New Ballpark Factors and Import Process - Page 4

Garlon · 01-04-2021, 03:04 PM

As for the player adjustments, if the quadratic factors are used then the neutralized player statistics which are normalized to a 750-run environment as per how the runs were created in that specific season should be used. If the discrete factors are used then the process as I described earlier should be used. The spreadsheet that you see that generates park factors is just something someone made up. The real events are the true park factors. There are no absolute park factors for any stadium. It is only about how they compare to the other stadiums in a league with the same conditions of a given season. If you make Wrigley Field the ballpark for every team in your league, then no matter what the park factors are for Wrigley they will end up playing as 1.000 as the park factors do not influence the league totals.

The stolen base percentage issue can easily be addressed with strategy settings. If you want the best stealers to increase their percentage set stealing bases to very often. What happens in OOTP is that on Normal or Rarely the weaker bases stealers are less likely to attempt a SB and therefore the caught stealing has to get redistributed somewhere. On Rarely you will have your SB leaders get more SB attempts but lower success rate, and on Very Often your leaders will be lower but have a better success rate. My suggestion is to simply increase the strategy setting to improve the success rates of the top players.

Brad K · 01-04-2021, 03:09 PM

The developers have stated there's a problem with base stealing and that it will be fixed.

Lukas Berger · 01-05-2021, 07:29 AM

Quote:

Originally Posted by Brad K

The developers have stated there's a problem with base stealing and that it will be fixed.

I don't think we stated that, at least if you're referring to what I said a while back. What I said is just that we are going to look at things to determine if there is a problem or not.

Lukas Berger · 01-05-2021, 07:37 AM

Quote:

Originally Posted by Brad K

Ignoring the first adjustment that needs to be made to the run production function is a serious problem. This is the first time I've had a negative impression of the quality of the game, and its huge.

I'll preface this by saying that this is my understanding, not totally official gospel, Markus or someone who knows better could correct me if I'm wrong.

That being said, neutralized stats were more useful/necessary in previous versions of the game, long ago when we had a different algorithm for generating ratings, but the way we deal with the normal stats now and translating them into ratings is essentially neutralizing them already and if you're using actual neutralized stats, it's a bit redundant.

We didn't remove the feature totally for one reason or another, but it's not really necessary or even particularly desirable to play using the neutralized stats with any relatively recent version of OOTP.

Brad K · 01-05-2021, 12:22 PM

Quote:

Originally Posted by Lukas Berger

I don't think we stated that, at least if you're referring to what I said a while back. What I said is just that we are going to look at things to determine if there is a problem or not.

Well then you're on your way. It shouldn't be any effort to confirm the data I posted. Thanks for your efforts. I understand the need to confirm outside data.

Brad K · 01-05-2021, 12:26 PM

Quote:

Originally Posted by Lukas Berger

I'll preface this by saying that this is my understanding, not totally official gospel, Markus or someone who knows better could correct me if I'm wrong.

That being said, neutralized stats were more useful/necessary in previous versions of the game, long ago when we had a different algorithm for generating ratings, but the way we deal with the normal stats now and translating them into ratings is essentially neutralizing them already and if you're using actual neutralized stats, it's a bit redundant.

We didn't remove the feature totally for one reason or another, but it's not really necessary or even particularly desirable to play using the neutralized stats with any relatively recent version of OOTP.

That's all very interesting. So the ratings are pre-neutralized!

My view is that neutralized stats and park factors went hand in hand. Do you recommend park factors?

Brad K · 01-05-2021, 12:29 PM

PS I really appreciate the developers occasionally jumping into discussions. Correctly, its not done "all the time" just at especially important times.

Lukas Berger · 01-05-2021, 02:55 PM

Quote:

Originally Posted by Brad K

That's all very interesting. So the ratings are pre-neutralized!

My view is that neutralized stats and park factors went hand in hand. Do you recommend park factors?

We use them by default, so I guess so, but tbh Garlon probably knows more about that specific area of the game than I do, so I would go with his recommendations on that without a qualm.

