Strategy Settings (Garlon & Others) - Page 2

Garlon · 09-17-2021, 02:12 PM

My suggestions that I posted are specifically for career mode with coaching off and injuries off.

When it comes to league stats the modifiers handle that. You cannot really affect that. My suggestions are what I use for getting nice distributions and results for great players without having to nerf too many players along the way because of sample sizes with the import settings. The rotation mode of strict will provide better winning percentages for the best pitchers at the expense of a few starts per year. You can of course change anything to your taste. These are only suggestions and may give others a starting point for their settings.

Garlon · 09-17-2021, 02:16 PM

My suggestions are to look at a set of key players like Wagner, Cobb, Ruth, DiMaggio, Williams, Gwynn, Bonds, and Trout. If these players are significantly underperforming that means the entire league being squeezed.

Brad K · 09-18-2021, 07:19 AM

Quote:

Originally Posted by jdw

Follow up question:

How badly does accuracy get messed up if L/R is turned on?

Why is there an assumption accuracy is good with splits turned off?

A single overall rating for Crawford and Backman would be lower than their split rating against RHP. Maybe its low enough someone else gets the job against RHP and they don't get to play except as a PH.

Brad K · 09-18-2021, 07:46 AM

Garlon, every time you post I learn something. The intricacies of the game are amazing.

You've put a lot of work into developing the settings you suggest and they are very useful to those who play the game the way you do. For those of us to play a different way, they're interesting and informative, but not necessarily useful.

Information that would be universally useful is the answers to the questions in post #3. I don't expect you to know everything but if you don't know these answers if you can beckon one of the devs to return and answer these questions a lot of people would appreciate it.

Garlon · 09-18-2021, 09:20 AM

When you use real stats the game makes an adjustment to the HR rates of players based on their home park in the era_ballparks file. This is some internal OOTP adjustment so that you do not get so many extreme HR seasons. If you swap out the era_ballparks file to a neutral version of the file then players will import differently when using real stats because the game actually references that file. The platers get adjusted according to their historical home ballpark regardless of what team they end up playing for in your own game.

When using neutralized stats there is no additional internal OOTP adjustment as all of the adjustments were done when those neutralized stats were put together so when you use neutralized stats the players will import the same way even if you change the era_ballparks file.

Keep in mind the import and strategy settings can affect the results. My goal is to find a nice balance by having the best hitters play up to what you would expect and having the best pitchers perform accordingly too.

If you want to do strict replays with real transactions then you may want to just use the default game settings.

Garlon · 09-18-2021, 10:31 AM

These are the results from a recent 1871-2020 game using my suggestions. Trades were disabled.

OOTP vs Real Value

500 HR: 31 vs 28

50+ HR seasons: 44 vs 46

300+ Wins: 20 vs 24

250+ Wins: 65 vs 48

jdw · 09-18-2021, 02:46 PM

Quote:

Originally Posted by Brad K

Why is there an assumption accuracy is good with splits turned off?

A single overall rating for Crawford and Backman would be lower than their split rating against RHP. Maybe its low enough someone else gets the job against RHP and they don't get to play except as a PH.

This is where it gets to be a catch-22 with players who are platooned quite a bit in real life.

Because he was limited against LHP for much of his career, Backman's overall numbers are far closer to his vs RHP numbers than they would be if he played full-time against all pitching. What it does is make Backman an okay hitter to use against both RHP and LHP.

Same goes for Crawford.

Backman Career
.294/.364/.362 vs RHP
.275/.349/.339 vs ALL
.165/.258/.201 vs LHP

Carwford 1972-76
.295/.381/.439 vs RHP
.285/.368/.426 vs ALL
.231/.299/.355 vs LHP

.275/.349/.339 is perfectly decent for a 2B from 1980-93. There will be some 5-year recalc where Backman is better than that, likely 1984, 1986 and 1988.

In turn, .285/.368/.426 in 1972-76 with 4 of those seasons in Dodgers Stadium before they moved the plate in 1977 is Really Valuable as a full-time player due to that OBP while chipping in some mid-level EBH pop for that period. Using neutralized stats, there's some nice kickers in there.

