Right aging and random talent modifiers for modern day MLB League

[Cras]

And I am reading the modifiers right, in that a higher number will speed up those processes, and lowering the number slows them down?

[phightin]

Just got done another 20 year sim using .900 as the aging modifier and leaving dev at default 1.00 I must say I have seen some improvements.

The average age of the top 105 hitters is 27.86 which I can live with. There were also 34 total hitters in the top 105 that were 30+ in age. Some were even in their late 30s as well. Still though there were only 6 hitters aged 33 or older in the top 105 so that does seem a bit low. There also was conversely 25 hitters in the top 105 25 and under in age which was a nice reasonable amount to see.

The overall percentage of 30+ players was 32% close to the minimum but between that and the low number of 33 and above players I think I may try another go at it and lower the aging modifier again to around 0.85. The only thing I'm worried about is that it will have an impact on the 25 and under guys coming through. There needs to be a balance between all three age groups.

[Ben E Lou]

I don't know how to do those fancy-dancy charts that you people do, but my spidey sense tells me that I may have stumbled upon something pretty good. I'd love for someone to try this against the charts:

RATINGS: 20%
CURRENT YEAR: 40%
PREVIOUS YEAR: 40%
2 YEARS AGO: 0%

Sabermetric Lineups

BATTER AGING: .500
BATTER DEV: .875
PITCHER AGING: .750
PITCHER DEV: .750

INJURIES: OOTP Classic

I've run a test career with these settings for ~200 years, giving me a chance to look over a lot of players who flamed out in their late 20s, and guys who lasted to their 40s. It "seems" like a good balance.

On Opening Day 2206, average age of batters in the majors is 28.8. Average age of pitchers is 29.3.

[Ben E Lou]

Actually, I'll sim 2206 and output the csvs. Maybe one of y'all can give soem feedback from those.

[Ben E Lou]

Heh...if I knew which CSVs contained the data needed, that would be better.

[Ben E Lou]

OK. I think I've figured out the correct csvs for batting, and narrowed them down to the last ~20 years. Here they are:

http://www.younglifenorthdekalb.com/...sting_csvs.zip

[henry296]

Skydog,

A couple of questions about your league setup:

1. How many levels of minors?
2. How many draft rounds?
3. How many batters qualfiied for batting title (3.1 PA)

[Ben E Lou]

Originally Posted by henry296

Skydog,

A couple of questions about your league setup:

1. How many levels of minors?
2. How many draft rounds?
3. How many batters qualfiied for batting title (3.1 PA)

1. Standard Minors for Fictional (AAA,AA,A,SS-A,R)
2. 20 round draft (players generated for 25 rounds)
3. 120 players (out of 24 teams) qualified in 2206. From 2187-2206 (20 seasons), an average of 117.2 qualified.

[Eiskrap]

Originally Posted by SkyDog

Actually, I'll sim 2206 and output the csvs. Maybe one of y'all can give soem feedback from those.

Using my template I get the following, where the red represents MLB data from 07-10 from the Lahman, and the blue is the data from OOTP.





To me it looks a lot better than default data, however there is still a bit too much of a decline in skills between 31-34 for me.

Its interesting in the Lahman data the BA actually improves in ages 30-35. I'm assuming this is down to only good, proven 30-35 year old players actually play. It appears that the players are getting better when in fact you just don't have the lower quality players in these age ranges. Where as if you have a 25 year old hitting .250 you still throw him out there hoping he'll improve, and he is cheap.

I wouldn't expect to see OOTP mirror this completely, however I still think it drops off more than it should.

[henry296]

Originally Posted by SkyDog

1. Standard Minors for Fictional (AAA,AA,A,SS-A,R)
2. 20 round draft (players generated for 25 rounds)
3. 120 players (out of 24 teams) qualified in 2206. From 2187-2206 (20 seasons), an average of 117.2 qualified.

Thanks.

The qualifiers looks really good. If you gross it up to 30 teams it is the same as MLB in 2011. I assume that is average fatigue?

