Without Accurate Sim Stats, Classic Rosters are not Classic ....

[squashbuggie]

First off - SCEA needs to let us globally edit ratings to speed things up. I think that JTommy and his calculator had the 1975 roster w/ best sim stats, until I perfected the 94/07 sim stats. When I say accurate, I mean 95%-98% exact sim stats to the ERA of ball for those three Classic Rosters.

The most important sim stats to perfect are:

*AL/NL Team ERA's

*K/9

*BB/9

*FLD. PCT.

If you recreate a roster, that is a lotta work. However, if your roster has stats when simming that are only 75% accurate, how worthy is that roster to be called Classic.

[Caulfield]

Re: Without Accurate Sim Stats, Classic Rosters are not Classic ....

I do wish service time stats and coaches names were editable but I still thoroughy enjoy reliving the Eighties. They are definitely ''Classic'' to me. but I have no problem referring to them as retro, historic or vintage either. I'm not one to quibble semantics. (Okay, I am but thats for another thread on another day )

[squashbuggie]

Re: Without Accurate Sim Stats, Classic Rosters are not Classic ....

I do wish service time stats and coaches names were editable but I still thoroughy enjoy reliving the Eighties. They are definitely ''Classic'' to me. but I have no problem referring to them as retro, historic or vintage either. I'm not one to quibble semantics. (Okay, I am but thats for another thread on another day )

Are you one for accurate sim stats - or do you prefer names that over-perform ?

[Madden08PCgmr]

Re: Without Accurate Sim Stats, Classic Rosters are not Classic ....

I respect that opinion. But after years of Madden-ing, I'll say this.

I despise guys that say "who's going to make such and such?" or request 'whatever'!

I prefer the legwork get done. If I want insane/perfect accuracy, I'll make it happen.

Even if a roster is less than perfect, at least the lions-share of the work is done rather than building everything from scratch.


Sent from my iPhone using Operation Sports

[Caulfield]

Re: Without Accurate Sim Stats, Classic Rosters are not Classic ....

Are you one for accurate sim stats - or do you prefer names that over-perform ?

I prefer accurate stats but sometimes a Brady Anderson will throw a wrench in the works.
which is fine, I can live with that so long as there is only one Brady a year.

[squashbuggie]

Re: Without Accurate Sim Stats, Classic Rosters are not Classic ....

I respect that opinion. But after years of Madden-ing, I'll say this.

I despise guys that say "who's going to make such and such?" or request 'whatever'!

I prefer the legwork get done. If I want insane/perfect accuracy, I'll make it happen.

Even if a roster is less than perfect, at least the lions-share of the work is done rather than building everything from scratch.


Sent from my iPhone using Operation Sports

I agree that people that expect rosters are not good. It is a lotta work and why do all that work and have inaccurate sim stats. Hope ur roster making is going well and thanks for ur contributions and those of ALL the roster makers.

[Madden08PCgmr]

Re: Without Accurate Sim Stats, Classic Rosters are not Classic ....

I agree that people that expect rosters are not good. It is a lotta work and why do all that work and have inaccurate sim stats. Hope ur roster making is going well and thanks for ur contributions and those of ALL the roster makers.

Thank you.

Pretty slow without an external editor, but I'm over the halfway point now.

Curious how it will all turn out, being a first-timer. I'm sure I'll tweak this a bunch.


Sent from my iPhone using Operation Sports

[squashbuggie]

Re: Without Accurate Sim Stats, Classic Rosters are not Classic ....

Thank you.

Pretty slow without an external editor, but I'm over the halfway point now.

Curious how it will all turn out, being a first-timer. I'm sure I'll tweak this a bunch.


Sent from my iPhone using Operation Sports

what kinda roster are you making ?

[Madden08PCgmr]

Re: Without Accurate Sim Stats, Classic Rosters are not Classic ....

what kinda roster are you making ?

1990 40-man.

I gathered enough info to do full 80-90 man, but my tinkering with 14/15 led me to believe I'd prob play seasons and not franchise, so why go that far?.. Seems to be the general consensus of the board too.

.. I plan to go forward off the build and try to get through the entire decade.

Sorely missing an editor though! It would be great to just blanket certain positions (pitchers) and equipment. I would guess that would shave half my time off. But, it is what it is I suppose.


Sent from my iPhone using Operation Sports

[Caulfield]

Re: Without Accurate Sim Stats, Classic Rosters are not Classic ....

1990 40-man.

I gathered enough info to do full 80-90 man, but my tinkering with 14/15 led me to believe I'd prob play seasons and not franchise, so why go that far?.. Seems to be the general consensus of the board too.

.. I plan to go forward off the build and try to get through the entire decade.

Sorely missing an editor though! It would be great to just blanket certain positions (pitchers) and equipment. I would guess that would shave half my time off. But, it is what it is I suppose.


Sent from my iPhone using Operation Sports

I didnt know if there was a 1990 thread, so I guess this will suffice for the time being:
40 man is excellent, with a smattering of free agents too, if its not asking too much.
preferably these-

Darrell Evans / Buddy Bell / Chris Speier / Jim Sundberg / Manny Trillo / Tony Armas / Dwayne Murphy / Keith Moreland / Glenn Hubbard / Jim Rice / Mike Schmidt
Tommy John / Doyle Alexander / Bob Forsch / Rick Rhoden / Mike Krukow / Shane Rawley / Bob Stanley / Steve Trout / Kent Tekulve

[Madden08PCgmr]

Re: Without Accurate Sim Stats, Classic Rosters are not Classic ....

Thanks for that!

I only have about a dozen FAs on the list so far.

I've been filling the expansion teams with minor leaguers who weren't active in the majors after '89. (another time consuming aspect)


Sent from my iPhone using Operation Sports

[JTommy67]

Re: Without Accurate Sim Stats, Classic Rosters are not Classic ....

