OOTP25 Historical Features

[Brad K]

Hopefully we'll know someday.

[Reed]

Originally Posted by LansdowneSt

Ha! I was just waiting for Garlon to speak on the neutralized stats he does. Not sure anyone else felt the need to insert themselves to speak for him! I’ve seen the regression on real and Maris might hit better if neut stats are used threads. None of what you said was criticism per se. I just know I ran test after test and it uses real stats very well too. That’s how I did my test runs personally.

Did you use park factors or neutralized park factors in your test?
In the past some outliers (Maris, Ruth, etc.) were toned down but with the new rating system they aren't being toned down?

IMO
If you want player to try to duplicate their IRL stats regardless what team they are playing on in OOTP, use real stats/neutral park factors.
If you want to see how Maris would do in a different park, Neutralized stats with park factors.
if you want to see how a player would be without park factors influencing his real life stats OR OOTP stats, then neutralized stats and neutralized park factors.
And finally you can let them play on their real team and use neutralized stats with real park factors and the results should be close to IRL stats.
I never use real stats with park factors but that is just me.

Yes the manual should be updated.

[LansdowneSt]

I test in Replay mode, turn coaches and development lab off, and the rest defaults. So, that’s real stats. That’s for testing. I actually play all sorts of ways.

[Brad K]

To me neutralized stats with park factors are the way to go. I expect and want some variation but that puts the target of the ratings on what they did real life if they play for the same teams and what they would have done playing for different teams.

Now if I could just get an answer to whether Real Stats are neutralized against the game's park factors then I'd be comfortable with my database choice.

[Reed]

Originally Posted by Brad K

To me neutralized stats with park factors are the way to go. I expect and want some variation but that puts the target of the ratings on what they did real life if they play for the same teams and what they would have done playing for different teams.

Now if I could just get an answer to whether Real Stats are neutralized against the game's park factors then I'd be comfortable with my database choice.

Agree with this.

[Jeff1787]

I am doing a random debut (RD) league, and there isn't a lot of info about which way to run the league. I am still doing testing with neutralized vs real stats and using real ballpark factors. I am also trying 5 yr vs 3 yr recalc.

It seems that double weight of current years stats shouldn't have any impact in RD.

Hopefully we will soon get a definitive answer in the adjust hitter and pitcher numbers.

[Reed]

Originally Posted by Jeff1787

I am doing a random debut (RD) league, and there isn't a lot of info about which way to run the league. I am still doing testing with neutralized vs real stats and using real ballpark factors. I am also trying 5 yr vs 3 yr recalc.

It seems that double weight of current years stats shouldn't have any impact in RD.

Hopefully we will soon get a definitive answer in the adjust hitter and pitcher numbers.

It is definitely good to test out a league before starting it in earnest.
For my random game 3 year double weight recalc, I am currently using 600/300 for adjust/weaken for batters and 80/40 for pitchers. 5 year recalc certainly has advantages, especially for career totals.
The way I test, create my league up to starting point. Backup League. Sim out 25 years or more. Check some of the better players to see if BA and HRs seem reasonable. Check better pitchers KO and walk rates. If they look good then just restore the back up and start your league. If I the stats don’t look right, restore the backup, modify the adjust/weaken settings and sim out 25 years again and check the stats.


Do you use one particular year for your stat modifiers or just use current year modifiers?
Anyway let us know what you come up with and your ideal setup.

[Jeff1787]

I use current year modifiers.

Is there a fast way to sim 25 years?

[LansdowneSt]

Go to managers office, turn on do not disturb. Then go to play until… specified date. In Game Settings, upper right, turn off things you don’t need… WPA graphs, replays etc.

[Brad K]

Originally Posted by Jeff1787

I am doing a random debut (RD) league, and there isn't a lot of info about which way to run the league. I am still doing testing with neutralized vs real stats and using real ballpark factors. I am also trying 5 yr vs 3 yr recalc.

It seems that double weight of current years stats shouldn't have any impact in RD.

Hopefully we will soon get a definitive answer in the adjust hitter and pitcher numbers.

It takes a lot of tests to discern reality amid the noise. Lots of tests. And yet whoever wrote the code has the answer. This is beyond the fog of war necessary for a good game.

