Right aging and random talent modifiers for modern day MLB League

[Eiskrap]

My graphs don't have VORP, just the stats you mention. You'd have to check with one of the other guys as to how they did their VORP graphs.

[phightin]

Originally Posted by BusterKing

So those last charts are of a result from these settings ?

RATINGS: 80%
CURRENT YEAR: 15%
PREVIOUS YEAR: 4%
2 YEARS AGO: 1%

Sabermetric Lineups

BATTER AGING: .250
BATTER DEV: 1.000
PITCHER AGING: .375
PITCHER DEV: .875

INJURIES: OOTP Classic
TALENT CHANGE RANDOMNESS: 67

Been swamped with work and finishing my masters degree the past week but plan on running a test sim with these tonight. Out of curiousity I've never used Sabermetric Lineups, anyone care to share to impact anticipated that they will have.

Will report back on my findings sometime tonight or tomorrow.

[phightin]

Just got done a 10 year sim using the same settings I posted excepted I tweaked the AI eval to 75/20/4/1

All I have to say is FINALLY!!!!!!!! I will post hard data tomorrow but numbers and just the overall feel of the game FINALLY seems real. The age distributions are matching up as good as they probably can be, and most of all there's guys having long careers in the MLB. Im not talking necessarily about hall of fame guys either, but more so guys who are pretty decent that grind it out for 12-13 years. At the same time theres also a balance of younger players. Guys are actually warranting their huge contracts somewhat too and even thats more realistic. For instance I saw several Pujols type players in their early 30s who were granted huge contracts as FAs. These guys rewarded those teams with 2-3 great years followed by hurting those teams the last 2-3 years by starting to drop off. Love realism.

I do anticipate there being a higher influx of HOFers mainly because of the original modern day players with their pre-existing stats but figure it will level off as more of those guys retire.

However, I am finally at peace now and have found settings, at least with OTTP13 that should work and allow me to continue playing.

[Mets Man]

Originally Posted by phightin

Just got done a 10 year sim using the same settings I posted excepted I tweaked the AI eval to 75/20/4/1

All I have to say is FINALLY!!!!!!!! I will post hard data tomorrow but numbers and just the overall feel of the game FINALLY seems real. The age distributions are matching up as good as they probably can be, and most of all there's guys having long careers in the MLB. Im not talking necessarily about hall of fame guys either, but more so guys who are pretty decent that grind it out for 12-13 years. At the same time theres also a balance of younger players. Guys are actually warranting their huge contracts somewhat too and even thats more realistic. For instance I saw several Pujols type players in their early 30s who were granted huge contracts as FAs. These guys rewarded those teams with 2-3 great years followed by hurting those teams the last 2-3 years by starting to drop off. Love realism.

I do anticipate there being a higher influx of HOFers mainly because of the original modern day players with their pre-existing stats but figure it will level off as more of those guys retire.

However, I am finally at peace now and have found settings, at least with OTTP13 that should work and allow me to continue playing.

I'm not sure these settings work for historical leagues very well. The reason being, when you import players each year, alot of the star players come in with high initial ratings. These young players (that could be anywhere between 18-24 years old) post up MVP/Cy Young numbers in their first year. So, I had to lower development speed down even more, just so that these young 18-24 year olds don't dominate my league as they have been.

[phightin]

Originally Posted by Mets Man

I'm not sure these settings work for historical leagues very well. The reason being, when you import players each year, alot of the star players come in with high initial ratings. These young players (that could be anywhere between 18-24 years old) post up MVP/Cy Young numbers in their first year. So, I had to lower development speed down even more, just so that these young 18-24 year olds don't dominate my league as they have been.

I wouldn't say these would be ideal for historical settings as you mentioned. I think my intention from the start of things was to create settings that would mirror a modern day MLB Universe. Most guys, even great college guys coming out of the draft still will spend at least a couple years in the minors before theyre up. Highschoolers at least 3-4 years.

[The Wolf]

I'm thinking about trying stats-only with feeder leagues for the first time. Anyone have any idea about what settings would work best with the rating evaluation zeroed out?

[phightin]

Originally Posted by The Wolf

I'm thinking about trying stats-only with feeder leagues for the first time. Anyone have any idea about what settings would work best with the rating evaluation zeroed out?

