Historical Player Ratings

[jberg57]

Has anyone played with these settings? What works for you?

[pstrickert]

It kind of depends. Tell us more about your plans for your league.

[quillenl]

OOTP has a stamina problem... so I would never recommend single season for stamina calc. I use 3 year recalc,, or career depending on what I am going for. Of all those settings, only stamina has a rule with me... which is "don't use single season".

The pool of starter capable players, in my experience, will be too small if you allow injuries. The number of guys who pitched just a couple of games, got shelled, then went back to the bullpen (or minors) accounts for quite a few starts... and most of these guys will have too low a stamina to start using single season recalc,

Mileage may vary, and if you have been using Current Season only for some amount of time I'd be interested in knowing if your experience has been different. Be nice if this is something that has changed in recent versions that I never knew due to being locked in an old way of doing something out of habit.

edit: For everything else... it def depends on what I am looking to get out of the league. I alternate between single season and 3 year recalc in general quite a bit.

[fredbeene]

you mean the touted HELP and YouTube video still doesn't cover this topic??

[The Game]

Originally Posted by fredbeene

you mean the touted HELP and YouTube video still doesn't cover this topic??

Thats what the boards are for. Chill.


Personally in historical I use 1 yr recalc. I hated several versions ago when Smoltz was a CL for 5 years. Eckersly is left as a starter longer then he was IRL and in 1 sim he retired prior to 1990. I have not noticed a stamina problem except in modern times with it set to Low. I use normal in most leagues, High pre 1980.

In many leagues I use entire career for imports, when its not a yearly historical. But I like having Tim Raines be able to play some 2b and Mantle as a bad SS.

[quillenl]

Originally Posted by The Game

I have not noticed a stamina problem except in modern times with it set to Low. I use normal in most leagues, High pre 1980.

I should have been more specific... I meant stamina rating for players in how it affects their ability to be in the starting rotation. For example, on a 1-10 scale I would be looking for 4 stamina and 3 pitches as a minimum. Generally OOTP doesn't generate enough players who meet this criteria compared to any given historical season.

Regardless, I went and did some testing and it is not nearly the issue I recall it being. Some changes to stamina calc between OOTP17 and 23 seem to be resulting in slightly more players capable of starting... but it the total number of players with stam >3 was virtually the same regardless of period used for calculating stamina so I withdraw my opinion on avoiding single season recalc for stamina.

Issue as I remember it:
Using 1979 as an example, you had 211 players start at least 3 games. I would expect all those players to have at least a 4 stamina so they could be expected to pitch at least 5 innings. Using Talent Development Engine you would only see about 174 players generated with 3+ stamina for 1979, with 184 generated when using Recalc based on Real Stats (Don't know why it is different, but it is).

You end up having a smaller pool of players getting starts on sims than what actually occurred in MLB for any particular season due to fewer options for what the AI considers a reasonable starter. My test sim saw 162 players get at least 3 starts, 75% of the number of players expected. In my sim, for example, Dave LaRoche ended up with 9 starts watching his ERA balloon from 1.8 to 6.2 which would have been avoided with a larger pool of players having stam >3.

Short version based on what I checked last night:
Don't sweat the Stamina Recalc type if your concern is the total number of starters available... the impact is minimal. Choose it for the goal of your league same as you would for the other stat recalc settings.

[Brad K]

Unless it's been changed for 23, "Real Stats" does not work as explained in the manual. I recommended neutralized stats.

[The Game]

Originally Posted by Brad K

Unless it's been changed for 23, "Real Stats" does not work as explained in the manual. I recommended neutralized stats.

Is that only for historical replays or for all historical players including imports? I used neutralized fora long time but I have used real for 3-4 versions and thousands of imports. Most of my imports do well.

[Brad K]

The real stats data base does not work as described in the manual. However you access it, historical replay, random debut, etc, it doesn't work as described in the manual (unless there 's a change not documented in the release notes).

One discrepancy is easy to check. Import Roger Maris 1961 rating on single year and see how many HRs the edit page says he'll hit with the ratings he has. Import Mickey Mantle the same way. If Mantle shows more HRs than Maris then the database is still adjusted to prevent the game from producing real life performances the devs consider to be outliers.

Meaning "Real Stats" still aren't real stats. They're just made with real stats. Like Cheeze Whiz is made with real cheese. Should be named Statz Whiz!!!

[jcard]

Originally Posted by Brad K

The real stats data base does not work as described in the manual. However you access it, historical replay, random debut, etc, it doesn't work as described in the manual (unless there 's a change not documented in the release notes).

One discrepancy is easy to check. Import Roger Maris 1961 rating on single year and see how many HRs the edit page says he'll hit with the ratings he has. Import Mickey Mantle the same way. If Mantle shows more HRs than Maris then the database is still adjusted to prevent the game from producing real life performances the devs consider to be outliers.

Meaning "Real Stats" still aren't real stats. They're just made with real stats. Like Cheeze Whiz is made with real cheese. Should be named Statz Whiz!!!

I will point out that Ratings Based On Real Stats does NOT necessarily mean Ratings Will Produce Real Life Stats. All it says is that Real Stats are the INPUT used to determine the ratings; the target statistical output those ratings produce through the game engine are determined by the goal of the programmer. That objective may or may not be the reproduction of the actual stats.

