Problems With OOTP Potential Ratings

Charlie Hough · 04-27-2010, 09:31 PM

Previously I've only played with recalc in my historical leagues, but I've been running some test leagues using the player development engine instead.

Here are my settings:
Recalc off
Player development on
Potential based on remaining years of career
Potential based on players at the same position
Scouting off

Unfortunately, OOTP seems to make miscalculations when issuing some of the potential ratings. The vast majority of players appear to receive realistic potential ratings, but others are completely divorced from their actual value in the remaining years of their careers. For example, Nolan Ryan and Steve Carlton were imported into my game with one-star ratings for potential.

Why is OOTP doing this? How could it generate a one-star potential rating for two pitchers of the caliber of Nolan Ryan and Steve Carlton?

Another problem is the stamina and roles assigned to pitchers when recalc is turned off. In real life, many pitchers who became All-Star or Hall of Fame starters began their careers as relievers. Since their early career stats are from relief appearances, OOTP always imports these pitchers as relievers and gives them low stamina. That's realistic for their early careers, but I rarely see any of these relievers develop into starters. Since there is no potential rating for stamina, it seems like these pitchers are doomed to remain in the bullpen, and the only guys who are starters for years on end are the guys who import as starters in the first place.

I know there are settings for basing pitcher stamina on a full career, but that seems unrealistic and unfair to me. A better option would be to have a potential rating for stamina and let players naturally develop to that point or possibly fail and remain relievers.

Does anyone have any insight on this? Any suggestions? Have changes been recommended?

Charlie Hough · 04-30-2010, 05:17 PM

Can anyone comment on this or offer any insight? Markus? Anyone?

actionjackson · 05-01-2010, 07:22 PM

Quote:

Originally Posted by Charlie Hough

Previously I've only played with recalc in my historical leagues, but I've been running some test leagues using the player development engine instead.

Here are my settings:
Recalc off
Player development on
Potential based on remaining years of career
Potential based on players at the same position
Scouting off

Unfortunately, OOTP seems to make miscalculations when issuing some of the potential ratings. The vast majority of players appear to receive realistic potential ratings, but others are completely divorced from their actual value in the remaining years of their careers. For example, Nolan Ryan and Steve Carlton were imported into my game with one-star ratings for potential.

Why is OOTP doing this? How could it generate a one-star potential rating for two pitchers of the caliber of Nolan Ryan and Steve Carlton?

Another problem is the stamina and roles assigned to pitchers when recalc is turned off. In real life, many pitchers who became All-Star or Hall of Fame starters began their careers as relievers. Since their early career stats are from relief appearances, OOTP always imports these pitchers as relievers and gives them low stamina. That's realistic for their early careers, but I rarely see any of these relievers develop into starters. Since there is no potential rating for stamina, it seems like these pitchers are doomed to remain in the bullpen, and the only guys who are starters for years on end are the guys who import as starters in the first place.

I know there are settings for basing pitcher stamina on a full career, but that seems unrealistic and unfair to me. A better option would be to have a potential rating for stamina and let players naturally develop to that point or possibly fail and remain relievers.

Does anyone have any insight on this? Any suggestions? Have changes been recommended?

I think something is seriously broken with the "set potential ratings based on..." setting. It appears to not mean a damn thing whether recalc is on or off. One possibility that I just thought of is the "base roles/positions on..." setting (real life stats or AI evaluation). My suggestion is to try setting that to AI evaluation but I'm totally spitballing here. If it's set to real life stats I'm wondering if Ryan and Carlton will import craptacularly and with only two pitches because the AI looks at their first season and a bit and labels them relievers. Then, because recalc is off these two HoFers never get reset to become the studs they actually were.

Absolutely 1000% does not make it right and in fact points to serious issues with the "set potential ratings based on..." setting which, if I'm not mistaken, only appears during the initial setup. There is absolutely no doubt that setting potential ratings according to remaining years of career should be the most realistic way to play a historical game, but until the underlying flaws with this setting are fixed, it will continue to wreak havoc on any attempts to play a historical game and that IMHO is unacceptable. Either take the setting out or make it work, but don't leave it as is. This setting is absolutely pivotal to everything that is part of GMing. It affects drafting, trades, free agency, rule V, waivers, retirement etc etc etc, so not only does the "recalc trumps all" crap need fixing, but so does this setting and post haste - like in the first major patch.

I think it would go a long way towards improving the historical AI (outside of in-game play) and I envision and anticipate an even more tremendous baseball sim and a much more complete historical baseball sim if and when it is fixed.

