Accuracy in Historical Best-Of League

Max_Fischer · 12-22-2017, 07:44 PM

Hi everyone! So I'll try to make a long story short here so this isn't a super long post. Basically, I just bought the game a week or so ago, and I am trying to put together something I've always wanted to do: to simulate a league (or really, leagues) consisting of the greatest teams of all-time. What I have done is come up with a long list of the best teams, mostly based on win %, but also with all WS winners thrown in. Then I want to group them into leagues of 8, simulate the league using OOTP, have the top 1 or 2 teams advance, and then continue on like that. I want the seasons to be standard length (150 to 160 games) to give the true best teams a chance to rise to the top, rather than just doing short best-of-7 series or whatever. The thing is, I'm getting some weird results and I'm wondering if it's because I'm doing something wrong.

Just so you know exactly what I did, I went to New Custom Game --> Create Fictional League (1 sub league, 1 division, 8 total teams) --> then Switch to historical team selection on Step 2. Then as a test I picked a team near the very top of the list (1931 A's, went 107-45, win % of .704, 7th best since 1901), and then filled out the league with the bottom 7 teams in my list, based on win %. These are mostly teams who are on the list due to having won the WS, despite not being that great (think 2006 Cards). All of them have a win % on the order of .550 or lower. Needless to say, they are pretty much objectively worse than the '31 A's, with records anywhere from 22 to 29 games worse.

Just to be complete and tell you everything I did, in Step 2 I also unchecked the Minor Leagues, then on Step 3 I selected 2017 settings. I unchecked "Automatically adjust league strategy when advancing to new year" and "Hold Inaugural Fantasy Draft". Went with a 168 game season and series length of 4, so everyone would play each other 24 times (12 home/12 away). In Step 4 I set Scouting Accuracy to 100% Accurate, and deselected "Enable automatic evolution of the league", left the rest at default. In Step 5 I chose GM and Manager, selected Play in Commissioner Mode, Cannot be fired, and Start unemployed, and then started the game. Then I clicked Auto-Play Until Next Year.

So what happened? The 1931 A's finished last. They finished DEAD last, 40 games behind the league winner. I thought maybe it had to do with the fact that I selected 2017 Settings, so I ran it again with 1931 Settings, but with basically the same result (last place A's, thought not quite 40 games back).

I figure there are 3 possibilities:

1) Maybe it's reasonable to believe that the 1931 A's would finish last in such a league, and my preconceived notions are wrong.

2) There is some setting in OOTP that I am not considering that would make for a more realistic sim (I'm thinking this is likely since I am a newb).

3) Even though OOTP lets you set up such leagues, maybe it is not really designed for this sort of cross-era simulation and so can't be expected to provide accurate results?

If you bothered to read all this then I tremendously thank you. Any thoughts?

Hrycaj · 12-22-2017, 08:10 PM

I'm not an ootp expert by any means but I have done basically what you have done more than once. Here are a few things to consider.

1. You only ran it twice and that is a low sample size considering.
2. Did injuries affect the course of events? They can play a major role.
3. 2017 settings are not ideal for a team from 1931. The A's of that year did not have a bullpen like teams today have. However, based on your settings I believe the AI will treat the team like it is from 2017. I also believe that the ratings are normalized for the setting you chose. So you may not have the guys that can go full games like they would have Back then. Or it may be a combination of both factors. Like I said I'm not an expert but those are a few things that I thought of that may have contributed to your results.

Spritze · 12-22-2017, 11:16 PM

A teams wins don't matter in OOTP but the runs it scored and gave up in whatever the run environment was do.

Neutralized stats were created for the task you are attempting as every team is put on a level playing field, a 750 run environment 1871-2016. Batting, pitching and fielding stats are all adjusted.

TheMaus2 · 12-23-2017, 01:00 AM

Perhaps pythagorean W-L might be a better metric?

Max_Fischer · 12-23-2017, 11:30 AM

Quote:

Originally Posted by Spritze

Neutralized stats were created for the task you are attempting

I'm not sure what this is. Are neutralized stats something I can enable in the game?

Spritze · 12-23-2017, 02:47 PM

Quote:

Originally Posted by Max_Fischer

I'm not sure what this is. Are neutralized stats something I can enable in the game?

Yes, when you create your game you can choose neutralized or real stats.