It's important to remember that, whatever our ability to figure out how the sim engine works and use that to our advantage when creating historical rosters, there are still some limiting factors.

With respect to fielding percentages, the fielding ability attribute controls more than just error tendency. It also affects the athleticism of the player in the field and constrains our ability to accurately reproduce acrobatic players who might have been prone to more errors. Secondly, the default rating scales do not produce enough defensive errors even for the default SCEA rosters. Beyond that, error distributions among the positions are wrong as are throwing errors vs. fielding errors (which can be discerned by holding accuracy or fielding at zero and checking errors). This means that to reproduce these positional error distributions, some positions (on average) would have to be rated WAY lower than others, so much so that you'd quickly find the cpu managers favoring backup second-basemen over the starting shortstop, for example.

Additionally, fly ball to ground ball ratios are not always accurate depending on your roster year, and assists for outfielders are out of the normal range in the sim engine. This further complicates things with respect to defense.

Another important note on defense: baseball historians have concluded (rightly so, I think), that historical fielding percentages were much lower because official scorers were a lot tougher than they are today. For example, in the 1900s and 10s it was common to charge an error on any ball that was not successfully converted into an out. This meant that if a speedster hit a grounder to shortstop who fielded the ball cleanly and fired to first but the runner beat it out, then the shortstop could be charged with an error. Since that time, scorers have slowly and steadily become more lenient.

It means that if Eddie Collins or even Alan Trammel played under today's scoring conventions, they'd likely have a much higher fielding percentages. And, of course, MLB The Show uses modern scoring methods. After realizing this in full, I changed my calculator to rescale defensive ratings to modern ranges. Another reason is that, going back in time, defensive ratings have to get really, really, low to reproduce historical fielding percentages, and you eventually reach a point where zero isn't even low enough.

There are also numerous relationships in the sim engine that might go unnoticed. For example, lowering fielding ratings increases run production, which means more runs scored and RBIs. It also affects innings pitched by starters. (More unearned runs = more pitches thrown)

There are other statistics that are not reproduced accurately in the default rosters, such as GIDP, sac bunts, intentional walks, etc...all of which have their own particular effect in the sim engine.

Another problem is that 99 stealing aggression does not reproduce the frequency of steal attempts historically seen by some of the great basestealers of the game's speedster eras (early 1900s and 1980s).

All this means that in order to reproduce historical statistics as closely as possible, you have to begin to deviate extensively from standard ranges. It's not an easy affair, I know that many on this forum in the past including myself have studied this rather extensively.

All things considered, I still think their sim engine is the best in a sports game and affords us quite a bit of opportunity. Until we get a quicker way to edit rosters, however, I think the roster makers are doing as good a job as possible.

[squashbuggie]

Re: Without Accurate Sim Stats, Classic Rosters are not Classic ....

It's important to remember that, whatever our ability to figure out how the sim engine works and use that to our advantage when creating historical rosters, there are still some limiting factors.

With respect to fielding percentages, the fielding ability attribute controls more than just error tendency. It also affects the athleticism of the player in the field and constrains our ability to accurately reproduce acrobatic players who might have been prone to more errors. Secondly, the default rating scales do not produce enough defensive errors even for the default SCEA rosters. Beyond that, error distributions among the positions are wrong as are throwing errors vs. fielding errors (which can be discerned by holding accuracy or fielding at zero and checking errors). This means that to reproduce these positional error distributions, some positions (on average) would have to be rated WAY lower than others, so much so that you'd quickly find the cpu managers favoring backup second-basemen over the starting shortstop, for example.

Additionally, fly ball to ground ball ratios are not always accurate depending on your roster year, and assists for outfielders are out of the normal range in the sim engine. This further complicates things with respect to defense.

Another important note on defense: baseball historians have concluded (rightly so, I think), that historical fielding percentages were much lower because official scorers were a lot tougher than they are today. For example, in the 1900s and 10s it was common to charge an error on any ball that was not successfully converted into an out. This meant that if a speedster hit a grounder to shortstop who fielded the ball cleanly and fired to first but the runner beat it out, then the shortstop could be charged with an error. Since that time, scorers have slowly and steadily become more lenient.

It means that if Eddie Collins or even Alan Trammel played under today's scoring conventions, they'd likely have a much higher fielding percentages. And, of course, MLB The Show uses modern scoring methods. After realizing this in full, I changed my calculator to rescale defensive ratings to modern ranges. Another reason is that, going back in time, defensive ratings have to get really, really, low to reproduce historical fielding percentages, and you eventually reach a point where zero isn't even low enough.

There are also numerous relationships in the sim engine that might go unnoticed. For example, lowering fielding ratings increases run production, which means more runs scored and RBIs. It also affects innings pitched by starters. (More unearned runs = more pitches thrown)

There are other statistics that are not reproduced accurately in the default rosters, such as GIDP, sac bunts, intentional walks, etc...all of which have their own particular effect in the sim engine.

Another problem is that 99 stealing aggression does not reproduce the frequency of steal attempts historically seen by some of the great basestealers of the game's speedster eras (early 1900s and 1980s).

All this means that in order to reproduce historical statistics as closely as possible, you have to begin to deviate extensively from standard ranges. It's not an easy affair, I know that many on this forum in the past including myself have studied this rather extensively.

All things considered, I still think their sim engine is the best in a sports game and affords us quite a bit of opportunity. Until we get a quicker way to edit rosters, however, I think the roster makers are doing as good a job as possible.

very good response and thank you for ur expertise. It took many hours to perfect the 94/07 set. However, to see the results where League ERA's - K/9 - BB/9 - and SB totals and % are almost exact to that era .... Man - It is exciting.