Question on your Tests

[coljesep]

Are you using the same ballpark/city for all teams? If not, wouldn't park factors influence performance?

[CAT]

Most definitely. I have tested both ways. But I would probably try a test with a symmetrical park with altitude turned off.
When I ran my no ratings test, I used a domed stadium to prevent element interference.


I have been a tad busy lately and thus no posts, but I plan to post very soon.

[jcbarr]

My tests have all been run with varying stadiums and cities. I did this because that is the way that the league is going to play so I didn't think it would be logical to test any other way. Also you have to make sure that you have an ini file that can cope with all the cities and the stadiums. Just my two cents though.

[Knuckleball Lover]

Also you have to make sure that you have an ini file that can cope with all the cities and the stadiums.

This.

I mean, of course ballparks and altitude impacts things. Just like it impacts real baseball.

Unless your league is going to be all one city and stadium - test in the environment you're going to be playing in. What good is getting a test in a 0 altitude, neutral park filled league only to see it all go crazy when you bring in all the stadiums and cities?

[coljesep]

I don't think it was ever answered when I asked but for those continously running tests on pb.ini, what's your eventual goal? I know the idea is to get the coding to be as realistic as you can.... but then what? Or is that just the project?

[Knuckleball Lover]

I don't think it was ever answered when I asked but for those continously running tests on pb.ini, what's your eventual goal? I know the idea is to get the coding to be as realistic as you can.... but then what? Or is that just the project?

For me, that was the goal - to get the stats as realistic as possible with as many ratings mixes as possible, making the pb.ini as close to universal as possible. This is especially the focus around player ratings that you commonly see, like 70/70/60 or 70/90/50, 80/40/50, etc.

Also wanted to make it so that the usual "traps" don't have to apply, for example, make it so players didn't need a very high PH to have a high batting average or pitchers had to have very high AS or pitch ratings to succeed, etc. Likewise pb.ini "traps" like turning off altitude to keep Coors Field in check, having GF mean very little, having too many/too few of a batted ball type, making defense meaningful, etc.

[coljesep]

So you mean regardless of player ratings, so if a 60 PH is the top in your league and a 90 ph is tops in my league they'd have the same production etc

I kind of gave up on PB.ini a few yrs ago. I like where my #s are, it's not too crazy but people understand it's a 10 yr old sim. It was something I tried real hard on a few yrs back and got to the point where I was "happier" with my results and just left em because I don't want the league to constantly have to adjust to new settings etc

[CAT]

I see where you a going Mike. Good points

Testing in a neutral environment does have its advantages when testing for homeruns especially. I did do this with my 8 team league since the sample size was small.

Since we are trying to get as close to MLB average results (or AL or NL average if using only one league) either one park style (preferably a park with a factor of 100) or use each team's home park since that is what we are trying to emulate. Using real parks is almost moot because half of the stadiums in BBpro don't exist anymore or are not used for baseball anymore.

So we are actually testing with parks that don't exists.

But my main focus has been contact and walks and not just homeruns or other special hits. This has worked for me thus far. When these are balanced, everything else falls into place.

I ran my last round of tests "offline" because I found an error in my testing format and had to fix it and in the process of "fixing" it I actually I destroyed my pb.ini file again and had to jump back few versions to get my stuff back.)

Anyway, I focused more on balancing the balls in play with the strikeouts and walks and if the home runs were not ridiculous, I new I was ontrack.

As you posted though, if the file works and the league is consistent, leave it alone.

Here are the parks that are no more.
Yankee
Shea
Comiskey
Metrodome (2010)
Astrodome
Veterans
Three Rivers
Kingdome
Candlestick
Qualcomm
Riverfront
Tigers Stadium
Atlanta Fulton
Busch

[jcbarr]

I don't think it was ever answered when I asked but for those continously running tests on pb.ini, what's your eventual goal? I know the idea is to get the coding to be as realistic as you can.... but then what? Or is that just the project?

My goal was very simple. I wanted to get a pgend.dat file and a pb.ini file that would generate players and produce stats that were comparable to major league averages. That way I could effectively run a sim without having to ever edit a player. I could simply focus on running the games and tinkering with teams. I did this in preparation for the league that I am opening very soon.

It is fun too just to run test after test and see what you get, but I did it ultimately to be able to start a league that will have consistency in both player ratings and player stats.

In my instance my ini file might not work with your rating sets. It is VERY possible that it will HAVE to be used in conjunction with my pgend.dat file, but I'm fine with that because it means that I don't have to tinker with player ratings every year that the ammy pool is released.