I'm not sure exactly. I know that the problems in the past with ratings eval zeroed out others have run into have been with the AI handling of free agency and players that were still highly rates being left in FA over a course of a few years.

I have finally come to terms on my ideal settings thanks to help of the people on this post. I must say even with the age modifiers toned down to the extremes that they were upon examining a 20 year sim players over thirty rated top 105 in VORP were actually still slightly down. I can live this however it really does prove my point of ageing being off with the game. I think others on here do not mind or choose to ignore it which is fine and their own preference but it's definitely there.

The last step now is figuring out whether I want to keep lineups traditional or sabremeteric and also if i want the injury rate to progress as it does throughout a season irl. This would be keeping injuries high during spring training april and may, classic june july, and then very low august september and playoffs. The only thing im worried about is having a lower rate of injuries that hurts the ageing modifiers and in turn would allow for too many older players to still be around as time goes on.

[phightin]

In the final process of fixing the ageing modifiers to my likeing. I do think after several test sims and seeing the results batting ageing at .250 and development at 1.000 is about as perfect as you will get. As far as pitching goes it has been a struggle to say the least.

For pitching the at .375 and development at .875 let me first start by saying development is close to percfect. The distribution of guys 25 and under is about spot on and while every now and then a great 20-21 old will arive its a very rare. Most high school drafted pitchers will spend at least 4-5 years in the minors developing which is also realistic. As far as ageing goes however it has been a struggle. You see at .375 with development at .875 you get probably too many 30+ guys hanging around. However the problem is while you have about an extra surplus of 30+ guys there still isnt enough guys 33+ and 35+ in the game so its a tough problem.

[phightin]

After a solid 2-3 days of simming I have reached my final conclusions of what I feel to be the perfect ageing and development modifiers for the game (or as close to perfect as you can get) in a modern day MLB setting.

Batting Ageing: .250
Batting Development: 1.000
Pitcher Ageing: .375
Pitcher Development: .900
Talent Change: 67

AI EVAL Option: 40/30/20/10 although I still think you could use a ratings heavy one and it wont make a huge difference.

Using this both batting and pitching age distribution reflect close to real life numbers. For pitchers you have a rare few guys pitch into their 40s. I upped the development from .875 because there were too many 30+ guys hanging around and this seems to have fixed the problem.

So hopefully this is my final post with this issue. I thank everyone for their help. It certainly has been a process.

[Mets Man]

Originally Posted by phightin

After a solid 2-3 days of simming I have reached my final conclusions of what I feel to be the perfect ageing and development modifiers for the game (or as close to perfect as you can get) in a modern day MLB setting.

Batting Ageing: .250
Batting Development: 1.000
Pitcher Ageing: .375
Pitcher Development: .900
Talent Change: 67

AI EVAL Option: 40/30/20/10 although I still think you could use a ratings heavy one and it wont make a huge difference.

Using this both batting and pitching age distribution reflect close to real life numbers. For pitchers you have a rare few guys pitch into their 40s. I upped the development from .875 because there were too many 30+ guys hanging around and this seems to have fixed the problem.

So hopefully this is my final post with this issue. I thank everyone for their help. It certainly has been a process.

Thanks for the research, very useful.

I've been toying around with these settings myself. I don't really go by hard data in my analysis, so I imagine I would be less accurate than you.

But first difference, I have my AI Evaluation Settings as: 32/44/16/8 and the Current Ratings turned off, but Star Overall Ratings on. The reason being, I want to make it like real life where most people judge current ability on current performance. Yes, scouting has a bit to do with how good you think a player current is, but actual statistical performance has a way of swaying/biasing opinion. That's why I also have the Star Overall Ratings take into account AI Evaluations, so good statistics will inflate a player's overall value. Also, when it comes to contract time, I feel that in real life, statistics are what people mostly go by when assessing how much a person is worth, not necessarily scouts opinion.

