Ballpark factors and league totals

No Pepper · 08-11-2006, 08:39 AM

I created a fictional universe and have been tweaking league total modifiers over several sim years to get numbers to where I'd like them. However in doing this, I've "forgotten" to assign ballparks to the teams.

I notice that the game generates a fictional ballpark for each team with unique dimensions and ballpark factors.

My question is, if I were to adjust these factors, would that change my league totals or just the distribution? Meaning, if I have 100 league homeruns as determined by my league total modifiers, if I change my ballpark factors, I'll still have 100 league homeruns? Or would changing ballpark factors alone alter my league totals?

1998 Yankees · 08-11-2006, 09:08 AM

No Pepper,

I'm no expert on OOTP2006 at this point, but I did do a lot of thinking, and editing, of ballpark factors because after scanning the guide and this forum extensively, I realized that no one has any idea of how much to scale these numbers and why.

Some sources seem to base ballpark factors on recent player performances; I think the fellow who provided the MLB stadium data (Halos 17: http://www.ootpdevelopments.com/boar...d.php?t=123302) quoted Bill James for his ballpark factors, but why would Comerica Park have a triples factor of 203 (2.030)? Maybe there were some fast in-the-gap hitters on the Tigers in recent years? I would not want that to be the reason why my fictional Tigers hit twice as many triples as any one else.

Other approaches to ballpark factors have involved Excel formulas that calculate the factors from the stadium dimensions. This approach seemed more logical to me but I noticed that the results varied according to whose model I used. Or what version.

One thing that I think I can be sure on; the stadium dimensions are there for reference only; your changing the wall distances and heights has no effect on the factors.

If you want some advice from a mere "bat boy" at this point, you may want to consider doing what I did which is to set all ballpark factors to 1.000 to avoid upsetting the balance of your results. At least until OOTP clarifies what the proper range of variation should be and what can be expected at the extremes.

Some folks in this forum wisely say that you should test, test, test to get the model right, and I agree up to a point. I just don't want to have to test every last detail of this model, down to individual ballpark factors.

Carplos · 08-11-2006, 09:27 AM

Quote:

My question is, if I were to adjust these factors, would that change my league totals or just the distribution? Meaning, if I have 100 league homeruns as determined by my league total modifiers, if I change my ballpark factors, I'll still have 100 league homeruns? Or would changing ballpark factors alone alter my league totals?

Yeah, they will affect the league totals. I've done the same thing as you and then had to readjust totals to get them back in line where I want.

No Pepper · 08-11-2006, 09:32 AM

Thanks Yankee. I'd set factors to 1 but the game already has generated park factors (non-1.00 values) for my parks, and I'm afraid if I change those to a standard 1.00 across the board, it'll affect the league numbers I got while adjusting league total modifiers. Of course, you are right: test, test, test.

I'm not focused on where particular hits occur, just that they do to match the league totals I want.

The game guide says very little about park factors, but does say dimensions "also impact the results of games played...such as judging a player's ability
to rob a home run."

Carplos: Arrrgh!!! haha Interesting that the game generates unique park factors though. I've been thinking about releasing my league as a template but if what you say is true, releasing a template (as far as I understand it to be) won't be worth squat because park factors will always be different, or, have to be edited by the player. A quickstart is limiting too for a fictional universe, unless if you release and delete all players and refill. Hrmmm, workarounds!

andymac · 08-11-2006, 09:33 AM

Quote:

Originally Posted by Carplos

Yeah, they will affect the league totals. I've done the same thing as you and then had to readjust totals to get them back in line where I want.

Yea, you want to make sure that your park factors balance if possible (I'm not sure if OOTP does this at default or not, might have to test it when I get a chance). Because if, for instance, the HR factors for all the parks average out to be 1.05 or something than your league is likely going to have about 5% more home runs than you set it to have in the league totals. Obviously they are other factors and luck that wouldn't make that exact, but that is the general idea behind how they work.

RonCo · 08-11-2006, 11:50 AM

Andymac is correct. Actual statistical performance is based on:

1) League totals
2) Park factors (which are different in OOTP than in Sabermetrics)
3) Talent levels in the league

Talent levels are constantly changing, and can be completely out of whack if you're using edited players/rosters--something I haven't heard many folks talk about.

And if you want to use league totals to "accurately" adjust composite stats, the system is designed to assume that the league average of any OOTP park factor is 1.00. [i.e. all doubles factors average out to 1.00] If this isn't true, then you'll get odd skewings of totals required to acheive your goals. This in itself isn't bad so long as you know what you're doing.