Crawford and Backman aren't at that level overall if they were playing full-time. But they are pretty valuable against RHP when their plate appearances are limited against LHP.

jdw · 09-18-2021, 03:14 PM

Quote:

Originally Posted by joefromchicago

Garlon and I have had several conversations about this topic and I won't belabor it here. If you want to get more in depth, take a look at this thread.

Joe - I have read several of the threads where you and Garlon (and others) go around the bases on pitchers stamina and rotation. I take that stuff into consideration when making tweaks here and there to Garlon's recommendations in the era I like to play (70s though mid-90s). It's always great stuff to read.

Garlon · 09-18-2021, 03:35 PM

Regarding the batting vs RHP and vs LHP, BigRod and I looked at the data in the database and realized it was incomplete so we decided we should rebuild those stats since we now have complete data for every batter every season since 1901. This is not yet in the game though as this needs to be formatted properly. There are some hurdles with this that needs to be worked out with player ids if I remember so I was not able to get to this.

Regarding pitcher stamina, I think that Normal may be the best option as this setting applies to relievers too. Yes, finding the right setting for pitcher stamina has been something I have spent quite a bit of testing on and I think the results seem best on Normal. When you go above normal this can make relievers stronger, which in turn can diminish results from starting pitchers.

I used to prefer start highest rested to get those extra starts from the #1 starters, but I think you get better winning percentages from your top pitchers on strict rotation because it makes the #5 starter make 32 starts instead of 24 starts and that has an effect across the league.

There are many settings and many combinations that are possible to test but these seem to give me good results.

Garlon · 09-18-2021, 03:54 PM

From that same league I recently mentioned:

OOTP vs Real

Anson: .319/.377/.418 vs .333/.395/.445

Delahanty .357/.428/.520 vs .346/.410/.505

Wagner .322/.381/.465 vs .327/.390/.466

Cobb .358/.430/.499 vs .366/.432/.512

Ruth .337/.470/.617 vs .342/.472/.690

DiMaggio .338/.413/.595 vs .325/.396/.579

Williams .338/.477/.623 vs .344/.480/.634

Mays .308/.391./.539 vs .302/.384/.557

Aaron .305/.369/.532 vs .305/.374/.555

Clemente .319/.352/.477 vs .317/.359/.475

Morgan .272/.389/.409 vs .271/.392/.427

Gwynn .352/.393/.484 vs .338/.388/.459

Boggs .325/.406/.434 vs .328/.415/.443

Bonds .298/.428/.582 vs .299/.444/.607

Trout .308/.433/.584 vs .304/.418/.582

Ask for players and I will post them.

jdw · 09-18-2021, 04:13 PM

Great stuff, Garlon. I look forward to the eventual version of OOTP that has the splits in it. That will be cool.

Brad K · 09-18-2021, 04:24 PM

Quote:

Originally Posted by jdw

This is where it gets to be a catch-22 with players who are platooned quite a bit in real life.

No catch 22 at all if you use splits. Assuming the splits simulate real life.

Give the players the right ratings and the AI manages will play them as they were played real life.

Brad K · 09-18-2021, 04:25 PM

Quote:

Originally Posted by jdw

Great stuff, Garlon. I look forward to the eventual version of OOTP that has the splits in it. That will be cool.

22 has splits although I think there are some inaccuracies. Much better than the randomized splits in older versions.

Brad K · 09-18-2021, 04:40 PM

Quote:

Originally Posted by Garlon

When you use real stats the game makes an adjustment to the HR rates of players based on their home park in the era_ballparks file. This is some internal OOTP adjustment so that you do not get so many extreme HR seasons. If you swap out the era_ballparks file to a neutral version of the file then players will import differently when using real stats because the game actually references that file. The platers get adjusted according to their historical home ballpark regardless of what team they end up playing for in your own game.

When using neutralized stats there is no additional internal OOTP adjustment as all of the adjustments were done when those neutralized stats were put together so when you use neutralized stats the players will import the same way even if you change the era_ballparks file.

Keep in mind the import and strategy settings can affect the results. My goal is to find a nice balance by having the best hitters play up to what you would expect and having the best pitchers perform accordingly too.

If you want to do strict replays with real transactions then you may want to just use the default game settings.

OK, so real stats are neutralized against the park factors in the game's era_ballparks file. Neutralized stats are neutralized against... what... well, real life park factors, and those should be the same as the ones in era_ballparks. So what we have are things named differently that do the same thing.