My development modifiers as:
Batting Ageing: .235
Batting Development: 1.000
Pitcher Ageing: .375
Pitcher Development: .880
Talent Change: 38

My justification? Well, I feel that batters drop off a cliff too predictably once they hit their 30s. Yes, I suppose in real life, most batters do drop off in their 30s. I just feel that I like to still have some sustaining ability for some players into their early 30s. With .250, I felt that 30+ year old players were obsolete, so when it came to long term contract time, you never bothered to give a contract to a player into their 30s. I wanted it such that you'd have some of the best players in your league to be 31 to 32 years old. And maybe have a few players who were still playing at a high level when they're 33-36, although rarely. I guess its my old school thinking. When I think about players in their 80s like Wade Boggs, Tony Gwynn and a few others that were still hitting over .300 into their mid to late 30s. Also, the steroid era with old guys still performing (the steroid era went a little too far with aging stars).

I suppose modern day, people are aging very quickly so that most people start to become horrible by 32 or 33 years old.

With the pitching modifiers, I didn't want to go as far as you did in terms of old pitchers. I like that I have old pitchers that can still be effective. From what I understand, pitchers develop later than batters in real life. I'm wondering, don't they have sustaining ability into their 30s? Aren't old pitchers (I mean 32-34 years old) still very effective in real life? I'm thinking about Roy Halladay the last couple of years.

I always thought that a ballplayer's (Batter and Pitcher) prime was in the ages 27-32 roughly. So, I would naturally assume that a player who is 32 years old would remain elite. I found that with your development settings (.250, 1.000, .375, .900), most players that were 31 and 32 were way off their primes and into the decline phase. To the extent that all the Top Players in the league were between 25-29. Almost nobody in their 30s cracked the top list of players.

Finally, I don't like the Talent Change Randomness, I find it's way too prevalent on the high settings. I fiddled around with it and even found that lowering to as low as 30 still produces lots of randomness. I do like diamonds in the rough and some late draft picks, sleepers becoming stars. But I don't like the fluctuations that occur with high settings because it makes it such that there aren't many consistent good performers, to the extent that you don't become accustomed to the good players in your league because they keep changing so often. I like having household names in my league with occasional new household names emerging and some players having off years. But overall, I feel that game is almost random enough to produce that even without the Talent Change Randomness.

[Mets Man]

Oh, just to add to my above post.

With the Pitcher Development settings (.375, .875) that was originally proposed. I find that yes, there are probably too many old pitchers (in their mid to late 30s) that are good, but those are mostly relievers. Look at the starters and you'll find that there aren't that many really good 33+ year old starters in the league. Relievers you'll find good to elite 36-40+ year olds.

I don't know if that reflects real life or not. Do a lot of relievers sustain their ability into their late 30s in real life?

[phightin]

I rate by VORP so all of the top 20-30 pitchers are usually starters. Interesting enough the average age of top pitchers IRL has gotten considerably lower over time. Doing several sims the average age of the top 10 pitchers in the MLB for 2012 was 27 top 20 27.9. When you look at real life with guys like Hamels, Kershaw, Strasburg, etc it makes perfect sense. While guys like Halladay can certainly be effective into their mid-30s they are the exception rather than the rule.

I find with .375 and .900 for pitchers you get as close to a realistic distribution that you can. For instance the average age of the top 10 pitchers in 2035 was 27.3 top 20 was 28.5. There certainly were several starters in the top 42 of my league who were in their 30s about 15 in the top 42 and several in their mid to late 30s one even in their 40s. When looking at the overall top 105 pitchers through VORP the average of guys aged 30+ is 40%, you get that consistently with those settings. Also the peak for most guys is 27-29 which I can live with.

I pointed out the problem is though whether you have the settings at .375 .875 or use .900 there arent enough guys aged 33+ and 35+. However there is really no easy solution to this. If you use .875 for development I found there to be seasons where you had over 50% of pitchers in the top 105 for VORP aged 30+ and too many guys overall throughout the league pitching into their 40s. Pitchers even good ones with long careers should start to retire in their late 30s, 40 should be rare. This in my mind was way too many even if you got extra guys to cover the 33+ 35+ range.

With .900 I found it at least keeps the numbers down to 40% overall each year for the 30+ guys and close to a realistic distribution. While you still can not cover the age ranges 33+ 35+ guys at least are having realistic career spans and their peak is 27-32 which is good enough for me.

