Right aging and random talent modifiers for modern day MLB League

[phightin]

I'm in the process of trying to test out some age and talent modifiers in a modern day MLB world to try to overcome some of the flaws in realistic game play. As most of you know in the OTTP world players will usually flame out, going from good or great players to relative worthless minor leaguers rather quickly when players reach their late 20s. I'm not as concerned about this as I do not think there is a way to overcome that issue but more with trying to get this to occur during players early 30s (32-33) rather than their late 20s (28-29) which should be their prime.

My first question would be if anyone has experimented with this yet and looked into the future to see the results?

My plan is basically to use vorp to look at the top 5% or so of the league from 2011 and establish an average age of the top players. I then want to use that as guide to hopefully hit that target area each year in the game. Right now I will only do hitters as I'm not sure what to do with pitchers yet.

Does anyone have any suggestions on possible age/talent modifier settings that may work best? I will probably lower all of them to try to find a balance.

OTTP is a great game but unfortunately more of a platform for fictional leagues or user controlled ones. Realisitic gameplay and ageing like this has always been a problem.

Note

Out of request I have gone back and updated my original post with the final modifiers that myself and others have come to a consensus on for a real life modern day MLB setting. They are:

Batter Ageing: .250
Batter Development: 1.000
Pitcher Ageing: .375
Pitcher Development: .900

Talent Change: 67

AI Eval: I have experimented with both a ratings heavy one and a stat heavy one and there does not seem to be much of a difference. Right now I am using 30/40/20/10 but I have used 70/20/6/4 even and seen similar results.

Hope this helps and happy OOTPing!

[Mets Man]

Originally Posted by phightin

I'm in the process of trying to test out some age and talent modifiers in a modern day MLB world to try to overcome some of the flaws in realistic game play. As most of you know in the OTTP world players will usually flame out, going from good or great players to relative worthless minor leaguers rather quickly when players reach their late 20s. I'm not as concerned about this as I do not think there is a way to overcome that issue but more with trying to get this to occur during players early 30s (32-33) rather than their late 20s (28-29) which should be their prime.

My first question would be if anyone has experimented with this yet and looked into the future to see the results?

My plan is basically to use vorp to look at the top 5% or so of the league from 2011 and establish an average age of the top players. I then want to use that as guide to hopefully hit that target area each year in the game. Right now I will only do hitters as I'm not sure what to do with pitchers yet.

Does anyone have any suggestions on possible age/talent modifier settings that may work best? I will probably lower all of them to try to find a balance.

OTTP is a great game but unfortunately more of a platform for fictional leagues or user controlled ones. Realisitic gameplay and ageing like this has always been a problem.

I disagree that this is a problem only for MLB leagues. I think this can be a problem for fictional leagues too. It's not about having players perform like their real life counterparts. Its a question of realistic aging patterns. You expect most guys in your league to decline in their 30s in both MLB and fictional leagues. Yes, some players will have talent drops during their prime years, but it shouldn't be the norm (for both MLB and fictional leagues).

You could make the argument that perhaps it is funner for gameplay-sake to have such randomness, since you just never know what's going to happen. But this problem also hurts the gameplay because it takes away the decision of ever giving a big long term contract to any good player. Never give a long term contract to a player coming off a good year - he most definitely will not live up to it with the talent change randomness the way it is.

[phightin]

I agree its a problem with all areas of the game. It's simply OOTP's development system.

Most people won't like to hear this but I wish they would switch to some sort of regneration formulamatic pattern like FM utilizes. Guys develop and usually peak between 27-31 and then as they retire they are regenerated into the game. It would be the only real way to balance out the development and talent in distribution in the game imo. The current system just doesnt cut it.

[RchW]

I have a very good experience with development and aging. My age distribution is slightly young compared to MLB, but it is difficult to tweak as the results of a tweak take some time.

The other thing is I see a lot of posts but no one is posting data as to the age distribution in their leagues. Without some data how can we identify the problem.

Just to give you a small sample

Batters 35+ in game 64 in MLB 2011 55

Batters 30-34 in game 131 in MLB 2011 144

Pitchers 35+ in game 53 in MLB 2011 48

Pitchers 30-34 in game 124 in MLB 2011 137

Pitchers 17-25 in game 108 in MLB 2011 146
Batters 17-25 in game 144 in MLB 2011 162

It's by no means perfect, but at least I have data. Show me some data!!!