The whole thing is poorly understood around the community, though...or maybe it's just me. I had to work on it for weeks and weeks, and read a ton of board posts before the mechanism sunk into my thick head.

andymac · 08-11-2006, 11:54 AM

It would be nice if the game somewhere totaled up the park factors and told you how much they are changing things. A "league park factor" of sorts.

1998 Yankees · 08-11-2006, 12:23 PM

Quote:

Originally Posted by No Pepper

Interesting that the game generates unique park factors though.

Yes, and what's even more interesting is that if you look closely at the generated ballpark factors, they really don't make sense given the stadium dimensions. For example, a short right-field with a low wall may have a left-handed HR factor of less than 1.000. That truly says to me that these factors are random. I hear what you say about not wanting to upset your balance by editing them, once you've started up your template. That, and think about what I went through editing 368 stadiums (30 MLB, 180 minor, 86 foreign, 72 independent) reverting all ballpark factors to 1.000!

RonCo · 08-11-2006, 01:31 PM

Stadium dimensions are meaningless, though as the manual says, a low fence will result in some fly outs being "saved home runs" and long or short fences will change the sugar-coating of the play-by-play to result in longer or shorter HR travel distances. The number of HR/2B/3B etc. will not be influenced at all by the dimensions of a ballpark--and, as you mentioned, the created versions will often be fundamentally improbable.

This springs, again, from starting with a poor model and then trying to make it do more than it was designed for.

I note, that some people like this approach because, for example, they think it allows them to base a league on the moon and get "accurate" 900 foot HRs. Despite the fact that this assumption is not correct, I'll give them their argument...but I note that a model that would fit everyone's needs would be one that determined fly ball distance based on a single point of reference (say sea level at 80 deg F on a windless day), then allowed modification for gravity (or altitude).

Zeyes · 08-11-2006, 02:40 PM

Quote:

Originally Posted by andymac

It would be nice if the game somewhere totaled up the park factors and told you how much they are changing things. A "league park factor" of sorts.

FWIW, it's easy to do that by exporting to csv and work on parks.csv in Excel (or similar). Just calculate the average of each factor type, and you'll have what you're looking for (presuming that each ballpark hosts the same number of games in your league, which is probably true for 99.9% of all conceivable setups).

BTW, for the fictional league I just created to test this, the average ballpark factors come out like this:

Avg: 1.013
Avg LHB: 1.019
Avg RHB: 1.008
2B: 1.004
3B: 0.990
HR: 1.013
HR LHB: 1.017
HR RHB: 1.011

No idea if it's just coincidence, but that looks pretty close to 1 for all of them, so perhaps OOTP does try to balance the league overall. Anyway, no matter if it actively balances the values or just repeatedly picks values along a normal distribution or somesuch, I would guess the league size also plays a role, with a 30-team league likely having a higher probability to get averages close to 1 than a league with, say, 8 teams.

Quote:

Originally Posted by No Pepper

I've been thinking about releasing my league as a template but if what you say is true, releasing a template (as far as I understand it to be) won't be worth squat because park factors will always be different, or, have to be edited by the player. A quickstart is limiting too for a fictional universe, unless if you release and delete all players and refill. Hrmmm, workarounds!

As an, erm, workaround

you could release the parks.dat file for a properly tuned league along with the template. People would have to create a league based on your template, then close the game, replace parks.dat and reload the league for that to work, though. And they couldn't make any changes to your template (particularly the league structure) that would affect the team layout, I guess.

andymac · 08-11-2006, 02:40 PM

Quote:

Originally Posted by RonCo

but I note that a model that would fit everyone's needs would be one that determined fly ball distance based on a single point of reference (say sea level at 80 deg F on a windless day), then allowed modification for gravity (or altitude).

Yep, this would be the ideal solution.

andymac · 08-11-2006, 02:44 PM

Quote:

Originally Posted by Zeyes

No idea if it's just coincidence, but that looks pretty close to 1 for all of them, so perhaps OOTP does try to balance the league overall. If so, I would guess the league size also plays a role, with a 30-team league likely having a higher probability to get averages close to 1 than a league with, say, 8 teams.

My guess is that the random generation is based around 1. So, it would make sense that most times the numbers are going to come out close to 1. However, there is always the chance that you are going to get a set of parks that is skewed a significant amount in one direction or the other. I hadn't thought about using the csv export to check that which wouldn't be difficult but I maintain it would be nice if either the game generated that automatically and/or made sure that its' factors balanced.

