manage against a virtual Casey Stengel or John McGraw or Leo Durocher. I'd also love to try to out-GM Branch Rickey.



modred -> RE: Winning is Cool, but... (8/29/2009 10:54:44 AM)

To get back on track with the AI - I would love to see separate XML files for the AI GM logic and be able to assign different types to different CPU teams. Also, if managers are going to make a comeback the same thing for them. Thus, in the end I could create a series of XML files that contain my custom general managers and field managers. This would make it easier for non-windows programmers to contribute items.



modred -> RE: Winning is Cool, but... (8/29/2009 10:57:28 AM)

Also, I haven't looked at this closely in the most recent versions but previously the AI GM would make too many offers to middle to lower of the road guys early in free agency, this guys would accept, and then the team would be out of roster slots and money while there were stars still left out there. The stars would still be left after the offseason free agency to be scooped up by me - the human team - for cheap.

The game itself also seems to create too many filler players which confuse the AI.



buchanan_jm -> RE: Winning is Cool, but... (8/29/2009 1:49:56 PM)

I've just finished my 10th season; my team won 122 games in a 154 game season.  The AI needs to be a little harder.  I think the biggest issues are rosters, as has been mentioned.  Carrying just one catcher is stupid.  One AI team did this, and the catcher got injured 2 days before the World Series.  I just don't think that would happen to a real life team.  Batting the best player first is goofy, as KG mentions.  Related is the AI penchant for pinch hitting for star players.  And then substituting in odd ways.  During a 3 game series, you might get the same 3 outfielders playing, but one day they'll be lined up LF, RF, CF and the next they'll be CF, LF, RF.  Why the change?  It doesn't seem to have to do with defensive ability.

Trading doesn't bother me as much.  Great deals can sometimes be made, but, for me, it's usually only at the deadline.  Drafting, too, can sometimes be odd, but the AI has a method.  It always seems to draft the most ready now player.  That leaves lots of good pitchers--albeit 3 years away--in round 5.  What bothers me more is there's usually a pretty good player or two never signed during the free agency period.  Believe it or not, one year the AI teams left Joe D unsigned after he had played just 2 years in the league.  Joe D is an odd case, but Wally Berger was unsigned one year too when still in his prime.  It's just hard to imagine this happening in real life.



markvacc1 -> RE: Winning is Cool, but... (8/29/2009 4:09:17 PM)

One thing I have been wondering about is the left vs right hitting and pitching. Doesn't seem like it matters much, at least not to the ai. Anybody want to shed some light on this?

Also, I have been simming and relief pitching is awfully neglected. I've gone thru a whole season with some relief pitchers on the major league roster all season and only pitch 5 innings. It's been in the 70s and realize RPs have been used more in the latter years, but still should be more than that in the 70s.

Wish we could have a faster FA period, especially if not going for any or already signed what you were after. When click skip the rest, it should fly by.

Wouldn't mind seeing somebody start a real player league in the very beginning with an owner for each team and as a new year approaches with new teams coming in, a call goes out for an owner for expansion. Or all the teams listed and somebody says they will take Texas, which was the Senators first and he waits til that season approaches to get involved. Of course the Commish would have to contact him to see if he is still in since the waiting period may take quite awhile.



dneely -> RE: Winning is Cool, but... (8/31/2009 6:40:56 PM)

I thought I would way-in on this topic and add my 2 cents.

First off I have not been playing as much recently but hope as life and teaching in college slow down a bit I will be able to find more time and have more interest. The following all are my observations since the first incarnations of Puresim.