I think with more rounds, you can get by with higher injury settings because you have less risk when minor league prospects get hurt.

[SunDevil]

Originally Posted by SkyDog

I don't know how to do those fancy-dancy charts that you people do, but my spidey sense tells me that I may have stumbled upon something pretty good. I'd love for someone to try this against the charts:

RATINGS: 20%
CURRENT YEAR: 40%
PREVIOUS YEAR: 40%
2 YEARS AGO: 0%

Sabermetric Lineups

BATTER AGING: .500
BATTER DEV: .875
PITCHER AGING: .750
PITCHER DEV: .750

INJURIES: OOTP Classic

I've run a test career with these settings for ~200 years, giving me a chance to look over a lot of players who flamed out in their late 20s, and guys who lasted to their 40s. It "seems" like a good balance.

On Opening Day 2206, average age of batters in the majors is 28.8. Average age of pitchers is 29.3.

SkyDog,

Thank you for running this test. Could you also list out the Target Change Randomness value that you used?

Thanks.

[Ben E Lou]

Originally Posted by sundevil

skydog,

thank you for running this test. Could you also list out the target change randomness value that you used?

Thanks.

100

[Ben E Lou]

Originally Posted by Eiskrap

Using my template I get the following, where the red represents MLB data from 07-10 from the Lahman, and the blue is the data from OOTP.





To me it looks a lot better than default data, however there is still a bit too much of a decline in skills between 31-34 for me.

Its interesting in the Lahman data the BA actually improves in ages 30-35. I'm assuming this is down to only good, proven 30-35 year old players actually play. It appears that the players are getting better when in fact you just don't have the lower quality players in these age ranges. Where as if you have a 25 year old hitting .250 you still throw him out there hoping he'll improve, and he is cheap.

I wouldn't expect to see OOTP mirror this completely, however I still think it drops off more than it should.

Thanks. I'm doing this for a v13 release of my NABA setup. I know I can't get it perfect, but I might try to improve it just a hair more before release. I'm going to run one more sim with a few more changes. If I get data to you around 9-10pm EDT, would you be able to evaluate it again, or should I just let it run overnight and post it in the morning?

[Ben E Lou]

Originally Posted by henry296

I assume that is average fatigue?

Yes.

[Eiskrap]

Originally Posted by SkyDog

Thanks. I'm doing this for a v13 release of my NABA setup. I know I can't get it perfect, but I might try to improve it just a hair more before release. I'm going to run one more sim with a few more changes. If I get data to you around 9-10pm EDT, would you be able to evaluate it again, or should I just let it run overnight and post it in the morning?

Leave it until the morning, I'm in the UK so won't be able to look at it until the morning.

[kehowey]

Originally Posted by Cras

And I am reading the modifiers right, in that a higher number will speed up those processes, and lowering the number slows them down?

yes

[Ben E Lou]

OK. Last check, Eiskrap.

http://www.younglifenorthdekalb.com/ootp/last_check.zip


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

On Opening Day 2171, average age of players in the majors is 28.6.

I went to a much heavier ratings-based evaluation because I noticed that the AI teams were leaving solid players in their mid to late 20s sitting in FA all year, presumably because they had never cracked the starting lineup and therefore had no stats. I'm talking guys rated 60-50-60 who were solid fielders, but had either been in the minors or blocked by a star in the majors, so they didn't have stats. And really, this is the way that I prefer the AI to play when I'm actually simming out a career. Given that *I* am gonna sit a guy who had a .850 OPS last year but dropped 20 points in all three categories in the offseason, I want the AI to have that same advantage. I can sacrifice some realism there.

[Eiskrap]

Updated graphs as below:







In terms of distribution of PAs, I don't think we are going to see any more improvement...

[luger]

Wow, nice work. I had been running somewhat similar settings and getting close but I think bumping up the ratings in the AI Evaluation may be the key. Bravo!

[Ben E Lou]

Originally Posted by Eiskrap

Updated graphs as below:







In terms of distribution of PAs, I don't think we are going to see any more improvement...

Sweet. That's good enough for me. Appreciate the evaluation.