Garlon · 09-18-2021, 04:58 PM

If you use real stats there is an OOTP adjustment that Markus built into the game so that when using real stats you would not get excessive single season HR results.

The neutralized stats are a complete rebuild of the stats that are indeed neutralized against home park factors for BA/2B/3B/HR by home ballpark and whether the player bats L, bats R, or bats Both. I do not know the way Markus handles the adjustments for real stats, but I can give you some insight on the neutralized stats file.

Consider a player who had 40 HR and played in a home park with a 1.250 HR factor for RHB in a league with 8 teams. We figure half of the games are at home and half on the road. The entire league ballpark HR factors must average to 1.000 and there are 8 teams so that is 8 * 1.000 = 8.000 The home ballpark we said is 1.250, so the other 7 ballparks combines are 8 - 1.250 = 6.750. We then divide this by 7 to get 6.750/7 = 0.964 as the average of the road ballparks for that player. So the player had a HR environment of (1.250 + 0.964)/2 = 1.107. This means that player had a 10.7% advantage for HR that season. So we take 40/1.107 = 36.13 = 36. This player has 36 HR for their neutralized stats that season.

This same process is done for BABIP/2B/3B too as there are park factors for LHB and RHB for each category. These factors were built specifically for OOTP from the historical data for those stadiums in order to do this process.

The game will not reference the era_ballparks file when importing players from the neutralized stats file, so if you change the era_ballparks file it will not affect the player imports or their ratings.

I suggest taking a look at several players real and neutralized stats and decide for yourself. There is actually very little difference for most players. The biggest difference are Ruth and Ott as far as I know since they played in the friendliest HR environments so they end up getting the biggest adjustments.

This is done for pitcher HR allowed too. So when you look at pitcher neutralized stats they will be the same as their real stats except for slight differences in HR allowed though.

There game currently does have a version of the Bats L and Bats R historical data but it does not go all the way to 1901.

Garlon · 09-18-2021, 05:39 PM

A reason I became interested in neutralized stats was the Willie Davis entry on pgs 740-743 of the Bill James Historical Abstract. Since Willie Davis spent most of his career with the Dodgers there is the question of whether he would have been a .300 in a neutral environment. Well, Bill James translates his career stats to a 750-run environment using the quadratic formula and has him as a .302 hitter.

The quadratic method was used for previous generations of the neutralized stats for OOTP, but the new neutralized stats do not use that method, rather it uses the discrete BABIP/2B/3B/HR park factors for Bats R or Bats L or overall for Bats Both. If you use baseball reference there is an option to neutralize the player stats to different run scoring environments using the quadratic method. The default there is 723 runs per 162, whereas Bill James used 750 runs per 162 in his book.

Willie Davis Real stats .279/.311/.412 OPS .723

BBR 723 runs scored per season .297/.329/.436 OPS .765

OOTP .283/.314/.431 OPS .745

Where Davis gains by adjusting his stats through discrete factors is his slugging.

2B/3B/HR

Real 395/138/182
OOTP 422/161/205

This is because Dodger stadium was not really affecting batting average as much as slugging.

The quadratic method assumes the run environment affects everything proportionately.

So with the neutralized stats we do not have Davis as a .300 hitter, but he does gain nearly 20 points of slugging.

In my recent game he was also +92 ZR in LF and +99 ZR in CF, so that is 191 runs saved on defense, which is about 19 wins.

luckymann · 09-19-2021, 11:48 AM

Quote:

Originally Posted by Garlon

A reason I became interested in neutralized stats was the Willie Davis entry on pgs 740-743 of the Bill James Historical Abstract. Since Willie Davis spent most of his career with the Dodgers there is the question of whether he would have been a .300 in a neutral environment. Well, Bill James translates his career stats to a 750-run environment using the quadratic formula and has him as a .302 hitter.

The quadratic method was used for previous generations of the neutralized stats for OOTP, but the new neutralized stats do not use that method, rather it uses the discrete BABIP/2B/3B/HR park factors for Bats R or Bats L or overall for Bats Both. If you use baseball reference there is an option to neutralize the player stats to different run scoring environments using the quadratic method. The default there is 723 runs per 162, whereas Bill James used 750 runs per 162 in his book.