NoOne · 12-23-2017, 02:56 PM

a 1901 team in 1931 isn't the same beast, essentially. ratings for players are all dialed in relative to statistical environments defined by the totals/modifiers etc.

so, you took players that 'should' do various things in 1901, changed the league totals and modifiers and their competition. in this particular case with the totals/modifiers used, 1901 teams are inferior to 1931 teams (*not always the case)

neutralized stats is an option. it's during game creation and also in game -- historical tab or wherever in setting, check manual for location. (tab left of 'stats and ai') it's better than nothing.

one thing to note, if you fill a league with all amazing teams, someone has to be the heel. after anythign you do, if they are all ~equivalent teams in quality, it will be quite random as to whom is successful vs unsuccessful.

and no matter what, if you mix teams from different eras and neutralize etc etc, performing as they did originally did isn't likely. new competition, new environment, a bunch of processing of ratings... not saying it cannot be fun or that it may work better than what i am leading you to believe, merely tempering expectations to be very open-minded.

you can work with resulting totals/modifiers to tweak thigns, if needed.

Max_Fischer · 12-26-2017, 11:35 AM

Thank for the replies everyone. The suggestion on neutralized stats seem to have helped. But I agree that there's no way to get all these teams to perform exactly as they did in their own era when mixing them together like this.

Thanks again!

Brad K · 01-03-2018, 12:13 PM

My theory is the poor results of the As is due to a large number of stronger teams being relatively close in overall talent with the As alone significantly below that.

That sort of talent gap often occurs when baseball expands. Evidence of such a gap is when more teams are over 500 than below and when the last place team loses more games than the first place team wins.

Skrabec1 · 01-04-2018, 06:58 PM

I thought there was some issue with the neutralized stats due to a field change in OOTP 18?

Spritze · 01-04-2018, 10:12 PM

Quote:

Originally Posted by Skrabec1

I thought there was some issue with the neutralized stats due to a field change in OOTP 18?

Currently there are no reported issues.

12-22-2017, 07:44 PM	#1
Max_Fischer Bat Boy Join Date: Dec 2017 Posts: 8	Accuracy in Historical Best-Of League Hi everyone! So I'll try to make a long story short here so this isn't a super long post. Basically, I just bought the game a week or so ago, and I am trying to put together something I've always wanted to do: to simulate a league (or really, leagues) consisting of the greatest teams of all-time. What I have done is come up with a long list of the best teams, mostly based on win %, but also with all WS winners thrown in. Then I want to group them into leagues of 8, simulate the league using OOTP, have the top 1 or 2 teams advance, and then continue on like that. I want the seasons to be standard length (150 to 160 games) to give the true best teams a chance to rise to the top, rather than just doing short best-of-7 series or whatever. The thing is, I'm getting some weird results and I'm wondering if it's because I'm doing something wrong. Just so you know exactly what I did, I went to New Custom Game --> Create Fictional League (1 sub league, 1 division, 8 total teams) --> then Switch to historical team selection on Step 2. Then as a test I picked a team near the very top of the list (1931 A's, went 107-45, win % of .704, 7th best since 1901), and then filled out the league with the bottom 7 teams in my list, based on win %. These are mostly teams who are on the list due to having won the WS, despite not being that great (think 2006 Cards). All of them have a win % on the order of .550 or lower. Needless to say, they are pretty much objectively worse than the '31 A's, with records anywhere from 22 to 29 games worse. Just to be complete and tell you everything I did, in Step 2 I also unchecked the Minor Leagues, then on Step 3 I selected 2017 settings. I unchecked "Automatically adjust league strategy when advancing to new year" and "Hold Inaugural Fantasy Draft". Went with a 168 game season and series length of 4, so everyone would play each other 24 times (12 home/12 away). In Step 4 I set Scouting Accuracy to 100% Accurate, and deselected "Enable automatic evolution of the league", left the rest at default. In Step 5 I chose GM and Manager, selected Play in Commissioner Mode, Cannot be fired, and Start unemployed, and then started the game. Then I clicked Auto-Play Until Next Year. So what happened? The 1931 A's finished last. They finished DEAD last, 40 games behind the league winner. I thought maybe it had to do with the fact that I selected 2017 Settings, so I ran it again with 1931 Settings, but with basically the same result (last place A's, thought not quite 40 games back). I figure there are 3 possibilities: 1) Maybe it's reasonable to believe that the 1931 A's would finish last in such a league, and my preconceived notions are wrong. 2) There is some setting in OOTP that I am not considering that would make for a more realistic sim (I'm thinking this is likely since I am a newb). 3) Even though OOTP lets you set up such leagues, maybe it is not really designed for this sort of cross-era simulation and so can't be expected to provide accurate results? If you bothered to read all this then I tremendously thank you. Any thoughts? Last edited by Max_Fischer; 12-22-2017 at 07:46 PM.

