Negro League Player Ratings - A Proposed Solution

Charlie Hough · 11-27-2023, 05:51 PM

This is not a bug, but it's an unfortunate consequence of real-life stats and how the adjust/weaken settings work in OOTP. However, we need a fix for this in OOTP 24 or OOTP 25, to ensure proper and fair representation of NeL players in OOTP.

Since Negro League teams played far fewer official games than NL or AL teams, players do not have the same amount of AB or IP. As a result, OOTP's default adjust/weaken settings cause many NeL players to be rated far worse than they should be. A large number of players are adjusted or weakened, so this effectively causes NeL players to have ratings that are closer to AAA, AA or even A level.

It also causes many people to believe that the NeLs are still being treated as minor leagues in OOTP. There have been a number of threads about this over the years, and the general belief expressed in these threads is that OOTP is treating the NeLs unfairly.

Even if you edit your adjust/weaken settings, many NeL players will still be negatively affected.

We need to fix this by having OOTP automatically lower the adjust/weaken settings when they're applied to relevant NeL seasons. It could apply a multiplier to the adjust/weaken AB and IP thresholds, to lower them based on the relative difference in the number of games played between the NeLs and MLB for that season. For example, in many seasons, NeL teams played around 80 games, compared 154 for MLB, so a multiplier of 0.52 could be applied to the AB and IP thresholds in the adjust/weaken settings, and that would hopefully resolve this problem.

There could be other ways to fix this, but I wanted to post this in the bugs forum, to hopefully get some immediate attention to this for OOTP 24 or OOTP 25.

Pelican · 12-06-2023, 12:48 PM

I agree with this suggestion. I have seen the same distorted effects among NeL players who had limited PA or IP, or where stats were incomplete.

I gather, since there is no "Historical" tab for Negro Leagues, there is no way - at present - to set "Adjust/Weaken" for "minor" leagues to different values than for MLB. Ideally, those levels of PA and IP could be reduced for NeL players, under a "Historical" tab, to reflect the shorter seasons, and missing stats in some cases.

Or, what the game engine could do, playing off "Global" MLB settings, is to pro-rate the settings, based on the games for which records are available in a NeL season. In other words, to make it easy, if the NeL Season were effectively 77 games, the "Global" settings for "Adjust/Weaken" would be reduced by half.

So, if IP were set at 100/50 [AB or IP] for "Adjust/Weaken" for MLB, the values for NeL would be 50/25. The variable would change based on the number of games played in the NeL season. At some point, maybe seasons under fifty or forty games, the reduction would need to stop, to avoid bizarre results from very limited stats.

In an ideal world, there would be greater weight given, even to limited stats, for postseason games between NeL and MLB players. This could be done on a career basis, as well. Seamheads breaks out NeL player performance against MLB players, during the annual October barnstorming games that occurred in the 1920's through the 1940's.
Maybe something like "give double [or triple] weight to stats against MLB teams". The performance of the best NeL players in these games is striking. Although a limited sample, it does provide the clearest marker for equalizing NeL stats to MLB play.

thehef · 12-06-2023, 04:00 PM

This is a great suggestion. However, it should apply to all minor leagues as well as Negro Leagues. Many minor leagues throughout history played significantly fewer games than the 154-162 MLB season. Usually this was by design but sometimes it was because a league folded...

Currently OOTP's milbleagues.csv file indicates 345 minor league league-seasons that had 90 or fewer scheduled games. This is not a direct correlation to how many games were actually played, but it's a good indicator. So for most of those season-leagues and depending upon what your adjust/weaken settings are, you're going to have anywhere from most to all of your players (in these leagues) adjusted or weakened. Unless I'm misunderstanding how A/W works... which is a very real possibility

Maybe the solution could be something like this, just using hitter AB's as the example:

- Gamer sets Adjust to 200 AB's and Weaken to 50 AB's
- OOTP applies this to MLB
- OOTP looks at the milbleagues.csv file to get the number of games for that season league and does some simple math to determine what the A/W settings should be for that league. The calc would simply be [milbleague games played] / [mlb games played] * [adjust number]
- OOPT applies the normalized Adj/Wkn numbers to each minor league

See the image below for some examples...

The # of games played in the milbleagues.csv file would need to be reviewed because as best as I can tell those numbers represent either scheduled games, or some combination of scheduled & actual. That number would need to be reflective of actual games played in order for this to work...I'll also note that there are two fields in this csv file that have to do with # of games: GamesScheduled and Season Length. For purposes of this example I'm using the former; I don't know exactly how OOTP uses these fields or if it uses one and not the other... but there are two fields and perhaps one could be used for scheduled games to be played (to correspond to season length/games-to-be-played in OOTP), and the other can be used to indicate the number of games actually played...

(In the example below of the PCL, it played more games than MLB. In my image below, the adjustment increases Adj/Wkn. However, the calculation could include a cap to where the Adj/Wkn settings wouldn't be higher than MLB's... Or not...)

TomVeal · 12-10-2023, 01:07 PM

There's a problem with this corrective measure: The A/W thresholds are supposed to represent a large enough sample size to provide a valid representation of the player's real life ability. That size doesn't change with the length of the season. Stats based on 50 plate appearances aren't indicative of very much.