Willie Davis Real stats .279/.311/.412 OPS .723

BBR 723 runs scored per season .297/.329/.436 OPS .765

OOTP .283/.314/.431 OPS .745

Where Davis gains by adjusting his stats through discrete factors is his slugging.

2B/3B/HR

Real 395/138/182
OOTP 422/161/205

This is because Dodger stadium was not really affecting batting average as much as slugging.

The quadratic method assumes the run environment affects everything proportionately.

So with the neutralized stats we do not have Davis as a .300 hitter, but he does gain nearly 20 points of slugging.

In my recent game he was also +92 ZR in LF and +99 ZR in CF, so that is 191 runs saved on defense, which is about 19 wins.

I am using your Quadratic Factors in my historical Pirates save Garlon. They are fantastic, and well worth all of the effort to calculate and enter them in at the start of each season. It is a bit annoying how it messes up the previous WAR scores, but that's not of your doing.

Thanks for everything you do, sir, and your fine way of describing how best to apply them. It is much appreciated by many, myself most definitely included.

G

Garlon · 09-19-2021, 12:31 PM

I may have a copy of the new format of the era_ballparks file with the quadratic park factors. I rebuilt the quadratic factors too for this game even though the game only loads the discrete values. The original quadratic factors were 5-yr values and I rebuilt the entire file using 3yr values.

Brad K · 09-19-2021, 01:51 PM

Quote:

Originally Posted by Garlon

I suggest taking a look at several players real and neutralized stats and decide for yourself. There is actually very little difference for most players. The biggest difference are Ruth and Ott as far as I know since they played in the friendliest HR environments so they end up getting the biggest adjustments.

I'm frustrated.

1. We're being told to import players and evaluate the effects for ourselves when the people who designed the adjustments are right here and could tell us what the effects are.

2. The neutralized file output doesn't match the real file (neutralized against era_ballparks) output. Why? Well, because apparently the park factors in era_ballparks don't match the park factors used for the neutralized file. So someone using the neutralized file is taking ratings adjusted for one environment and using them in another. "Pick the one you like better" isn't the way to reasonable outcomes.

3. Neither real nor neutralized stats match the common understanding of their names. Real stats aren't real because they're neutralized then adjusted to reduce peak home runs. Neutralized aren't neutralized for the game because they're neutralized against something other than the park factors the game uses and because with optimize for career play they reduce peak home runs.

Brad K · 09-19-2021, 02:35 PM

I fully understand the need to prevent Maris from hitting 75 HRs in 1961. And I understand the reluctance to rate him for 61 and then use the game engine to impose a cap.

A decision to reduce peak ratings makes sense. However people playing with optimized for single season set have a reasonable expectation Maris will hit in the 50s. And that isn't likely to happen. People playing with career settings have a reasonable expectation single season and career home run records will be approached.

09-17-2021, 02:12 PM	#21
Garlon Hall Of Famer Join Date: Jun 2004 Posts: 4,268	My suggestions that I posted are specifically for career mode with coaching off and injuries off. When it comes to league stats the modifiers handle that. You cannot really affect that. My suggestions are what I use for getting nice distributions and results for great players without having to nerf too many players along the way because of sample sizes with the import settings. The rotation mode of strict will provide better winning percentages for the best pitchers at the expense of a few starts per year. You can of course change anything to your taste. These are only suggestions and may give others a starting point for their settings. Last edited by Garlon; 09-17-2021 at 02:13 PM.

09-18-2021, 07:46 AM	#24
Brad K Banned Join Date: May 2016 Location: St Petersburg Florida USA Posts: 6,693 Infractions: 0/2 (4)	Garlon, every time you post I learn something. The intricacies of the game are amazing. You've put a lot of work into developing the settings you suggest and they are very useful to those who play the game the way you do. For those of us to play a different way, they're interesting and informative, but not necessarily useful. Information that would be universally useful is the answers to the questions in post #3. I don't expect you to know everything but if you don't know these answers if you can beckon one of the devs to return and answer these questions a lot of people would appreciate it. Last edited by Brad K; 09-18-2021 at 07:47 AM.