Zeyes · 08-11-2006, 02:50 PM

Quote:

Originally Posted by andymac

My guess is that the random generation is based around 1. So, it would make sense that most times the numbers are going to come out close to 1. However, there is always the chance that you are going to get a set of parks that is skewed a significant amount in one direction or the other.

Yep, hence my suspicion that more parks = better fit, at least on average. It would be nice if the game does actively balance things, but I kinda doubt it...although the slight deviations in my test might be intentional, who knows. Anyway, it wouldn't be hard for the game to balance things...just generate the parks as usual, then calculate the averages for each factor and use those to scale the generated values to 1.

Quote:

Originally Posted by andymac

I hadn't thought about using the csv export to check that which wouldn't be difficult but I maintain it would be nice if either the game generated that automatically ...

I did start my post with "FWIW".

I agree it would be nice to have that available in-game, but I wouldn't consider it a huge priority, given it's not much work to do it externally right now.

No Pepper · 08-11-2006, 04:08 PM

Thanks guys, its always good to see these kinds of discussions. If the game randomizes park factors but they average around 1 depending on league size, I guess its "safe to assume" that if I do change them to suit my ballparks, maintaining the same average, I shouldn't drastically see a change in my league totals. But if I do deviate from that average, I should expect change.

Thanks for the csv export and parks.dat suggestions, I'll have to learn how that works.

More learning than actual playing, or are the two synonymous?

Curtis · 09-28-2006, 04:34 PM

I realize it's just a statistical aberration, but when I was setting up one of my early leagues, every team in 'my' division (the northeast) ended up with parks with .8-.95 ballpark factors, while every team in California ended up with 1.1-1.25 factors. I thought that gave a little extra flavor to the game.

08-11-2006, 08:39 AM	#1
No Pepper All Star Starter Join Date: Jan 2003 Location: Milwaukee, WI Posts: 1,570	Ballpark factors and league totals I created a fictional universe and have been tweaking league total modifiers over several sim years to get numbers to where I'd like them. However in doing this, I've "forgotten" to assign ballparks to the teams. I notice that the game generates a fictional ballpark for each team with unique dimensions and ballpark factors. My question is, if I were to adjust these factors, would that change my league totals or just the distribution? Meaning, if I have 100 league homeruns as determined by my league total modifiers, if I change my ballpark factors, I'll still have 100 league homeruns? Or would changing ballpark factors alone alter my league totals? __________________ Uniforms compatible with OOTP23/24 Historical Major League Baseball 1901-current Historical Major League Baseball 1871-1900 Historical Federal League Historical Negro Leagues Last edited by No Pepper; 08-11-2006 at 08:40 AM.

08-11-2006, 09:32 AM	#4
No Pepper All Star Starter Join Date: Jan 2003 Location: Milwaukee, WI Posts: 1,570	Thanks Yankee. I'd set factors to 1 but the game already has generated park factors (non-1.00 values) for my parks, and I'm afraid if I change those to a standard 1.00 across the board, it'll affect the league numbers I got while adjusting league total modifiers. Of course, you are right: test, test, test. I'm not focused on where particular hits occur, just that they do to match the league totals I want. The game guide says very little about park factors, but does say dimensions "also impact the results of games played...such as judging a player's ability to rob a home run." Carplos: Arrrgh!!! haha Interesting that the game generates unique park factors though. I've been thinking about releasing my league as a template but if what you say is true, releasing a template (as far as I understand it to be) won't be worth squat because park factors will always be different, or, have to be edited by the player. A quickstart is limiting too for a fictional universe, unless if you release and delete all players and refill. Hrmmm, workarounds! __________________ Uniforms compatible with OOTP23/24 Historical Major League Baseball 1901-current Historical Major League Baseball 1871-1900 Historical Federal League Historical Negro Leagues Last edited by No Pepper; 08-11-2006 at 09:37 AM.