[Brad K]

Originally Posted by LansdowneSt

Ha! I was just waiting for Garlon to speak on the neutralized stats he does.

Tap tap. Is this thing on?

[32koufax]

Originally Posted by Reed

Agree with this.

1961 Replay using Neutralized Stats and Park Factors... Hitting Stats! Very pleased with overall numbers for both teams and individual players.

[32koufax]

[Reed]

For my league that I recently started I am using neutralized stats and park factors. Until told differently I think that is the safest way to go to see how players would have performed with various teams or their real life team.
I think at 1 time Markus did some adjusting to real stats accounting for home field and did some adjusting to players with outlier HRs. I don’t know if that is the case anymore.

PS. I AM NOT RESPONSIBLE IF YOU CHOOSE TO PLAY THIS WAY!!! (I think I’m going to add this to my signature)

[Brad K]

Originally Posted by 32koufax

1961 Replay using Neutralized Stats and Park Factors... Hitting Stats! Very pleased with overall numbers for both teams and individual players.

Do you mean every park factor at 1.000?

[Reed]

Originally Posted by Brad K

Do you mean every park factor at 1.000?

Sorry should have been clearer. I am NOT using neutral (1.000) park factor’s for my current league.

Usually I do play with neutralized player stats and neutralized (1.000) park factors. I think that gives me a better indication of how players compare.

[32koufax]

Originally Posted by Brad K

Do you mean every park factor at 1.000?

I did not have park factors at 1.000.

[32koufax]

Home run totals were more realistic. No one put up crazy numbers. Orioles Gentile did hit 62, as did Maris, and Colavito hit 60. Mick hit 47. One thing I noticed is that Maris hit well over .300 in several replays using real stats. In 3 replays using neutralized stats, his B.Ave was in line w/ his real life average.

[thehef]

This is some really good stuff!

Originally Posted by Reed

IMO
If you want player to try to duplicate their IRL stats regardless what team they are playing on in OOTP, use real stats/neutral park factors.
If you want to see how Maris would do in a different park, Neutralized stats with park factors.
if you want to see how a player would be without park factors influencing his real life stats OR OOTP stats, then neutralized stats and neutralized park factors.
And finally you can let them play on their real team and use neutralized stats with real park factors and the results should be close to IRL stats.
I never use real stats with park factors but that is just me.

Some questions, though:

Originally Posted by Reed

If you want to see how Maris would do in a different park, Neutralized stats with park factors.

Which scenario would this be best for? I'm assuming for where you don't have players on their historical teams, but not sure. If so, what happens where Maris or Mantle just happens to be end up on the Yankees?

Originally Posted by Reed

if you want to see how a player would be without park factors influencing his real life stats OR OOTP stats, then neutralized stats and neutralized park factors.

How would this one play out, a) players on their historically correct teams, and b) players not on their historical teams?

The other two scenarios makes good sense to me

Thanks!

[Reed]

Originally Posted by thehef

This is some really good stuff!



Some questions, though:


Which scenario would this be best for? I'm assuming for where you don't have players on their historical teams, but not sure. If so, what happens where Maris or Mantle just happens to be end up on the Yankees?



How would this one play out, a) players on their historically correct teams, and b) players not on their historical teams?

The other two scenarios makes good sense to me

Thanks!

Assuming you are using 1 year recalc., as played lineups, etc..
Using neutralized stats with real park factors, players on their real teams, players should end up close to their real life stats (allowing for randomness of course).
Using neutralized stats, real park factors, players not on historical team, players resulting stats will reflect what their stats would have been if they had played on this team in real life instead of their historical team.

I haven't tried this but it would be easy and interesting to start a game in 1961, 1 year recalc.. Neutralized stats. real park factors. Go into Maris's player editor and look at expected results. Then change park factors to real and see if it changes. Change his team in editor and see if his expected results change. You might even be able to change neutralized stats to real stats and do the same. The various results would be interesting IMO
OK, I did try the expected out and kept getting about 41-43 HRs for Maris. Then played a season out using as played line ups, 1 yr recalc. Neutralized stats, real ball park factors. In my extremely low sample size. Maris had 61 HRs and Mantle had 62. I haven’t figured out why the expected was so low.