09-18-2021, 03:54 PM	#30
Garlon Hall Of Famer Join Date: Jun 2004 Posts: 4,268	From that same league I recently mentioned: OOTP vs Real Anson: .319/.377/.418 vs .333/.395/.445 Delahanty .357/.428/.520 vs .346/.410/.505 Wagner .322/.381/.465 vs .327/.390/.466 Cobb .358/.430/.499 vs .366/.432/.512 Ruth .337/.470/.617 vs .342/.472/.690 DiMaggio .338/.413/.595 vs .325/.396/.579 Williams .338/.477/.623 vs .344/.480/.634 Mays .308/.391./.539 vs .302/.384/.557 Aaron .305/.369/.532 vs .305/.374/.555 Clemente .319/.352/.477 vs .317/.359/.475 Morgan .272/.389/.409 vs .271/.392/.427 Gwynn .352/.393/.484 vs .338/.388/.459 Boggs .325/.406/.434 vs .328/.415/.443 Bonds .298/.428/.582 vs .299/.444/.607 Trout .308/.433/.584 vs .304/.418/.582 Ask for players and I will post them. Last edited by Garlon; 09-18-2021 at 03:56 PM.

09-18-2021, 04:58 PM	#35
Garlon Hall Of Famer Join Date: Jun 2004 Posts: 4,268	If you use real stats there is an OOTP adjustment that Markus built into the game so that when using real stats you would not get excessive single season HR results. The neutralized stats are a complete rebuild of the stats that are indeed neutralized against home park factors for BA/2B/3B/HR by home ballpark and whether the player bats L, bats R, or bats Both. I do not know the way Markus handles the adjustments for real stats, but I can give you some insight on the neutralized stats file. Consider a player who had 40 HR and played in a home park with a 1.250 HR factor for RHB in a league with 8 teams. We figure half of the games are at home and half on the road. The entire league ballpark HR factors must average to 1.000 and there are 8 teams so that is 8 * 1.000 = 8.000 The home ballpark we said is 1.250, so the other 7 ballparks combines are 8 - 1.250 = 6.750. We then divide this by 7 to get 6.750/7 = 0.964 as the average of the road ballparks for that player. So the player had a HR environment of (1.250 + 0.964)/2 = 1.107. This means that player had a 10.7% advantage for HR that season. So we take 40/1.107 = 36.13 = 36. This player has 36 HR for their neutralized stats that season. This same process is done for BABIP/2B/3B too as there are park factors for LHB and RHB for each category. These factors were built specifically for OOTP from the historical data for those stadiums in order to do this process. The game will not reference the era_ballparks file when importing players from the neutralized stats file, so if you change the era_ballparks file it will not affect the player imports or their ratings. I suggest taking a look at several players real and neutralized stats and decide for yourself. There is actually very little difference for most players. The biggest difference are Ruth and Ott as far as I know since they played in the friendliest HR environments so they end up getting the biggest adjustments. This is done for pitcher HR allowed too. So when you look at pitcher neutralized stats they will be the same as their real stats except for slight differences in HR allowed though. There game currently does have a version of the Bats L and Bats R historical data but it does not go all the way to 1901. Last edited by Garlon; 09-18-2021 at 05:03 PM.

09-17-2021, 02:16 PM	#22
Garlon Hall Of Famer Join Date: Jun 2004 Posts: 4,268	My suggestions are to look at a set of key players like Wagner, Cobb, Ruth, DiMaggio, Williams, Gwynn, Bonds, and Trout. If these players are significantly underperforming that means the entire league being squeezed.

09-18-2021, 09:20 AM	#25
Garlon Hall Of Famer Join Date: Jun 2004 Posts: 4,268	When you use real stats the game makes an adjustment to the HR rates of players based on their home park in the era_ballparks file. This is some internal OOTP adjustment so that you do not get so many extreme HR seasons. If you swap out the era_ballparks file to a neutral version of the file then players will import differently when using real stats because the game actually references that file. The platers get adjusted according to their historical home ballpark regardless of what team they end up playing for in your own game. When using neutralized stats there is no additional internal OOTP adjustment as all of the adjustments were done when those neutralized stats were put together so when you use neutralized stats the players will import the same way even if you change the era_ballparks file. Keep in mind the import and strategy settings can affect the results. My goal is to find a nice balance by having the best hitters play up to what you would expect and having the best pitchers perform accordingly too. If you want to do strict replays with real transactions then you may want to just use the default game settings.