08-11-2006, 11:50 AM	#6
RonCo Hall Of Famer Join Date: Aug 2003 Posts: 10,502	Andymac is correct. Actual statistical performance is based on: 1) League totals 2) Park factors (which are different in OOTP than in Sabermetrics) 3) Talent levels in the league Talent levels are constantly changing, and can be completely out of whack if you're using edited players/rosters--something I haven't heard many folks talk about. And if you want to use league totals to "accurately" adjust composite stats, the system is designed to assume that the league average of any OOTP park factor is 1.00. [i.e. all doubles factors average out to 1.00] If this isn't true, then you'll get odd skewings of totals required to acheive your goals. This in itself isn't bad so long as you know what you're doing. The whole thing is poorly understood around the community, though...or maybe it's just me. I had to work on it for weeks and weeks, and read a ton of board posts before the mechanism sunk into my thick head. Last edited by RonCo; 08-11-2006 at 01:23 PM.

08-11-2006, 11:54 AM	#7
andymac Hall Of Famer Join Date: Feb 2002 Location: Effingham, IL Posts: 5,725	It would be nice if the game somewhere totaled up the park factors and told you how much they are changing things. A "league park factor" of sorts. __________________ June Madness: Links FTB: andymac

08-11-2006, 01:31 PM	#9
RonCo Hall Of Famer Join Date: Aug 2003 Posts: 10,502	Stadium dimensions are meaningless, though as the manual says, a low fence will result in some fly outs being "saved home runs" and long or short fences will change the sugar-coating of the play-by-play to result in longer or shorter HR travel distances. The number of HR/2B/3B etc. will not be influenced at all by the dimensions of a ballpark--and, as you mentioned, the created versions will often be fundamentally improbable. This springs, again, from starting with a poor model and then trying to make it do more than it was designed for. I note, that some people like this approach because, for example, they think it allows them to base a league on the moon and get "accurate" 900 foot HRs. Despite the fact that this assumption is not correct, I'll give them their argument...but I note that a model that would fit everyone's needs would be one that determined fly ball distance based on a single point of reference (say sea level at 80 deg F on a windless day), then allowed modification for gravity (or altitude). Last edited by RonCo; 08-11-2006 at 02:52 PM.

08-11-2006, 09:08 AM	#2
1998 Yankees Hall Of Famer Join Date: Aug 2006 Location: Yankee Stadium, back in 1998. Posts: 8,645	No Pepper, I'm no expert on OOTP2006 at this point, but I did do a lot of thinking, and editing, of ballpark factors because after scanning the guide and this forum extensively, I realized that no one has any idea of how much to scale these numbers and why. Some sources seem to base ballpark factors on recent player performances; I think the fellow who provided the MLB stadium data (Halos 17: http://www.ootpdevelopments.com/boar...d.php?t=123302) quoted Bill James for his ballpark factors, but why would Comerica Park have a triples factor of 203 (2.030)? Maybe there were some fast in-the-gap hitters on the Tigers in recent years? I would not want that to be the reason why my fictional Tigers hit twice as many triples as any one else. Other approaches to ballpark factors have involved Excel formulas that calculate the factors from the stadium dimensions. This approach seemed more logical to me but I noticed that the results varied according to whose model I used. Or what version. One thing that I think I can be sure on; the stadium dimensions are there for reference only; your changing the wall distances and heights has no effect on the factors. If you want some advice from a mere "bat boy" at this point, you may want to consider doing what I did which is to set all ballpark factors to 1.000 to avoid upsetting the balance of your results. At least until OOTP clarifies what the proper range of variation should be and what can be expected at the extremes. Some folks in this forum wisely say that you should test, test, test to get the model right, and I agree up to a point. I just don't want to have to test every last detail of this model, down to individual ballpark factors.

08-11-2006, 04:08 PM	#14
No Pepper All Star Starter Join Date: Jan 2003 Location: Milwaukee, WI Posts: 1,570	Thanks guys, its always good to see these kinds of discussions. If the game randomizes park factors but they average around 1 depending on league size, I guess its "safe to assume" that if I do change them to suit my ballparks, maintaining the same average, I shouldn't drastically see a change in my league totals. But if I do deviate from that average, I should expect change. Thanks for the csv export and parks.dat suggestions, I'll have to learn how that works. More learning than actual playing, or are the two synonymous? __________________ Uniforms compatible with OOTP23/24 Historical Major League Baseball 1901-current Historical Major League Baseball 1871-1900 Historical Federal League Historical Negro Leagues

09-28-2006, 04:34 PM	#15
Curtis Hall Of Famer Join Date: Jul 2006 Location: Watertown, New York Posts: 4,567	I realize it's just a statistical aberration, but when I was setting up one of my early leagues, every team in 'my' division (the northeast) ended up with parks with .8-.95 ballpark factors, while every team in California ended up with 1.1-1.25 factors. I thought that gave a little extra flavor to the game.