Charlie Hough · 05-01-2010, 07:47 PM

The good news is that the "recalc trumps all" will be fixed in the next patch. Markus has announced this, so we only need to wait a few weeks to hopefully see the results.

Actually, I've found that the potential ratings are just fine when recalc is turned off, except for the occasional instances like those I mentioned. And I think some players get potential ratings that are too good based on their limited real life time in MLB. I think it would be a good idea for OOTP to adjust potential ratings for players with limited at bats or innings pitched over a relatively short career in the same way that it adjusts initial ratings for these players.

actionjackson · 05-02-2010, 12:14 AM

For those interested in this issue I'm putting together 8 different tests to look at Carlton and Ryan's ratings at the start of 1966. I've got the first three now and will post the other five when I get a chance. All eight use the setting base potential ratings on remaining years of career. The first four (fourth tomorrow) will use neutralized stats, while the final four will use real life stats. Here are the first three with screenshots. Carlton's ratings are on top while Ryan's are underneath in each case.

#1. Neutralized Stats/Recalc On/Base Pos/Roles on RL Stats
#2. Neutralized Stats/Recalc On/Base Pos/Roles on AI Eval
#3. Neutralized Stats/Recalc Off/Base Pos/Roles on RL Stats:

actionjackson · 05-02-2010, 03:51 PM

Now tests 4, 5, and 6. To wit:

#4. Neutralized Stats/Recalc Off/Base Pos/Roles on AI Eval
#5. Real Stats/Recalc On/Base Pos/Roles on RL Stats
#6. Real Stats/Recalc On/Base Pos/Roles on AI Eval

actionjackson · 05-02-2010, 05:22 PM

Finally tests 7, and 8:

#7. Real Stats/Recalc Off/Base Pos/Roles on RL Stats
#8. Real Stats/Recalc Off/Base Pos/Roles on AI Eval

As per usual Carlton as imported on the top and Ryan as imported on the bottom.

actionjackson · 05-02-2010, 06:02 PM

To sum up Charlie and others, I think for the time being Neutralized Stats is a viable workaround to prevent the AI from being a complete ditz and trading Nolan Ryan (or Carlton in '65, which I didn't check on in this little experiment) for peanuts and most likely watching his career path follow the career path of Tanyon Sturtze, or worse. I know using neutralized stats is completely unpalatable to you, but until Markus addresses the Recalc overruling everything issue and more specifically the problems associated with the Base Potential Ratings On... setting these problems are unfortunately going to continue. It also unfortunately demonstrates that my suggestion of using the AI's evaluation to base positions and roles on had a little bit of an effect, but not as much as I was hoping for.

Oh well, I gave it a shot.

Perhaps I was such a jerk in the other thread because I always use neutralized stats, so I couldn't see what the problem was. Sorry about that Mr. Hough.

I'll try to check my assumptions at the door because I know where that leads. In the future I'll try to use some data to back up my crustiness.

Spritze · 05-02-2010, 09:58 PM

Log on fire tossing: If one plays with "Retire According To History" = off but Recalc= On and Potential is based on remaining years of career, how does that potentially affect potentials? Is the 1967 version of Koufax given a potential rating of zero even though he may play 10 more years since he won't get injured and retire? Gehrig won't die so he'll play 6 more years or so? Does that potential potentially affect his trade value? If his potential is zero and that is used in trade evaluation couldn't one collect a bunch of undead/uninjured/unretired all-stars for next to nothing?

Or does the OOTP Development / Aging engine just take over?

Wondering.

04-27-2010, 09:31 PM	#1
Charlie Hough Hall Of Famer Join Date: May 2006 Posts: 3,644	Problems With OOTP Potential Ratings Previously I've only played with recalc in my historical leagues, but I've been running some test leagues using the player development engine instead. Here are my settings: Recalc off Player development on Potential based on remaining years of career Potential based on players at the same position Scouting off Unfortunately, OOTP seems to make miscalculations when issuing some of the potential ratings. The vast majority of players appear to receive realistic potential ratings, but others are completely divorced from their actual value in the remaining years of their careers. For example, Nolan Ryan and Steve Carlton were imported into my game with one-star ratings for potential. Why is OOTP doing this? How could it generate a one-star potential rating for two pitchers of the caliber of Nolan Ryan and Steve Carlton? Another problem is the stamina and roles assigned to pitchers when recalc is turned off. In real life, many pitchers who became All-Star or Hall of Fame starters began their careers as relievers. Since their early career stats are from relief appearances, OOTP always imports these pitchers as relievers and gives them low stamina. That's realistic for their early careers, but I rarely see any of these relievers develop into starters. Since there is no potential rating for stamina, it seems like these pitchers are doomed to remain in the bullpen, and the only guys who are starters for years on end are the guys who import as starters in the first place. I know there are settings for basing pitcher stamina on a full career, but that seems unrealistic and unfair to me. A better option would be to have a potential rating for stamina and let players naturally develop to that point or possibly fail and remain relievers. Does anyone have any insight on this? Any suggestions? Have changes been recommended?