12-22-2017, 08:10 PM	#2
Hrycaj All Star Starter Join Date: Oct 2010 Posts: 1,799	I'm not an ootp expert by any means but I have done basically what you have done more than once. Here are a few things to consider. 1. You only ran it twice and that is a low sample size considering. 2. Did injuries affect the course of events? They can play a major role. 3. 2017 settings are not ideal for a team from 1931. The A's of that year did not have a bullpen like teams today have. However, based on your settings I believe the AI will treat the team like it is from 2017. I also believe that the ratings are normalized for the setting you chose. So you may not have the guys that can go full games like they would have Back then. Or it may be a combination of both factors. Like I said I'm not an expert but those are a few things that I thought of that may have contributed to your results. __________________ Click on my signature to read about the great game of baseball in Normington.

12-22-2017, 11:16 PM	#3
Spritze OOTP Historical Czar Join Date: Dec 2001 Location: Bothell Wa Posts: 7,253	A teams wins don't matter in OOTP but the runs it scored and gave up in whatever the run environment was do. Neutralized stats were created for the task you are attempting as every team is put on a level playing field, a 750 run environment 1871-2016. Batting, pitching and fielding stats are all adjusted. __________________ It's madness, madness, I tell you! For the love of God, don't do it!

12-23-2017, 01:00 AM	#4
TheMaus2 All Star Starter Join Date: Feb 2014 Location: The OOTP Forums. Always. Posts: 1,951	Perhaps pythagorean W-L might be a better metric? __________________ I write a monthly newsletter on the Food Baseball Association. I also listen to music no one's ever heard of in hopes of looking cool and alternative.

12-23-2017, 02:56 PM	#7
NoOne Banned Join Date: Apr 2015 Posts: 7,273 Infractions: 0/1 (3)	a 1901 team in 1931 isn't the same beast, essentially. ratings for players are all dialed in relative to statistical environments defined by the totals/modifiers etc. so, you took players that 'should' do various things in 1901, changed the league totals and modifiers and their competition. in this particular case with the totals/modifiers used, 1901 teams are inferior to 1931 teams (*not always the case) neutralized stats is an option. it's during game creation and also in game -- historical tab or wherever in setting, check manual for location. (tab left of 'stats and ai') it's better than nothing. one thing to note, if you fill a league with all amazing teams, someone has to be the heel. after anythign you do, if they are all ~equivalent teams in quality, it will be quite random as to whom is successful vs unsuccessful. and no matter what, if you mix teams from different eras and neutralize etc etc, performing as they did originally did isn't likely. new competition, new environment, a bunch of processing of ratings... not saying it cannot be fun or that it may work better than what i am leading you to believe, merely tempering expectations to be very open-minded. you can work with resulting totals/modifiers to tweak thigns, if needed.

12-26-2017, 11:35 AM	#8
Max_Fischer Bat Boy Join Date: Dec 2017 Posts: 8	Thank for the replies everyone. The suggestion on neutralized stats seem to have helped. But I agree that there's no way to get all these teams to perform exactly as they did in their own era when mixing them together like this. Thanks again!

01-03-2018, 12:13 PM	#9
Brad K Banned Join Date: May 2016 Location: St Petersburg Florida USA Posts: 6,693 Infractions: 0/2 (4)	My theory is the poor results of the As is due to a large number of stronger teams being relatively close in overall talent with the As alone significantly below that. That sort of talent gap often occurs when baseball expands. Evidence of such a gap is when more teams are over 500 than below and when the last place team loses more games than the first place team wins.

01-04-2018, 06:58 PM	#10
Skrabec1 Minors (Single A) Join Date: Apr 2010 Posts: 74	I thought there was some issue with the neutralized stats due to a field change in OOTP 18?