[Knuckleball Lover]

So you mean regardless of player ratings, so if a 60 PH is the top in your league and a 90 ph is tops in my league they'd have the same production etc

In a way. I wouldn't be so bold as to say a 60 PH would hit 70 HR like a 90 PH. What I mean more is like if the top pitcher has 60 pitches, then he'll do pretty well relative to the other pitchers. 60 pitches won't just suck because they are 60 pitches.

I guess another way to say it is that the balance of power in the league is dictated more by the ratings then by the pb.ini file. The file doesn't lean towards making hitters or pitchers have the advantage (which is why I like to use MLB/AL/NL averages - I assume that to be the baseline standard to compare against), but instead, that depends on the players in the league and how their ratings interact. I just want the pb.ini to let the player skills show and "get out of the way". If a player fails, it's because of his own ability, not because the pb.ini "requires" a good hitter to have 70 PH or the like.

I also want it so that players could recreate other eras with the same file by generating different kinds of players. For example, a dead ball file might have higher GFs on the hitters and/or lower PHs. This might be the next project heh.

[Knuckleball Lover]

Anyway, I focused more on balancing the balls in play with the strikeouts and walks and if the home runs were not ridiculous, I new I was ontrack.

The three "true" outcomes if you believe in DIPS. Ks, HR, BBs. Especially on the pitching side.

[CAT]

To answer the question of my goal. It is like knuck said, to produce a serviceable pb.ini file that can stand an dip in overall rating from pitchers or batters without producing unrealistic results. (and more recently to create a pgend.dat file that can produce players without the need to tweak them after creation, I am on track with this as well.)

Now it is true that when talent becomes diluted, the overall stat output should reflect that. (e.g. expansion years in MLB) But there should be some physical limitations to what ratings can do. To put some clearly defined physical limitations on what players can produce without giving them too many limitations. In fact I am close to ending this round of tests, also historically speaking, it is time for me to disappear for 6 months anyway.

This is really my first time consistently testing with clearly defined benchmarks. Once I started with this approach, rest was gravy. In the past my testing was terrible and as a result the stats were terrible. It was more of an eye test instead of looking at what the ratings were actually producing. It was like this: sim a few games or a season and o.k. no one hit 100 homers, the K leader was in the 170s and so forth, then post the file and get flamed by the users. "Hey your file sucks!!". ( I still have a dump of the orignal pb.ini site which has some interesting commentary) Just never got around to posting the results, most of which is irrelevant now.

Now I am number crunching and I am able to see what areas need tweaking if necessary. Getting league average results carried by the median players and not the stars. Knuck has been very kind in trying this approach as well, although his files have always been great. In fact some of the adjustments I have made based on this in depth approach, has led to modifications that knuck had already made. Go figure. Another difference is that we are testing and posting results as we go. Which can be annoying when things mess up, which happened to me recently.

This round of tests has also allowed us to discover some more details about the pb.ini file that may not have been clearly understood.

Examples. (without going too in-depth)

Timing and pitch ID.

This was a big rant for me at one point in time. But it has been addressed. Weaker batters get a double whammy. They get a wider timing curve when they correctly identify the pitch!!

The fix was to ensure that while weaker batters should fail on Pitch ID more often the higher CH hitters, they were not given them an impossible timing curve.
Now a player with a 30 or even a 40 CH rating can succeed if they have enough power, (which would be the only reason why they would or should be playing)

Alsom, In the past we were able to uncover explanation errors in the original pb.ini.txt One is the id rating formula, It is missing a parenthesis. The G/F explanation is backwards, each point over 50 lowers the swing angle instead of raising it as explained in the file.

More recently, I discovered that the batspeed adjust lines affect the base bat speed, and is not added to a batters PH rating (as explained in the pbini.txt file). Not sure why I missed this before. But it appears to be the case, and this difference is huge since each mph increase or decrease has a big affect on the ball. About 5 to 10 feet according to some physics sites.

This current round has also allowed us to test in the post steroid era. This has allowed us not to try to get a batter to hit 70 homers for the first time, since that probably should not have happened anyway.

Good questions Mike, Thanks

[Peter]

I see where you a going Mike. Good points


Here are the parks that are no more.
Yankee
Shea
Comiskey
Metrodome (2010)
Astrodome
Veterans
Three Rivers
Kingdome
Candlestick
Qualcomm
Riverfront
Tigers Stadium
Atlanta Fulton
Busch

OK...... now all the good questions are out there, how about a simpletons question: (ha ha)

As the ballparks are dissappearing, is there anyone who knows how to create NEW ballparks?

[CAT]

I contacted Dynamix for the source code, especially for their parks, some time ago and they respectfully said. NO!!!

Without that info there is no way to create and compile the parks for the game to use.