The game does allow you to steal big name players IF you do it before they are anywhere near their prime! In my 1947 season I picked up Early Wynn and Richie Asburn for nothing....I have NEVER had a season where I won close to 122 games in a 154 game season. But I have always tried to shy away from abusing the AI in trades and pickups (most of the time....). I have cried out for era specific managers since day one! that would be the Crowning Jewel for Puresim if Shaun and others were able to create this. APBA was the first computer sim (to my knowledge) to allow the community to create managers with specific tendencies, era aspects, modeled and named for famous managers of the past and tons of modern day incarnations. I can remember playing in online leagues where your opponent would send you or ask you to use a specific manager for the series at your park. Was it perfect, no it was not, but it added a really nice extra element to playing out games and having some idea what your opponent manager was likely to do. Managers in the 1900-1920 era would rarely pull a starter, make extensive use of the bunt, steal, hit & run etc etc they played for one run at a time and had no idea what a closer was. The modern ones were like Tony LaRussa going to the bullpen constantly always looking for the righty on righty or lefty on lefty advantage. I don't use financing's because I HATE math and don't want to be bothered with that aspect plus it does not work in the two era's I enjoy the most - the dead ball era and the 1950's.

I am pleased that Shaun has stated that he wants to concentrate on the AI in the next version of Puresim. But having played virtually every name or type of computer baseball sim it is a very daunting task! I don't know of ANY game where the players are completely happy with the AI.

My advice:

era specific managers are the best way to go!
tighten up the AI in terms of how it rates players and, if possible, give it an advantage of looking years ahead in terms of player ratings based on career stats.
I don't know what to say about financials???
Add lefty vs rightly effects on some order.
Make sure the game has options like pulling the infield and outfield in, suicide squeeze, bringing just the 3b in on a bunt as well as the corners. And any others I may have missed!

Lastly keep this great game alive! I always said, and will continue to say, that this game is the closest one to any on the market as far as the ability to become the "Perfect" baseball sim. And I STRONGLY advise bring back and making it the best available for online leagues! Shaun if you do that, you can retire and turn Puresim into a full-time occupation if you desired...



Wrathchild -> RE: Winning is Cool, but... (8/31/2009 6:57:30 PM)

One option I like that is in Out of the Park is similar to what dneely is talking about. Based on the year that your universe is set in, you have the option of using historically accurate tendencies for that year, such as bunting, pulling starters, etc. You also have the option after you see how it sets this sliders of customizing them if you want. You could, for example, change the 1904 setting of Rarely Pull Starters to Frequently Pull Starters if you wanted to add a little variety. What would be cool is then having, on top of this, the generic or historical manager types, which would adjust these settings for that manager. For example, if the season setting for Hit & Run is Common, you might have a manager type or historical manage who does Hit & Run +10%, meaning that whatever season you are using him in, he's more likely to Hit & Run than the standard.



rich12545 -> RE: Winning is Cool, but... (8/31/2009 8:44:18 PM)

One thing about PureSim now is that it's the only one that can get as close to perfect as can be gotten.  Seem like the other two biggies, ootp and bm (which I have purchased), add just enough each year to get their annual sales.  Clay at bm admits it, that it's like buying a subscription every year.  I have really high hopes for ps the way Shaun is improving it without thinking about next year's sales.



GoldenGreek -> RE: Winning is Cool, but... (8/31/2009 9:16:09 PM)

Would financials not still be beneficial in a reserve clause league? The reserve clause let players become free agents if they held out for a year. Frankly, I would like to see players have different kinds of agents (adding no trade clauses, MLB contract vs. minor league contract, options, etc). Shoot, puresim could just take all of the cool stuff from Bill Beane front office baseball!



KG Erwin -> RE: Winning is Cool, but... (8/31/2009 10:12:41 PM)


quote:

ORIGINAL: GoldenGreek

Would financials not still be beneficial in a reserve clause league? The reserve clause let players become free agents if they held out for a year. Frankly, I would like to see players have different kinds of agents (adding no trade clauses, MLB contract vs. minor league contract, options, etc). Shoot, puresim could just take all of the cool stuff from Bill Beane front office baseball!



One thing I'd like to see in the reserve clause era is something I'll call "limited financials". This means that the clubs each have a treasury that can be used to buy players from other teams, which was a common practice. In this fashion, teams like the Athletics can hold periodic "fire sales", which Connie Mack did indeed hold several times in his long career. Of course, this would require that each player have a minimum "cash value", but the money would NOT go into the player's pocket, as he was basically a commodity.