Brad K · 01-06-2021, 08:13 PM

Quote:

Originally Posted by Garlon

The spreadsheet that you see that generates park factors is just something someone made up.

That's quite dismissive. It appears to me some thought went into it along with some reasonably supportable calculations and formulas.

Quote:

The real events are the true park factors.

Depends. Clearly they represent what WAS. but the next question is does "what was" hit the center of the bell curve on possible outcomes. I don't see how it could be argued that a factor over 2 something on the Giants mentioned earlier does.

But with neutralized stats I supposed that will work out OK if its a completely historical games with all players on their historical teams. Lots of people don't play that way.

So start putting players who weren't Giants in that ballpark will, I think, give results that can't be defended as realistic.

Quote:

There are no absolute park factors for any stadium.

I understand that. But that's not a valid objection to a well made calculator. As long as the parks are comparatively realistic the season totals feature makes it right.

Garlon · 01-07-2021, 03:00 AM

You will get park factors for HR over 2 when you have so few HR in the league and when you have a park like Polo Grounds that is 277 down the LF line and Griffith Stadium that is 464 down the LF line in the same set of ballparks. You can put any players into these parks with these factors from the corresponding season and they will play realistically, but only with the corresponding import process as described and only if you are using that given season. The players do not need to be on their historical teams.

These factors are not meant to be used out of the context of that season. If you want the 1998 Yankees to play in the Polo Grounds, then perhaps have OOTP generate the factors or use the calculator someone made. If you want to replay history then these discrete factors are the best way to do so, but not until there is proper implementation. I am concerned that with the park factors that OOTP is applying them twice for outcomes, once on the pitcher and once on the batter, so if a park that has a 1.25 HR factor is actually ending up like 1.25*1.25 = 1.5625 if OOTP is calculating a 25% increase in the probability of a HR allowed by the pitcher and a 25% increase in the probability of the batter hitting a HR too. The factors should only get applied to the batter.

Brad K · 01-07-2021, 07:28 PM

Quote:

Originally Posted by Garlon

I am concerned that with the park factors that OOTP is applying them twice for outcomes, once on the pitcher and once on the batter, so if a park that has a 1.25 HR factor is actually ending up like 1.25*1.25 = 1.5625 if OOTP is calculating a 25% increase in the probability of a HR allowed by the pitcher and a 25% increase in the probability of the batter hitting a HR too. The factors should only get applied to the batter.

Why do you think that's possibly occurring?

I'm still puzzling over how stats can be neutralized in the ratings creation from a non neutralized database.

I am close to starting another game so need to make some decisions soon on settings. My current game is 1951 to the end of Barry Bonds career. Its 2003 and he's still playing.

Garlon · 01-08-2021, 12:58 PM

I think that the distribution of home and away results for HR was impossible based on the park factors I was using. Something is amiss with how the factors are being used in the probability outcomes.

The stats are being neutralized by referencing the era_ballparks file. Unfortunately, the neutralization process is flawed. If you put the neutral park factors file into the game you will see that your player imports are different.

Brad K · 01-08-2021, 06:04 PM

I don't understand. If the game is using era_ballparks to neutralize player ratings then the ratings should be different with a different era_ballparks file.

Garlon · 01-08-2021, 06:59 PM

Try the different era_balpark files, the original one with the quadratic factors, the one I posted with the discrete factors, and the neutral one with all the parks set to 1 for the factors. If you begin a new league with each of those files in place the player ratings will be different.

Brad K · 01-08-2021, 08:41 PM

Wouldn't this system be using the factors from the players home ball park as in the era_ballparks file? If so player ratings should be different when era_ballparks is different.

I'm not saying I favor this method - I don't - but I'm not seeing why this difference is a problem. What the neutralized database for players did (if I understand it correctly) is have neutralized stats based on historical park factors, essentially some version of era_ballparks. The calculation was done in advance rather than being done at player creation.

Brad K · 01-08-2021, 08:47 PM

Soooo... in my current game Willie Stargell hit 580 HRs and was a superstar even in Forbes Field. I wonder if my settings of neutralized stats caused this? Because it appears that he got neutralized twice. Plus at some point I imported Forbes from Silvan's to have the graphics. I don't recall if when importing I edited in the default era_ballparks factors.