09-18-2021, 10:31 AM	#26
Garlon Hall Of Famer Join Date: Jun 2004 Posts: 4,268	These are the results from a recent 1871-2020 game using my suggestions. Trades were disabled. OOTP vs Real Value 500 HR: 31 vs 28 50+ HR seasons: 44 vs 46 300+ Wins: 20 vs 24 250+ Wins: 65 vs 48

09-18-2021, 03:35 PM	#29
Garlon Hall Of Famer Join Date: Jun 2004 Posts: 4,268	Regarding the batting vs RHP and vs LHP, BigRod and I looked at the data in the database and realized it was incomplete so we decided we should rebuild those stats since we now have complete data for every batter every season since 1901. This is not yet in the game though as this needs to be formatted properly. There are some hurdles with this that needs to be worked out with player ids if I remember so I was not able to get to this. Regarding pitcher stamina, I think that Normal may be the best option as this setting applies to relievers too. Yes, finding the right setting for pitcher stamina has been something I have spent quite a bit of testing on and I think the results seem best on Normal. When you go above normal this can make relievers stronger, which in turn can diminish results from starting pitchers. I used to prefer start highest rested to get those extra starts from the #1 starters, but I think you get better winning percentages from your top pitchers on strict rotation because it makes the #5 starter make 32 starts instead of 24 starts and that has an effect across the league. There are many settings and many combinations that are possible to test but these seem to give me good results.

09-18-2021, 04:13 PM	#31
jdw All Star Starter Join Date: Nov 2003 Posts: 1,161	Great stuff, Garlon. I look forward to the eventual version of OOTP that has the splits in it. That will be cool.

09-18-2021, 05:39 PM	#36
Garlon Hall Of Famer Join Date: Jun 2004 Posts: 4,268	A reason I became interested in neutralized stats was the Willie Davis entry on pgs 740-743 of the Bill James Historical Abstract. Since Willie Davis spent most of his career with the Dodgers there is the question of whether he would have been a .300 in a neutral environment. Well, Bill James translates his career stats to a 750-run environment using the quadratic formula and has him as a .302 hitter. The quadratic method was used for previous generations of the neutralized stats for OOTP, but the new neutralized stats do not use that method, rather it uses the discrete BABIP/2B/3B/HR park factors for Bats R or Bats L or overall for Bats Both. If you use baseball reference there is an option to neutralize the player stats to different run scoring environments using the quadratic method. The default there is 723 runs per 162, whereas Bill James used 750 runs per 162 in his book. Willie Davis Real stats .279/.311/.412 OPS .723 BBR 723 runs scored per season .297/.329/.436 OPS .765 OOTP .283/.314/.431 OPS .745 Where Davis gains by adjusting his stats through discrete factors is his slugging. 2B/3B/HR Real 395/138/182 OOTP 422/161/205 This is because Dodger stadium was not really affecting batting average as much as slugging. The quadratic method assumes the run environment affects everything proportionately. So with the neutralized stats we do not have Davis as a .300 hitter, but he does gain nearly 20 points of slugging. In my recent game he was also +92 ZR in LF and +99 ZR in CF, so that is 191 runs saved on defense, which is about 19 wins.

09-19-2021, 12:31 PM	#38
Garlon Hall Of Famer Join Date: Jun 2004 Posts: 4,268	I may have a copy of the new format of the era_ballparks file with the quadratic park factors. I rebuilt the quadratic factors too for this game even though the game only loads the discrete values. The original quadratic factors were 5-yr values and I rebuilt the entire file using 3yr values.

09-19-2021, 02:35 PM	#40
Brad K Banned Join Date: May 2016 Location: St Petersburg Florida USA Posts: 6,693 Infractions: 0/2 (4)	I fully understand the need to prevent Maris from hitting 75 HRs in 1961. And I understand the reluctance to rate him for 61 and then use the game engine to impose a cap. A decision to reduce peak ratings makes sense. However people playing with optimized for single season set have a reasonable expectation Maris will hit in the 50s. And that isn't likely to happen. People playing with career settings have a reasonable expectation single season and career home run records will be approached.