[SpaceMonkey13]

Originally Posted by RchW

I have a very good experience with development and aging. My age distribution is slightly young compared to MLB, but it is difficult to tweak as the results of a tweak take some time...

Could you post your settings?

[Charlie Hough]

Yes, we see a constant stream of postings from people who claim that there are various flaws in OOTP with aging and development patterns, frequency of certain types of plays, etc.

Yet we almost never see anyone post any data sets with a sample size to back up their claims.

The reality is that the vast majority of 'flaws' that people perceive are actually a product of their own cognitive bias. For example, they believe that it is rare for players in their late 20's to experience a rapid decline in their productivity or ability. Yet we have countless cases from MLB history to show that this is a very common occurrence.

Now, if you can show full sets of data to compare the number of OOTP flame-outs to those in MLB history, under controlled and identical conditions, then we'll be getting somewhere.

A lot of effort has been expended over the years to get this game to produce results that mirror real life. The problem is that most people's perceptions of real life are based on their limited observations and hunches, and if they say the actual MLB statistics for these issues, they would be shocked at what they find.

[RchW]

Here is data from a more recent league, batters only. This is number of batters for a given age. It should be obvious what's wrong with my development and aging modifiers. Any one care to say what I should do with the settings?

What else do we see. In real life for hitters there is a big move to younger players and a big drop off after 33. Is MLB unrealistic?

Show data before making conclusions.

EDIT

This is a revised chart. I forgot to normalize the data.

[RchW]

Here is the pitcher data. Note how skewed it is to older pitchers. I certainly need to tweak this league quite a bit.

Note also how in real life pitchers over 33 are an endangered species.

[oman19]

*Raises hand eagerly* oh oh oh I know I know I know....you need to up your player dev modifiers

[Mets Man]

Originally Posted by RchW

Here is data from a more recent league, batters only. This is number of batters for a given age. It should be obvious what's wrong with my development and aging modifiers. Any one care to say what I should do with the settings?

What else do we see. In real life for hitters there is a big move to younger players and a big drop off after 33. Is MLB unrealistic?

Show data before making conclusions.

EDIT

This is a revised chart. I forgot to normalize the data.

I hear what you're saying about providing data, its something none of us have done, so you got us there.

It could all be perception, I suppose, but players in our game just appear to be dropping off at the ages 25-29, we're talking tons of players. And I'm not talking about simply declining - I'm talking about falling off completely like to 1 - 2 stars. These former stars are turning into minor leaguers within one or two years of their last good season. I've only named about 15 cases, but I've noticed many many more. Like I said, I haven't compiled data using the scientific method to track this, but its just something that we noticed.

I can send you data files, but its probably too late now, because I just recalculated my league. Nevertheless, you could ask me about any player from the 80s. And I'll look up what their scouting report was in the last year (before I recalculated). Try to name one player that you know and see if that player still had good ratings. Trust me, I've tried this and I'm hard pressed to find one guy who had decent ratings. Name one player in the late 80s that was playing in 1984. Wade Boggs? Nope, he didn't make it, he dropped off a looong time ago (he hasn't reached 30 years old yet either). Try another guy. Honestly, I don't know any players that are still good in my league that were around when I first created the league (1984). Every player that I started when I created to the league with is a dud now (young or old).

Maybe all of us are going by perception, but I guess when I did a historical league, I expected it to have some variation when compared to real life (in terms of who performed well and who didn't), but I didn't expect every notable player that I knew from the era to completely become a dud within the first few years.

[Mets Man]

Originally Posted by RchW

Here is data from a more recent league, batters only. This is number of batters for a given age. It should be obvious what's wrong with my development and aging modifiers. Any one care to say what I should do with the settings?

What else do we see. In real life for hitters there is a big move to younger players and a big drop off after 33. Is MLB unrealistic?

Show data before making conclusions.

EDIT

This is a revised chart. I forgot to normalize the data.

Perhaps your charts seem to look normal because the talent randomness is switching up the good players into bad players and the bad players into good players so the numbers in terms of talented players stays leveled and normalized.

What I'm saying is, perhaps for every Wade Boggs, Don Mattingly, Roger Clemens etc etc who turn to duds, there is a Rafael Santana, John Cerutti, Nelson Liriano who turns into a star.