11-27-2023, 05:51 PM	#1
Charlie Hough Hall Of Famer Join Date: May 2006 Posts: 3,640	Negro League Player Ratings - A Proposed Solution This is not a bug, but it's an unfortunate consequence of real-life stats and how the adjust/weaken settings work in OOTP. However, we need a fix for this in OOTP 24 or OOTP 25, to ensure proper and fair representation of NeL players in OOTP. Since Negro League teams played far fewer official games than NL or AL teams, players do not have the same amount of AB or IP. As a result, OOTP's default adjust/weaken settings cause many NeL players to be rated far worse than they should be. A large number of players are adjusted or weakened, so this effectively causes NeL players to have ratings that are closer to AAA, AA or even A level. It also causes many people to believe that the NeLs are still being treated as minor leagues in OOTP. There have been a number of threads about this over the years, and the general belief expressed in these threads is that OOTP is treating the NeLs unfairly. Even if you edit your adjust/weaken settings, many NeL players will still be negatively affected. We need to fix this by having OOTP automatically lower the adjust/weaken settings when they're applied to relevant NeL seasons. It could apply a multiplier to the adjust/weaken AB and IP thresholds, to lower them based on the relative difference in the number of games played between the NeLs and MLB for that season. For example, in many seasons, NeL teams played around 80 games, compared 154 for MLB, so a multiplier of 0.52 could be applied to the AB and IP thresholds in the adjust/weaken settings, and that would hopefully resolve this problem. There could be other ways to fix this, but I wanted to post this in the bugs forum, to hopefully get some immediate attention to this for OOTP 24 or OOTP 25. Last edited by Charlie Hough; 02-28-2024 at 06:06 PM.

12-06-2023, 12:48 PM	#2
Pelican Hall Of Famer Join Date: Mar 2021 Location: Wilmington, Delaware Posts: 2,934	I agree with this suggestion. I have seen the same distorted effects among NeL players who had limited PA or IP, or where stats were incomplete. I gather, since there is no "Historical" tab for Negro Leagues, there is no way - at present - to set "Adjust/Weaken" for "minor" leagues to different values than for MLB. Ideally, those levels of PA and IP could be reduced for NeL players, under a "Historical" tab, to reflect the shorter seasons, and missing stats in some cases. Or, what the game engine could do, playing off "Global" MLB settings, is to pro-rate the settings, based on the games for which records are available in a NeL season. In other words, to make it easy, if the NeL Season were effectively 77 games, the "Global" settings for "Adjust/Weaken" would be reduced by half. So, if IP were set at 100/50 [AB or IP] for "Adjust/Weaken" for MLB, the values for NeL would be 50/25. The variable would change based on the number of games played in the NeL season. At some point, maybe seasons under fifty or forty games, the reduction would need to stop, to avoid bizarre results from very limited stats. In an ideal world, there would be greater weight given, even to limited stats, for postseason games between NeL and MLB players. This could be done on a career basis, as well. Seamheads breaks out NeL player performance against MLB players, during the annual October barnstorming games that occurred in the 1920's through the 1940's. Maybe something like "give double [or triple] weight to stats against MLB teams". The performance of the best NeL players in these games is striking. Although a limited sample, it does provide the clearest marker for equalizing NeL stats to MLB play. __________________ Pelican OOTP 2020-? ”Hard to believe, Harry.” Last edited by Pelican; 12-06-2023 at 12:50 PM.

12-06-2023, 04:00 PM	#3
thehef Hall Of Famer Join Date: Jun 2006 Posts: 5,345	This is a great suggestion. However, it should apply to all minor leagues as well as Negro Leagues. Many minor leagues throughout history played significantly fewer games than the 154-162 MLB season. Usually this was by design but sometimes it was because a league folded... Currently OOTP's milbleagues.csv file indicates 345 minor league league-seasons that had 90 or fewer scheduled games. This is not a direct correlation to how many games were actually played, but it's a good indicator. So for most of those season-leagues and depending upon what your adjust/weaken settings are, you're going to have anywhere from most to all of your players (in these leagues) adjusted or weakened. Unless I'm misunderstanding how A/W works... which is a very real possibility Maybe the solution could be something like this, just using hitter AB's as the example: - Gamer sets Adjust to 200 AB's and Weaken to 50 AB's - OOTP applies this to MLB - OOTP looks at the milbleagues.csv file to get the number of games for that season league and does some simple math to determine what the A/W settings should be for that league. The calc would simply be [milbleague games played] / [mlb games played] * [adjust number] - OOPT applies the normalized Adj/Wkn numbers to each minor league See the image below for some examples... The # of games played in the milbleagues.csv file would need to be reviewed because as best as I can tell those numbers represent either scheduled games, or some combination of scheduled & actual. That number would need to be reflective of actual games played in order for this to work...I'll also note that there are two fields in this csv file that have to do with # of games: GamesScheduled and Season Length. For purposes of this example I'm using the former; I don't know exactly how OOTP uses these fields or if it uses one and not the other... but there are two fields and perhaps one could be used for scheduled games to be played (to correspond to season length/games-to-be-played in OOTP), and the other can be used to indicate the number of games actually played... (In the example below of the PCL, it played more games than MLB. In my image below, the adjustment increases Adj/Wkn. However, the calculation could include a cap to where the Adj/Wkn settings wouldn't be higher than MLB's... Or not...) Attached Images

12-10-2023, 01:07 PM	#4
TomVeal All Star Starter Join Date: Jul 2013 Location: Spanaway, Washington Posts: 1,242	There's a problem with this corrective measure: The A/W thresholds are supposed to represent a large enough sample size to provide a valid representation of the player's real life ability. That size doesn't change with the length of the season. Stats based on 50 plate appearances aren't indicative of very much.