ootp park factors question

[HonusWagner]

Who here has set park factors to 1.000 for all league parks? Am beginning to think this may be the way to go. Currently am preferring the notion that each park's dimensions / quirks add to the gameplay (especially in the classic parks from the '20s). But keep hearing things like dimensions not effecting HR. Haven't done a too-thorough overview on the stats to make any conclusions, and old threads may be outdated. However, am having some trouble visualizing how my park (Baker Bowl) is performing this year: apparently it is the slugger's paradise, at least for HR. This seems aligned with the generally smaller dimensions compared to many of the other league parks. But since there will be more RHB than LHB I am not seeing anything in the LF and L-CF dimensions to indicate why RHB performance should be taking hits (npi).

Then compare to Exposition Park - massive dimensions seem to reflect performance outcomes (rare HR, higher relative AVE, EBH).

Then compare to Shibe Park - not deep LF, RF but relatively fewer HR.

[MathBandit]

I think I might be misunderstanding something. Why would I want to set all park factors to 1.000- so that Coors Field and Fenway play exactly the same as T-Mobile Field and Petco Park do?

[HonusWagner]

Originally Posted by MathBandit

I think I might be misunderstanding something. Why would I want to set all park factors to 1.000- so that Coors Field and Fenway play exactly the same as T-Mobile Field and Petco Park do?

I guess the question is variance vs accuracy. If you simply want variance between parks then you're right, obviously. Setting all factors to a default removes arbitrariness, which is when the variances are not really accurate reflections of park characteristics. Also, in historical gameplay park factors which aren't terribly accurate could create a hinderance to accurate outcomes.

[Brad K]

What's an accurate outcome? That a player produces close to real life statistics regardless of where he plays?

[HonusWagner]

Originally Posted by Brad K

What's an accurate outcome? That a player produces close to real life statistics regardless of where he plays?

Yes, since in this example the player's performance would closely reflect RL performance. So, yes that's one potential example of accuracy, but not the only one.

[Brad K]

Different inputs but the same output. To each his own.

[RonCo]

Your park factors are set to hurt RHB (or to aid LHB more).

Per the user's manual:

Ballpark Factors
The Ballpark Factors section directly impacts how hitters perform in the selected ballpark, over a large sample size. Ballpark Factors are extremely important, as they directly affect the outcome of plays in a given park.

Ballpark factors are based on a "norm" of 1.000. That is, a ballpark with all 1.000 factors is essentially a "neutral" park where hitters will all perform similarly.


Ballpark Dimensions
The Ballpark Dimensions section describes the size of the selected ballpark.

Ballpark dimensions are purely cosmetic! It's the ballpark factors that actually influence the game engine. That is, making a ballpark with a 500-foot center field won't make it any harder to hit home runs to center if the home run factors are still 1.000.

So the important figure here for raw stats output is the park factor, not the dimensions. Making the dimensions right is valuable only to the point that it makes the play by play and game log descriptions of home runs feel right (or more immersive).

If you want "accurate" stats, get the factors right for each park. If you want "accurate" descriptions, get the dimensions right for each park.

[RonCo]

Some nuance applies here, though.

Variance of player's performance between ballparks is associated with the variance of park factors at each park. If you set all factors to 1.000, all players will perform "similarly" (meaning, according to their ratings against each other). But the same is essentially true if you set all park factors to any value. Put them all at 1.5 and everyone gets the same % boost to offense. Put them all at .5 and everyone gets the same % cut.

Adjust the fences everywhere (or anywhere), though, and players will simply hit home runs of shorter or longer distances.

[Pelican]

I see several issues here. One is what I call. The Chuck Klein issue. When the Phillies moved from Baker Bowl to Shibe Park, in the middle of the 1938 Season, it seemed to contribute to a sharp decline in HR, particularly by Klein. Or, he was getting old and slowing down.

Klein still has very good OOTP ratings in 1938, and of course part of that comes from hitting in a tiny park with a short RF (with a high fence). For half the season. Despite his power ratings, he is not slugging much at Shibe. That is how park ratings should affect the game.

But your park ratings look flawed to me. I don’t believe that Shibe was over 500 feet to CF, ever. LF also looks too deep.

One thing that bothers me is how the game recomputes park ratings based on a change in dimensions. Long story short, it you move back the fences, you’ll see an expected increase in HR, but when you move them back to where they were, the park ratings will be different from where you started. Don’t get that.

The other issue I have with the park ratings generator is that it associates 2B and 3B and average with large parks. Even a small park like Baker Bowl, which caused pop fly doubles and triples with its high RF fence, has ratings well below 1.00. And the AVG rating below 1.00 makes no sense. Teams (even the Phillies) raked at Baker Bowl.

I love the old ballparks and the crazy results from their odd irregular fences. (I like the new retro parks like Camden Yards and Citizens Bank for the same reason.) So I won’t be switching to across-the-board 1.000 settings anytime soon. But I can understand the temptation to do that, and neutralize the park factors. Chuck Klein would not have liked it.