If this is the case though, it doesn't make it realistic or functional. Yes, there should be a few cases of stars turning into duds AND some cases of duds turning into stars, BUT this should be the exception, NOT the norm. If indeed stars are switching places with the duds to make the league look completely different after a few years (in terms of who the top performers are) then that's still a problem. Sure from a statistical point of view you're numbers look good, they work out fine. But there's still a problem there. Doesn't anybody else think there is a problem when your league's top performers are no longer recognizable by the 5th year into your league?

For your MLB 2011 league, would you be able to check who the top performers are and see if we recognize them? Could you do this after like 6-8 years into the league. Are any of the stars of 2011 still stars by 2018? If you happen to find one player, then look inside his ratings and stats and see if he had sustained consistent ratings/stats (at a star level). See how many consistent years these guys put together before falling off track if they did fall off track. Do you have anybody that didn't fall off track? I'm not necessarily looking for Albert Pujols level performers. But just consistent star players who had sustained success (ratings and stats) through their prime years for at least 5 seasons. In real life, I can name tons of these type of cases (these type of consistent performers). Can you find any in your simulation?

[RchW]

Originally Posted by Mets Man

I hear what you're saying about providing data, its something none of us have done, so you got us there.

It could all be perception, I suppose, but players in our game just appear to be dropping off at the ages 25-29, we're talking tons of players. And I'm not talking about simply declining - I'm talking about falling off completely like to 1 - 2 stars. These former stars are turning into minor leaguers within one or two years of their last good season. I've only named about 15 cases, but I've noticed many many more. Like I said, I haven't compiled data using the scientific method to track this, but its just something that we noticed.

I can send you data files, but its probably too late now, because I just recalculated my league. Nevertheless, you could ask me about any player from the 80s. And I'll look up what their scouting report was in the last year (before I recalculated). Try to name one player that you know and see if that player still had good ratings. Trust me, I've tried this and I'm hard pressed to find one guy who had decent ratings. Name one player in the late 80s that was playing in 1984. Wade Boggs? Nope, he didn't make it, he dropped off a looong time ago (he hasn't reached 30 years old yet either). Try another guy. Honestly, I don't know any players that are still good in my league that were around when I first created the league (1984). Every player that I started when I created to the league with is a dud now (young or old).

Maybe all of us are going by perception, but I guess when I did a historical league, I expected it to have some variation when compared to real life (in terms of who performed well and who didn't), but I didn't expect every notable player that I knew from the era to completely become a dud within the first few years.

Lets be sure we are talking apples to apples here. Your earlier post referenced fictional leagues now you talk historical. My data refers to a modern day MLB fictional. It may I stress may apply to a historical league that doesn't have recalc.

Also keep in mind that a historical league with development on is a fictional league with recognizable names. There is no guarantee any of those names will develop how you like. It's just numbers to the AI

[CatKnight]

Originally Posted by Mets Man

What I'm saying is, perhaps for every Wade Boggs, Don Mattingly, Roger Clemens etc etc who turn to duds, there is a Rafael Santana, John Cerutti, Nelson Liriano who turns into a star.

If you're expecting the same players (more or less) to rise to the top as did historically, then your settings are wrong. You need to turn on recalc. Numbers are most reliable with a one year recalc, though most posts I recall reading recommend three years to balance out career paths.

Someone (I forget who) posted that if you're playing without recalc on, you're playing a fictional league that just happens to have recognizable teams and names. Now, there's nothing wrong with that....but nor can we be surprised when someone's career deviates very sharply from their historical performance in this case.

[RchW]

Originally Posted by Mets Man

Perhaps your charts seem to look normal because the talent randomness is switching up the good players into bad players and the bad players into good players so the numbers in terms of talented players stays leveled and normalized.

What I'm saying is, perhaps for every Wade Boggs, Don Mattingly, Roger Clemens etc etc who turn to duds, there is a Rafael Santana, John Cerutti, Nelson Liriano who turns into a star.

If this is the case though, it doesn't make it realistic or functional. Yes, there should be a few cases of stars turning into duds AND some cases of duds turning into stars, BUT this should be the exception, NOT the norm. If indeed stars are switching places with the duds to make the league look completely different after a few years (in terms of who the top performers are) then that's still a problem. Sure from a statistical point of view you're numbers look good, they work out fine. But there's still a problem there. Doesn't anybody else think there is a problem when your league's top performers are no longer recognizable by the 5th year into your league?