I do know that if I didn't they ran for the duration of Forbes because I didn't change the park name to exactly match what's in era_ballparks. Also Roberto Clemente had a 30 some WAR career which might be partly caused by Silvan's factors.

Brad K · 01-31-2021, 10:01 PM

Quote:

Originally Posted by Brad K

Soooo... in my current game Willie Stargell hit 580 HRs and was a superstar even in Forbes Field. I wonder if my settings of neutralized stats caused this? Because it appears that he got neutralized twice. Plus at some point I imported Forbes from Silvan's to have the graphics. I don't recall if when importing I edited in the default era_ballparks factors.

I did some experimentation. It seems the option for neutralized stats double neutralizes them. So it should be removed. Real stats are already neutralized unless Garlon's 1111 park factors file is used.

I don't see much use for non neutralized but some people might like it. If provided the era_ballparks file could be removed from the process of ratings creation.

Perhaps the two could be named with park factors and without park factors.

Garlon · 02-01-2021, 02:24 PM

I took a look at the season Ted Williams had in 1941 when he batted .406/.553/.735/1.287

The factors for LHB in 1941 Fenway Park were
BABIP: 0.942
2B: 1.186
3B: 0.776
HR: 0.893

After adjustments he gains 6 singles, gets 2 fewer doubles, and 2 more HR for the season, providing 10 more total bases. This gives him .419/.563/.757/1.319 for the season.

When OOTP uses neutralized stats the game should not be trying to make another neutralization adjustment based upon the ballpark factors since that defeats the purpose of the file.

Brad K · 02-09-2021, 07:36 AM

Can you imagine a guy walking up to the plate and the opposition knows he's going to touch first base safely over half the time? Incredible.

01-04-2021, 03:04 PM	#61
Garlon Hall Of Famer Join Date: Jun 2004 Posts: 4,274	As for the player adjustments, if the quadratic factors are used then the neutralized player statistics which are normalized to a 750-run environment as per how the runs were created in that specific season should be used. If the discrete factors are used then the process as I described earlier should be used. The spreadsheet that you see that generates park factors is just something someone made up. The real events are the true park factors. There are no absolute park factors for any stadium. It is only about how they compare to the other stadiums in a league with the same conditions of a given season. If you make Wrigley Field the ballpark for every team in your league, then no matter what the park factors are for Wrigley they will end up playing as 1.000 as the park factors do not influence the league totals. The stolen base percentage issue can easily be addressed with strategy settings. If you want the best stealers to increase their percentage set stealing bases to very often. What happens in OOTP is that on Normal or Rarely the weaker bases stealers are less likely to attempt a SB and therefore the caught stealing has to get redistributed somewhere. On Rarely you will have your SB leaders get more SB attempts but lower success rate, and on Very Often your leaders will be lower but have a better success rate. My suggestion is to simply increase the strategy setting to improve the success rates of the top players. Last edited by Garlon; 01-04-2021 at 03:09 PM.

01-04-2021, 03:09 PM	#62
Brad K Banned Join Date: May 2016 Location: St Petersburg Florida USA Posts: 6,693 Infractions: 0/2 (4)	The developers have stated there's a problem with base stealing and that it will be fixed. Last edited by Brad K; 01-04-2021 at 03:13 PM.

01-07-2021, 03:00 AM	#70
Garlon Hall Of Famer Join Date: Jun 2004 Posts: 4,274	You will get park factors for HR over 2 when you have so few HR in the league and when you have a park like Polo Grounds that is 277 down the LF line and Griffith Stadium that is 464 down the LF line in the same set of ballparks. You can put any players into these parks with these factors from the corresponding season and they will play realistically, but only with the corresponding import process as described and only if you are using that given season. The players do not need to be on their historical teams. These factors are not meant to be used out of the context of that season. If you want the 1998 Yankees to play in the Polo Grounds, then perhaps have OOTP generate the factors or use the calculator someone made. If you want to replay history then these discrete factors are the best way to do so, but not until there is proper implementation. I am concerned that with the park factors that OOTP is applying them twice for outcomes, once on the pitcher and once on the batter, so if a park that has a 1.25 HR factor is actually ending up like 1.251.25 = 1.5625 if OOTP is calculating a 25% increase in the probability of a HR allowed by the pitcher and a 25% increase in the probability of the batter hitting a HR too. The factors should only get applied to the batter. Last edited by Garlon; 01-07-2021 at 09:36 AM.*