Introducing this concept would definitely illustrate how the owners once held sway over their "commodities". It would also illustrate how Branch Rickey's concept of the farm system was revolutionary, in that he could stockpile players, sell or trade them off, and then sign others. This was how he turned the Cardinals, and later the Dodgers, into perennial contenders.



GoldenGreek -> RE: Winning is Cool, but... (9/1/2009 12:06:27 AM)

I wonder if it would be beneficial to have the option for larger minor league systems with full team rosters and a set schedule for each team, etc. Then you could set minor league ballparks and truly see how a player might look in a minor league system. PS doesn't have a rookie ball or short season, nor does it have a fall league. I am guessing this is not a 1.85 update request, but just something to ponder as a future thought to build into the system.



rich12545 -> RE: Winning is Cool, but... (9/1/2009 1:31:24 AM)

If you'd like to see how that works, download the ootp demo.  I think it works for some who want all that micromanagement and it doesn't work for others.  When I play ootp, I use only AAA with a high school feeder league.  Any more is just too much for my brain to take in.



KG Erwin -> RE: Winning is Cool, but... (9/1/2009 2:33:08 AM)

I agree with rich12545. PureSim doesn't overwhelm the player with micromanagement and minutiae, and I'd prefer to keep it that way. Wearing the hats of both GM and game-by-game dugout manager provides enough responsibility to keep me from getting bored.



Orcin -> RE: Winning is Cool, but... (9/8/2009 9:12:25 PM)


quote:

ORIGINAL: puresimmer

Let's light this thread up, because I am going to next focus on nothing but making the AI more challenging. Especially in real player associations where human players currently have too much of an advantage.



Shaun,

Before I get to my question, I need to explain my logic. To determine how the AI GM ranks players, I did the following:
1. I offered my best player to the other team, and he offered anyone on his roster in return. Counting down his list in order shows me how he ranks his own players.
2. I offered to trade for his #40 player, and he would accept 35 of my players in return (he won't accept fictional players). Counting down my list in order shows me how he ranks my players.

Given those assumptions and using the new "tougher" AI GM in 1.84 BETA 2 and beyond, should I be able to trade my #7 and #10 rated players in a 2-for-1 to get another team's #2 player? I asking to see if that is how you intended the logic to work, or is that still a lopsided trade in favor of the human in your opinion? I am not being critical of the AI here. I am just trying to understand if you "fixed" it this way on purpose, or if it is still one of the issues that you haven't or couldn't address?

In case more information is helpful...

his #2 is 22 with potential of 16 (developing)
my #7 is 34 with potential of 1 (past peak)
my #10 is 29 with potential of 1 (past peak) but on his list of desired positions

Anyone else that wants to weigh in, please feel free. My opinion is that the AI is still unable to handle a multi-player trade without getting robbed, and I am avoiding this deal for reasons of competitive balance.








kg_1007 -> RE: Winning is Cool, but... (9/8/2009 9:35:35 PM)

I think that baseball is a game of numbers, where over time, things "average out". While I am only now playing my first season, I know that my brother, who got me into this, has played the seasons 2001-2007, and just as I do, plays as the Cardinals. Early in the time frame , just as with the real life team, he was winning between 95-110 games, then, just as in real life, the 2006 version slipped, and the 2007 version tanked with injuries...so it was either a purely lucky coincidence to match real life, or very good simulation.The AI does have some issues, especially as I posted in a different topic about the Righty/Lefty matchups, but overall, judging from his experience in 7 + seasons, and mine in about 1/3 of a season, things should even out it seems to me.
By the way..a great community here, you all are very knowledgeable, thank you.



Page: [1]

Valid CSS!




Forum Software © ASPPlayground.NET Advanced Edition 2.4.5 ANSI
3.5625