04-30-2010, 05:17 PM	#2
Charlie Hough Hall Of Famer Join Date: May 2006 Posts: 3,644	Can anyone comment on this or offer any insight? Markus? Anyone?

05-01-2010, 07:47 PM	#4
Charlie Hough Hall Of Famer Join Date: May 2006 Posts: 3,644	The good news is that the "recalc trumps all" will be fixed in the next patch. Markus has announced this, so we only need to wait a few weeks to hopefully see the results. Actually, I've found that the potential ratings are just fine when recalc is turned off, except for the occasional instances like those I mentioned. And I think some players get potential ratings that are too good based on their limited real life time in MLB. I think it would be a good idea for OOTP to adjust potential ratings for players with limited at bats or innings pitched over a relatively short career in the same way that it adjusts initial ratings for these players.

05-02-2010, 12:14 AM	#5
actionjackson Hall Of Famer Join Date: Apr 2009 Location: Toronto, ON Posts: 6,179	For those interested in this issue I'm putting together 8 different tests to look at Carlton and Ryan's ratings at the start of 1966. I've got the first three now and will post the other five when I get a chance. All eight use the setting base potential ratings on remaining years of career. The first four (fourth tomorrow) will use neutralized stats, while the final four will use real life stats. Here are the first three with screenshots. Carlton's ratings are on top while Ryan's are underneath in each case. #1. Neutralized Stats/Recalc On/Base Pos/Roles on RL Stats #2. Neutralized Stats/Recalc On/Base Pos/Roles on AI Eval #3. Neutralized Stats/Recalc Off/Base Pos/Roles on RL Stats: Attached Images

05-02-2010, 03:51 PM	#6
actionjackson Hall Of Famer Join Date: Apr 2009 Location: Toronto, ON Posts: 6,179	Now tests 4, 5, and 6. To wit: #4. Neutralized Stats/Recalc Off/Base Pos/Roles on AI Eval #5. Real Stats/Recalc On/Base Pos/Roles on RL Stats #6. Real Stats/Recalc On/Base Pos/Roles on AI Eval Attached Images

05-02-2010, 05:22 PM	#7
actionjackson Hall Of Famer Join Date: Apr 2009 Location: Toronto, ON Posts: 6,179	Finally tests 7, and 8: #7. Real Stats/Recalc Off/Base Pos/Roles on RL Stats #8. Real Stats/Recalc Off/Base Pos/Roles on AI Eval As per usual Carlton as imported on the top and Ryan as imported on the bottom. Attached Images

05-02-2010, 06:02 PM	#8
actionjackson Hall Of Famer Join Date: Apr 2009 Location: Toronto, ON Posts: 6,179	To sum up Charlie and others, I think for the time being Neutralized Stats is a viable workaround to prevent the AI from being a complete ditz and trading Nolan Ryan (or Carlton in '65, which I didn't check on in this little experiment) for peanuts and most likely watching his career path follow the career path of Tanyon Sturtze, or worse. I know using neutralized stats is completely unpalatable to you, but until Markus addresses the Recalc overruling everything issue and more specifically the problems associated with the Base Potential Ratings On... setting these problems are unfortunately going to continue. It also unfortunately demonstrates that my suggestion of using the AI's evaluation to base positions and roles on had a little bit of an effect, but not as much as I was hoping for. Oh well, I gave it a shot. Perhaps I was such a jerk in the other thread because I always use neutralized stats, so I couldn't see what the problem was. Sorry about that Mr. Hough. I'll try to check my assumptions at the door because I know where that leads. In the future I'll try to use some data to back up my crustiness.

05-02-2010, 09:58 PM	#9
Spritze OOTP Historical Czar Join Date: Dec 2001 Location: Bothell Wa Posts: 7,253	Log on fire tossing: If one plays with "Retire According To History" = off but Recalc= On and Potential is based on remaining years of career, how does that potentially affect potentials? Is the 1967 version of Koufax given a potential rating of zero even though he may play 10 more years since he won't get injured and retire? Gehrig won't die so he'll play 6 more years or so? Does that potential potentially affect his trade value? If his potential is zero and that is used in trade evaluation couldn't one collect a bunch of undead/uninjured/unretired all-stars for next to nothing? Or does the OOTP Development / Aging engine just take over? Wondering.