[evanbarth]

So is this a change from prior versions. I've used 11 in the past and the modifiers were much higher like .600? Why the change?

Originally Posted by Mets Man

Thanks for the research, very useful.

I've been toying around with these settings myself. I don't really go by hard data in my analysis, so I imagine I would be less accurate than you.

But first difference, I have my AI Evaluation Settings as: 32/44/16/8 and the Current Ratings turned off, but Star Overall Ratings on. The reason being, I want to make it like real life where most people judge current ability on current performance. Yes, scouting has a bit to do with how good you think a player current is, but actual statistical performance has a way of swaying/biasing opinion. That's why I also have the Star Overall Ratings take into account AI Evaluations, so good statistics will inflate a player's overall value. Also, when it comes to contract time, I feel that in real life, statistics are what people mostly go by when assessing how much a person is worth, not necessarily scouts opinion.

My development modifiers as:
Batting Ageing: .235
Batting Development: 1.000
Pitcher Ageing: .375
Pitcher Development: .880
Talent Change: 38

My justification? Well, I feel that batters drop off a cliff too predictably once they hit their 30s. Yes, I suppose in real life, most batters do drop off in their 30s. I just feel that I like to still have some sustaining ability for some players into their early 30s. With .250, I felt that 30+ year old players were obsolete, so when it came to long term contract time, you never bothered to give a contract to a player into their 30s. I wanted it such that you'd have some of the best players in your league to be 31 to 32 years old. And maybe have a few players who were still playing at a high level when they're 33-36, although rarely. I guess its my old school thinking. When I think about players in their 80s like Wade Boggs, Tony Gwynn and a few others that were still hitting over .300 into their mid to late 30s. Also, the steroid era with old guys still performing (the steroid era went a little too far with aging stars).

I suppose modern day, people are aging very quickly so that most people start to become horrible by 32 or 33 years old.

With the pitching modifiers, I didn't want to go as far as you did in terms of old pitchers. I like that I have old pitchers that can still be effective. From what I understand, pitchers develop later than batters in real life. I'm wondering, don't they have sustaining ability into their 30s? Aren't old pitchers (I mean 32-34 years old) still very effective in real life? I'm thinking about Roy Halladay the last couple of years.

I always thought that a ballplayer's (Batter and Pitcher) prime was in the ages 27-32 roughly. So, I would naturally assume that a player who is 32 years old would remain elite. I found that with your development settings (.250, 1.000, .375, .900), most players that were 31 and 32 were way off their primes and into the decline phase. To the extent that all the Top Players in the league were between 25-29. Almost nobody in their 30s cracked the top list of players.

Finally, I don't like the Talent Change Randomness, I find it's way too prevalent on the high settings. I fiddled around with it and even found that lowering to as low as 30 still produces lots of randomness. I do like diamonds in the rough and some late draft picks, sleepers becoming stars. But I don't like the fluctuations that occur with high settings because it makes it such that there aren't many consistent good performers, to the extent that you don't become accustomed to the good players in your league because they keep changing so often. I like having household names in my league with occasional new household names emerging and some players having off years. But overall, I feel that game is almost random enough to produce that even without the Talent Change Randomness.

[Mets Man]

Originally Posted by phightin

I rate by VORP so all of the top 20-30 pitchers are usually starters. Interesting enough the average age of top pitchers IRL has gotten considerably lower over time. Doing several sims the average age of the top 10 pitchers in the MLB for 2012 was 27 top 20 27.9. When you look at real life with guys like Hamels, Kershaw, Strasburg, etc it makes perfect sense. While guys like Halladay can certainly be effective into their mid-30s they are the exception rather than the rule.

I find with .375 and .900 for pitchers you get as close to a realistic distribution that you can. For instance the average age of the top 10 pitchers in 2035 was 27.3 top 20 was 28.5. There certainly were several starters in the top 42 of my league who were in their 30s about 15 in the top 42 and several in their mid to late 30s one even in their 40s. When looking at the overall top 105 pitchers through VORP the average of guys aged 30+ is 40%, you get that consistently with those settings. Also the peak for most guys is 27-29 which I can live with.