For your MLB 2011 league, would you be able to check who the top performers are and see if we recognize them? Could you do this after like 6-8 years into the league. Are any of the stars of 2011 still stars by 2018? If you happen to find one player, then look inside his ratings and stats and see if he had sustained consistent ratings/stats (at a star level). See how many consistent years these guys put together before falling off track if they did fall off track. Do you have anybody that didn't fall off track? I'm not necessarily looking for Albert Pujols level performers. But just consistent star players who had sustained success (ratings and stats) through their prime years for at least 5 seasons. In real life, I can name tons of these type of cases (these type of consistent performers). Can you find any in your simulation?

I think your missing something. My charts are comparing my fictional league with real life baseball in 2011. My development and aging modifiers need work. If you look at the charts I have too many 30 somethings and not enough 24-28 year olds.

Edit: to be clear my charts are not normal. My age distribution is off compared to what it should be. If you can't see this please look again.

What I'm trying to say is that you cant just cherry pick names; you have to look at league wide data to establish what settings are best. You also must compare similar leagues. I can't and don't know if your league is set up the same way as mine.

[Eiskrap]

I started looking at this yestreday as well, here is my data!

I looked at it in a different way, and I've only done this for batters so far.

Using the MLB quick start (i'm assuming this is the way the game is supposed to work 'most accurately' and match modern day MLB). I simmed to 2030 and compared the players from that game to players from the 2007-2010 Lahman database. Ideally I would have tested more than one season, however with the player search page defaulting to page selectors with the MLB game I could only face doing it once last night.

The two things I looked at were the proportion of ABs by age, as well as high level performance by age (BA and OBP). Looking at these should give a decent indication of not only whether players were playing for long enough, but also if they were maintaining their level of performance correclty throughout their carears. I understand there are some flaws in this logic, but I think give a decent suggestion for aging and development modifiers that could be used.

So for my first test I ran a default MLB quick start until 2030 to see the distribution of batter and this is what I got:





Initial thoughts are:

Between the ages of 20-25 the proportion of ABs looks almost spot on!
Between the ages of 26-33 these players are getting too many ABs
Conversely after 33 the proportion of ABs is too low

BA on the whole looks too low, this should be expected with OOTP using 2011 modifiers and 2011 being a better year for pitchers than 2007-2010. However the biggest variance looks like it is from 33 onwards which would explain why these players are not seeing enough ABs, because they have aged too quickly and lost ability.

Conclusions:

Although the sample size of OOTP data is limited to one season, this backs up what most people are seeing, so I'm happy to run with it. Players are aging too fast, their skills are falling off too quickly especially in the 33+ age range. This is resulting in teams signing 33 year old palyers to 3-4 years deals but getting 2-3 seasons of a .220 BA before paying them $13m a year to sit in the minors.

Next Steps:

Not one to do things by half, I'm a fan of trying to find the opposite end of the spectrum first and then work towards the middle ground. I have run a sim until 2040 default MLB quickstart with aging modifiers at 0.3 and development 1.1 (to try and balance out the additional older players so we still seeing some 22/23 year olds.) I'll post results when I have collated, hopefully tonight.

[henry296]

Originally Posted by Mets Man

Perhaps your charts seem to look normal because the talent randomness is switching up the good players into bad players and the bad players into good players so the numbers in terms of talented players stays leveled and normalized.

What I'm saying is, perhaps for every Wade Boggs, Don Mattingly, Roger Clemens etc etc who turn to duds, there is a Rafael Santana, John Cerutti, Nelson Liriano who turns into a star.

If this is the case though, it doesn't make it realistic or functional. Yes, there should be a few cases of stars turning into duds AND some cases of duds turning into stars, BUT this should be the exception, NOT the norm. If indeed stars are switching places with the duds to make the league look completely different after a few years (in terms of who the top performers are) then that's still a problem. Sure from a statistical point of view you're numbers look good, they work out fine. But there's still a problem there. Doesn't anybody else think there is a problem when your league's top performers are no longer recognizable by the 5th year into your league?

For your MLB 2011 league, would you be able to check who the top performers are and see if we recognize them? Could you do this after like 6-8 years into the league. Are any of the stars of 2011 still stars by 2018? If you happen to find one player, then look inside his ratings and stats and see if he had sustained consistent ratings/stats (at a star level). See how many consistent years these guys put together before falling off track if they did fall off track. Do you have anybody that didn't fall off track? I'm not necessarily looking for Albert Pujols level performers. But just consistent star players who had sustained success (ratings and stats) through their prime years for at least 5 seasons. In real life, I can name tons of these type of cases (these type of consistent performers). Can you find any in your simulation?