[Syd Thrift]

The big issue with using parks with historical play isn't so much that the park factors aren't "right", it's that the game doesn't take them into account when generating ratings. To me, this becomes one more reason to play with fictional players instead of historical ones but I think that it'd be a lot better to apply those factors in as a whole, even if there were years where, like, Joe Dimaggio wasn't completely eaten by the Yankee Stadium left field or whatever. And while it's generally true that some players play better in parks than others, the game is simply not that granular. You play in Wrigley in the 70s, all lefties get like a 20% boost to HRs, not just guys who hit the alleys well, everyone. Pitchers might be a bit more tricky inasmuch as situational lefties play a ton of LHBs whereas lefty starters barely face any of them but I think you can come to a conclusion by looking at the splits by handedness.

As a side note, too, I would really like to add park factors for walks and strikeouts and maybe even errors (which I'm not sure are included in the Seamheads DB I use but they've got to be somewhere). That's less of a deal in the modern game where there are regulations all teams have to adhere to and a general sense of professionalism with the grounds that keep everything at a more or less even keel, but before I think 1969 teams still had distracting advertisements in the hitter's eye and I think you have to go well into the 70s to find fields that were uniformly manicured and so on. And Seamheads isn't opening for me right now but I have a general sense that gustiness in Candlestick would have made walks more of a thing there... not to mention pitching around players more having (probably) a heightened impact on walks in Coors.

[HonusWagner]

Yes exactly, the old parks have so much originality and character and it adds to the gameplay if these are reflected within reasonable accuracy by the game. But as of now am not that confident. Like so much that happened with this game the features remained undeveloped to their potential. My Phillies in Baker don't make much sense to me. I don't see why LF eats would-be base hits any more than any other field. I don't see how RF is not producing the wall-bouncers it apparently isn't.

So the park factors model seems to be this: using HR for example, the engine is set to generate x HR for the season. That total is divided over the number of parks. Park factors dictate where the HR will be generated. Therefore, the factors are zero-sum: a HR generated in a park means it cannot be generated in another park. And I think that's about it for the park factor model. And so increasing the HR factor in one park must decrease HR produced in another park. I mean, for a stats-generator it is a simple,, elegant and effective model. The downside is that it effectively completely eliminates the ballpark as a factor, which makes it also ironic.

I just can't get into fictional that much. Though it would seem that 1.000 would really be the way to go with any sort of historical because the park factors are largely fictional anyway.

[Brad K]

Originally Posted by Syd Thrift

As a side note, too, I would really like to add park factors for walks and strikeouts and maybe even errors (which I'm not sure are included in the Seamheads DB

The idea of walks and strikeouts changing due to the park is decent evidence all the park factors data is suspect. I think that perhaps even with three year averaging it's a situation where the noise still out powers the signal. After all, you can go through a multi year period where no parks changed yet park factors do.

[HonusWagner]

If all PFs were set to 2.000 for HR would HR output double? If I understand this correctly it wouldn't since the factors designate relative output. Player HR output for the season would be near RL.

[MathBandit]

Originally Posted by HonusWagner

If all PFs were set to 2.000 for HR would HR output double? If I understand this correctly it wouldn't since the factors designate relative output. Player HR output for the season would be near RL.

Yes, they would double if you told the game to double everyone's HRs.

Well, you're telling them to more than double half the players' HRs and increase the other half by a little bit- but the net effect is doubling the overall HRs.

[HonusWagner]

Originally Posted by MathBandit

Yes, they would double if you told the game to double everyone's HRs.

My understanding is Park factors only work to redistribute outputs. If a league is slated to generate 240 HR in the season across 8 parks then each park will generate around 30 HR. Otherwise, the doubling of HR as you claim would skew many of the other hitting performances also. It would be a mess.

Park factors exist to redistribute statistical outputs across the various parks, that's it.

[MathBandit]

Then your understanding seems to be incorrect. This is what happened when I booted up a standard 2023 game,set all parks to have 2.000 HR factor, and simulated the 2023 season.

[Brad K]

Interesting. That means when the game simulates the season to adjust the modifiers to the talent in the league it doesn't consider park factors. I wonder if there's anything else the sim doesn't consider. Maybe manager strategy?

[MathBandit]

It would be weird if it did take park factors into account. If I take the current MLB teams/players and have them play in parks where its 275 down the line and 300 feet to straightaway centre, I'd be pretty upset if HR numbers didn't skyrocket.

[jcard]

Originally Posted by MathBandit

Then your understanding seems to be incorrect. This is what happened when I booted up a standard 2023 game,set all parks to have 2.000 HR factor, and simulated the 2023 season.

Odd that some leader totals were not changed despite having assigned ballpark factors (doubles, triples, etc.). Perhaps in extreme contexts such as this there is a “crowding out” effect on some type of hits (doubles, triples) by others (singles, home runs).

Did you happen to note league totals as well as leaders?

[itsmb8]

My understanding is the modifiers set the desired statistical output, ratings adjust the distribution, and park factors adjust stat output after the fact.