01-05-2021, 12:29 PM	#67
Brad K Banned Join Date: May 2016 Location: St Petersburg Florida USA Posts: 6,693 Infractions: 0/2 (4)	PS I really appreciate the developers occasionally jumping into discussions. Correctly, its not done "all the time" just at especially important times.

01-08-2021, 12:58 PM	#72
Garlon Hall Of Famer Join Date: Jun 2004 Posts: 4,274	I think that the distribution of home and away results for HR was impossible based on the park factors I was using. Something is amiss with how the factors are being used in the probability outcomes. The stats are being neutralized by referencing the era_ballparks file. Unfortunately, the neutralization process is flawed. If you put the neutral park factors file into the game you will see that your player imports are different.

01-08-2021, 06:04 PM	#73
Brad K Banned Join Date: May 2016 Location: St Petersburg Florida USA Posts: 6,693 Infractions: 0/2 (4)	I don't understand. If the game is using era_ballparks to neutralize player ratings then the ratings should be different with a different era_ballparks file.

01-08-2021, 06:59 PM	#74
Garlon Hall Of Famer Join Date: Jun 2004 Posts: 4,274	Try the different era_balpark files, the original one with the quadratic factors, the one I posted with the discrete factors, and the neutral one with all the parks set to 1 for the factors. If you begin a new league with each of those files in place the player ratings will be different.

01-08-2021, 08:41 PM	#75
Brad K Banned Join Date: May 2016 Location: St Petersburg Florida USA Posts: 6,693 Infractions: 0/2 (4)	Wouldn't this system be using the factors from the players home ball park as in the era_ballparks file? If so player ratings should be different when era_ballparks is different. I'm not saying I favor this method - I don't - but I'm not seeing why this difference is a problem. What the neutralized database for players did (if I understand it correctly) is have neutralized stats based on historical park factors, essentially some version of era_ballparks. The calculation was done in advance rather than being done at player creation.

01-08-2021, 08:47 PM	#76
Brad K Banned Join Date: May 2016 Location: St Petersburg Florida USA Posts: 6,693 Infractions: 0/2 (4)	Soooo... in my current game Willie Stargell hit 580 HRs and was a superstar even in Forbes Field. I wonder if my settings of neutralized stats caused this? Because it appears that he got neutralized twice. Plus at some point I imported Forbes from Silvan's to have the graphics. I don't recall if when importing I edited in the default era_ballparks factors. I do know that if I didn't they ran for the duration of Forbes because I didn't change the park name to exactly match what's in era_ballparks. Also Roberto Clemente had a 30 some WAR career which might be partly caused by Silvan's factors.

02-01-2021, 02:24 PM	#78
Garlon Hall Of Famer Join Date: Jun 2004 Posts: 4,274	I took a look at the season Ted Williams had in 1941 when he batted .406/.553/.735/1.287 The factors for LHB in 1941 Fenway Park were BABIP: 0.942 2B: 1.186 3B: 0.776 HR: 0.893 After adjustments he gains 6 singles, gets 2 fewer doubles, and 2 more HR for the season, providing 10 more total bases. This gives him .419/.563/.757/1.319 for the season. When OOTP uses neutralized stats the game should not be trying to make another neutralization adjustment based upon the ballpark factors since that defeats the purpose of the file.

02-09-2021, 07:36 AM	#79
Brad K Banned Join Date: May 2016 Location: St Petersburg Florida USA Posts: 6,693 Infractions: 0/2 (4)	Can you imagine a guy walking up to the plate and the opposition knows he's going to touch first base safely over half the time? Incredible.