In my 1969-2011 test league, I definitely have some stars remain stars. For example, Jim Thome and A-Rod both ended up with over 500 homers, but Reggie Jefferson turned into one of the best players of all time with 773 HRs. Both Boggs and Gwynn ended up as HOFers. I think tonight, I'm going to try a more detailed tracking.

[RchW]

Eiskrap, nice work!

I'll have to check that out because my age approach has its problems too.

[phightin]

Originally Posted by Charlie Hough

Yes, we see a constant stream of postings from people who claim that there are various flaws in OOTP with aging and development patterns, frequency of certain types of plays, etc.

Yet we almost never see anyone post any data sets with a sample size to back up their claims.

The reality is that the vast majority of 'flaws' that people perceive are actually a product of their own cognitive bias. For example, they believe that it is rare for players in their late 20's to experience a rapid decline in their productivity or ability. Yet we have countless cases from MLB history to show that this is a very common occurrence.

Now, if you can show full sets of data to compare the number of OOTP flame-outs to those in MLB history, under controlled and identical conditions, then we'll be getting somewhere.

A lot of effort has been expended over the years to get this game to produce results that mirror real life. The problem is that most people's perceptions of real life are based on their limited observations and hunches, and if they say the actual MLB statistics for these issues, they would be shocked at what they find.

Yes because hundreds upon hundreds of posts with people having the same issues year after year is all based on pure speculation

Too many people here are too loyal to the game at times and cant see past it's flaws so that it can improve. Yes we all love the game that's why we're here but the inability to be able to critique it won't allow for improvement.

With that said I appreciate the people posting their data so far and plan to do the same today. The two problems though are that they simply assess the overall entire universe and do not measure the individual or rate of variance among players performing at peak levels in the game. For example, from the charts from Eiskrap prove what we are saying in overall performance dropoffs. However they and the others posted on here dont take into account rates of change among players performing highly and then dropping after a few years and rather look at the broad spectrum of data. So you could have a group 10-12 players aged 25-26 performing highly for a couple of years then all drop off after a a couple years and another 10-12 players from the same age range take their place.

This makes some sense as there should be a cycle but there should be an amount of consistency in the league as well. To measure this would take a lot of time so I'm not surprised no on here has yet.

I'm going to take some samples today and I would even argue that the results will show even more problems with ageing and development than Eiskrap's showed.

[RchW]

Originally Posted by phightin

Yes because hundreds upon hundreds of posts with people having the same issues year after year is all based on pure speculation

Too many people here are too loyal to the game at times and cant see past it's flaws so that it can improve. Yes we all love the game that's why we're here but the inability to be able to critique it won't allow for improvement.

With that said I appreciate the people posting their data so far and plan to do the same today. The two problems though are that they simply assess the overall entire universe and do not measure the individual or rate of variance among players performing at peak levels in the game. For example, from the charts from Eiskrap prove what we are saying in overall performance dropoffs. However they and the others posted on here dont take into account rates of change among players performing highly and then dropping after a few years and rather look at the broad spectrum of data. So you could have a group 10-12 players aged 25-26 performing highly for a couple of years then all drop off after a a couple years and another 10-12 players from the same age range take their place.

This makes some sense as there should be a cycle but there should be an amount of consistency in the league as well. To measure this would take a lot of time so I'm not surprised no on here has yet.

I'm going to take some samples today and I would even argue that the results will show even more problems with ageing and development than Eiskrap's showed.

So you really think that Charlie Hough and I are the problem and not part of the solution?

If you produce data that shows a problem Markus will look at it. No one is saying that issues don't exist but there is no way to take unsubstantiated observations and make them the basis for changing the game. But carry on you obviously don't need any help from data to reach your conclusions.

[phightin]

Am currently looking at the top 100 hitters in the league. I just simmed the entire 2012 mlb season using default settings here is what i found.

Average age of top 105 hitters = 29.09
Median = 29
avg age of top 10 = 27.6
20= 28.2
30= 28.43333
40 = 28.625
50 = 29
60 = 29
70 = 29.4
80 = 29.4375


I can agree with all of this obviously right now. 29 would be the prime year of hitters in general and i have no problem with the top 10-20 being in the 27-28 range.

Am currently performing a sim 10 years into the future. Be ready for age to drop off by 2-3 years IMO.