[LansdowneSt]

Originally Posted by Brad K

The real stats data base does not work as described in the manual. However you access it, historical replay, random debut, etc, it doesn't work as described in the manual (unless there 's a change not documented in the release notes).

One discrepancy is easy to check. Import Roger Maris 1961 rating on single year and see how many HRs the edit page says he'll hit with the ratings he has. Import Mickey Mantle the same way. If Mantle shows more HRs than Maris then the database is still adjusted to prevent the game from producing real life performances the devs consider to be outliers.

Meaning "Real Stats" still aren't real stats. They're just made with real stats. Like Cheeze Whiz is made with real cheese. Should be named Statz Whiz!!!

Part of that is because the Editor shows a modern, neutral environment. Also, 61 HR was an outlier season. If the Editor showed 61 as his baseline and he got an in-game outlier season, he'd be in the upper 70s. That's what I understood that Editor line to reflect - not the literal line that player delivered in that historic year.

[Brad K]

Originally Posted by LansdowneSt

Part of that is because the Editor shows a modern, neutral environment. Also, 61 HR was an outlier season. If the Editor showed 61 as his baseline and he got an in-game outlier season, he'd be in the upper 70s. That's what I understood that Editor line to reflect - not the literal line that player delivered in that historic year.

That's why I brought up the Mantle comparison. I don't recall the exact numbers but last time I checked this Mantle got rated for the low 50s in HRs and Maris got rated for the low 30s. This is not due to modern neutral etc. This is due to Real Stats not being real stats!

Another thing that came out a year or so ago in a discussion Garlon and the devs is that real stats aren't real in another sense. Contrary to what is stated in the manual they are neutralized. The neutralization is done using the game's park factors.

No answer was provided concerning what park factors were used to create neutralized stats. Are the park factors the same? They should be, right? And if so why are there two choices for a rating basis database?

If so then the difference between real and neutralized isn't what would be concluded from the meanings of the words. The difference is that what are considered outlier real life performances on home runs are reduced for rating purposes.

[Brad K]

Originally Posted by jcard

I will point out that Ratings Based On Real Stats does NOT necessarily mean Ratings Will Produce Real Life Stats.

Yes. But that should be due to randomness not design.

[jcard]

Originally Posted by Brad K

Yes. But that should be due to randomness not design.

Again, you are assuming that the proper function of ratings is to make the expected simulation outcome to equal the actual historical record. This is a value. This is a subjective standard- a perfectly reasonable one, to be sure, but a subjective one nonetheless.

More generally, I think that the discussion would be better served by framing the issue differently: it seems to me the actual disagreement with the games design is not with how it uses real stats, but rather how it defines the purpose of recalculation (one-year in particular).

[Hoiles]

Originally Posted by quillenl

OOTP has a stamina problem... so I would never recommend single season for stamina calc. I use 3 year recalc,, or career depending on what I am going for. Of all those settings, only stamina has a rule with me... which is "don't use single season".

I agree with this. For some reason, for starters with a limited amount of innings in a year (say <100 IP), instead of weakening their ratings a bit, they just make their stamina low and they become relievers. So the pool of starters is pretty much only those who started the entire year. There must be some sort of design reasoning behind this.

[Brad K]

Originally Posted by jcard

Again, you are assuming that the proper function of ratings is to make the expected simulation outcome to equal the actual historical record. This is a value. This is a subjective standard- a perfectly reasonable one, to be sure, but a subjective one nonetheless.

When I (and many others) choose "historical game" we have reason to expect historical performance subject to normal randomness and whatever TCR level we have chosen. We can properly complain when something called "Real Stats" doesn't deliver this BY DESIGN.

[Brad K]

Originally Posted by jcard

More generally, I think that the discussion would be better served by framing the issue differently: it seems to me the actual disagreement with the games design is not with how it uses real stats, but rather how it defines the purpose of recalculation (one-year in particular).

To me the issue is Real Stats aren't "real stats". They're like Cheeze Whiz, made with real cheese but not the real thing. Statz Whiz!

[jcard]

Originally Posted by Brad K

When I (and many others) choose "historical game" we have reason to expect historical performance subject to normal randomness and whatever TCR level we have chosen. We can properly complain when something called "Real Stats" doesn't deliver this BY DESIGN.

Two trips on the logical carrousel are enough for me.

[Brad K]

Originally Posted by jcard

Two trips on the logical carrousel are enough for me.

Sorry you're confused by realtiy!!!

[quillenl]

Originally Posted by Brad K

Unless it's been changed for 23, "Real Stats" does not work as explained in the manual. I recommended neutralized stats.

Neither Neutral nor Real Stats work the way it sounds like you desire from the later posts. Using your 1961 Mantle / Maris comparison... Mantle will generate as a significantly better HR hitter relative to Maris when playing Single Season replay for 1961 in Version 23 and 22. See attached preseason predictions using both methods (editor supports view of projected stats supports these predictions).

Neither here nor there except that this is not a reason to suggest Neutral over Real or vice versa.

On a side note, from what I remember, Real Stats were supposed to have been changed a few versions back such that there was no longer a reason to use Neutral Stats to adjust for stadiums they had played at historically. There is a slight different in the editor values for Real vs Neutral so maybe I misunderstood what was changed. I'd fully expected them to be equal when I checked.