I pointed out the problem is though whether you have the settings at .375 .875 or use .900 there arent enough guys aged 33+ and 35+. However there is really no easy solution to this. If you use .875 for development I found there to be seasons where you had over 50% of pitchers in the top 105 for VORP aged 30+ and too many guys overall throughout the league pitching into their 40s. Pitchers even good ones with long careers should start to retire in their late 30s, 40 should be rare. This in my mind was way too many even if you got extra guys to cover the 33+ 35+ range.

With .900 I found it at least keeps the numbers down to 40% overall each year for the 30+ guys and close to a realistic distribution. While you still can not cover the age ranges 33+ 35+ guys at least are having realistic career spans and their peak is 27-32 which is good enough for me.

I didn't mean that the Top 10 list of pitchers are relievers. That list will almost always put starters on there, very rarely will you see a reliever on that list no matter how good they are. What I meant was, I found not enough very good/elite starters were in their 30s, but there were too many very good/elite relievers that were over 35. If you sort with just starters, you'll be hard pressed to find a highly rated 34+ player. But if you sort by all relievers, you'll find plenty highly rated 34+ players.

I'm curious, if you found that there were too many old old pitchers that were too effective, why did you decide to edit the pitcher development modifier instead of the pitching aging modifier?

In most cases, whenever I feel that old players are too effective for too long, I lower the aging modifier. Whenever, I feel that young players are too effective, then I lower the development modifier. You seemed to have done the opposite. Any reason for that?

[Mets Man]

Also phighten,

Would you mind posting some graphs of your research? I was very curious to see how your distribution curves worked out.

[phightin]

Originally Posted by Mets Man

I didn't mean that the Top 10 list of pitchers are relievers. That list will almost always put starters on there, very rarely will you see a reliever on that list no matter how good they are. What I meant was, I found not enough very good/elite starters were in their 30s, but there were too many very good/elite relievers that were over 35. If you sort with just starters, you'll be hard pressed to find a highly rated 34+ player. But if you sort by all relievers, you'll find plenty highly rated 34+ players.

I'm curious, if you found that there were too many old old pitchers that were too effective, why did you decide to edit the pitcher development modifier instead of the pitching aging modifier?

In most cases, whenever I feel that old players are too effective for too long, I lower the aging modifier. Whenever, I feel that young players are too effective, then I lower the development modifier. You seemed to have done the opposite. Any reason for that?

I'll try to post some graphs tomorrow. I think IRL there are a good amount of relievers that are 30+. Since guys don't pitch as many innings they can get away with longer careers. If you look at the average MLB bullpen my guess would be on most teams you'll see a good amount of older guys barring injuries etc.

The reason why I adjusted the development modifier is because (from what I understand from others here) both effect the distribution of the other. For example there should be a balance in any league depending on what the modifiers are of X number of statistics being achieved depending on the size and scope of league number of players etc. By turning up development slightly it allows for a slightly greater amount of younger pitchers to ecel which cancels out a slight amount of the older pitchers. It's like an opposite rubber band effect if you're following me. I did not want to turn up the ageing modifier because that would have simply lowered the age distribution curve and cut down on the length of pitchers careers.

I see usually at least 1/3 of my top 40 pitchers 30+ or older with the majority being 27-29 which I think matches real life. The problem is at .375 .875 while maybe you get slightly more 30+ pitchers that are elite they start to over run your league in other areas. I believe IRL the average age of an MLB starter is right around 28.3 years of age but I would have to check again.

[phightin]

Won't give you a graph Mets Fan but I'll throw some hard data at you to hopefully show what I'm trying to emphasize.

Using MLB.com stats to simply look at the top 50 pitchers in the MLB last year based on ERA and qualifying (all have minimum 160 innings pitched).

Ages 2011 Top 50 MLB Starting Pitchers
Major League Baseball Stats | MLB.com: Stats
Average Age = 28.72

# of pitchers aged 30+ = 18

# of pitchers aged 33+ = 10

# of pichers aged 35+ = 6

As you can see there really are not that many guys 30+ and into their mid thirtys that are "top pitchers" IRL. The average age of a top pitcher in his prime last season was right around 28. and while certain pitchers can pitch well once they get to 33 and over in age there are very very few who actually can. Guys like Halladay, Lee, Hudson, Dickey, etc are all rare exceptions.

Interesting enough when looking at the same spectrum of data for my Modern Day MLB League in 2036 (.375 ageing, .900 Dev) Average age of the top 50 pitchers by ERA the average age is actually slightly HIGHER!

Average age = 29.26 so slightly higher

# of pitchers aged 30+ = 22 slightly higher

# of pitchers aged 33+ = 10 spot on

# of pitchers aged 35+ = 4 slightly lower

So as you can see all the age distributions for the top pitchers in the league using this setting, VORP, and I assume WAR as well will all measure out close to real life. For whatever reason, top MLB pitchers are trending younger and that's why there are a vast amount of top aces in the game.

Now with that said there most likely is a lower distribution, although I wouldn't say major but somewhat there, of older pitchers throughout the entire league. However, that's something I have to live with because as I said before if you crank down the Dev settings or age modifier guys start hangining on too long and there's too many good 33+ and 35+ top pitchers in the game. I would go through and check year by year periodically and I counted 50-55% at times pitchers aged 30+ in the top portion of my league which is off from real life. My personal preference is to take a top down approach when it comes to this stuff as I would rather get the that right than have my league dominated by older guys etc.

[GiantYankee]

Originally Posted by phightin

Won't give you a graph Mets Fan but I'll throw some hard data at you to hopefully show what I'm trying to emphasize.

Using MLB.com stats to simply look at the top 50 pitchers in the MLB last year based on ERA and qualifying (all have minimum 160 innings pitched).

Ages 2011 Top 50 MLB Starting Pitchers
Major League Baseball Stats | MLB.com: Stats
Average Age = 28.72

# of pitchers aged 30+ = 18

# of pitchers aged 33+ = 10

# of pichers aged 35+ = 6

As you can see there really are not that many guys 30+ and into their mid thirtys that are "top pitchers" IRL. The average age of a top pitcher in his prime last season was right around 28. and while certain pitchers can pitch well once they get to 33 and over in age there are very very few who actually can. Guys like Halladay, Lee, Hudson, Dickey, etc are all rare exceptions.

Interesting enough when looking at the same spectrum of data for my Modern Day MLB League in 2036 (.375 ageing, .900 Dev) Average age of the top 50 pitchers by ERA the average age is actually slightly HIGHER!

Average age = 29.26 so slightly higher

# of pitchers aged 30+ = 22 slightly higher

# of pitchers aged 33+ = 10 spot on

# of pitchers aged 35+ = 4 slightly lower

So as you can see all the age distributions for the top pitchers in the league using this setting, VORP, and I assume WAR as well will all measure out close to real life. For whatever reason, top MLB pitchers are trending younger and that's why there are a vast amount of top aces in the game.

Now with that said there most likely is a lower distribution, although I wouldn't say major but somewhat there, of older pitchers throughout the entire league. However, that's something I have to live with because as I said before if you crank down the Dev settings or age modifier guys start hangining on too long and there's too many good 33+ and 35+ top pitchers in the game. I would go through and check year by year periodically and I counted 50-55% at times pitchers aged 30+ in the top portion of my league which is off from real life. My personal preference is to take a top down approach when it comes to this stuff as I would rather get the that right than have my league dominated by older guys etc.

Thanks for your great work and analysis! I will be your numbers in my leagues.

[robc]

If we play in a different era we would need different aging and development modifiers correct? I imagine the age distribution has fluctuated during the years?

[phightin]

Originally Posted by robc

If we play in a different era we would need different aging and development modifiers correct? I imagine the age distribution has fluctuated during the years?

Yea different eras would have different career spans, primes, and distributions of players. TBH I'm not quite familar with how historical leagues work as I don't usually play them. My whole thing is taking an MLB modern day universe and keeping the same standards over the years as it progresses. I have no interest in my league fluctuating away from the current standards in the game etc.

The more I was thinking about it today one of the reasons we proably see more young starters in the game today is the rush to use some of these top prospects before they get unnecessary innings in the minors and risk injury etc. Ten years ago top pitching draft picks out of high school would usually spend at least a few years in the minors. Now and the Giants are probably the perfect example although several other teams do it too, they call these guys up after 2-3 years.