Thoughts on development and aging after a 200 year sim

[greenOak]

After the latest patch released I started a new standard MLB game and simmed 200 years into the future (all settings are standard except league evolution which was turned off, scouting accuracy which I set to 100%, and all ratings are on the 20-80 scale). Here are my thoughts:

PROSPECT DEVELOPMENT

Prospect development is better but is still slower than IRL. Currently in 2224, there are 3 21 year olds on the opening day roster (all under 30 OVR), 15 22 year olds (only 4 are 40 OVR+), and 25 23 year olds most of whom actually appear ready for the MLB. The number of 21-22 year olds seems reasonable, but the vast majority of them do not seem MLB ready. Excluding relievers who seem to have naturally high ratings, there isn't a single 21 year old rated over 30, and only eight 22 year olds are rated 40+.

Looking at the HOF, the vast majority of game generated hall of famers broke into the league at 21-24, with a couple being older and a couple getting their start playing ~90 MLB games at age 20. I didn't see one guy who played a significant amount of time at 19. Obviously teenagers are rare in the MLB, but I would've expected to see a few. IRL, hall of famers show a much wider spread in terms of when they entered the league. You'll see guys who broke into the big leagues as teenagers and guys who didn't make the MLB until their mid to late 20's.

PLAYER AGING

With the default settings, players don't really have a chance at breaking a lot of records because they don't get to play enough games in their careers. Among non generated players, the average age at retirement of the top 10 batters in WAR is 41.2 years old (Mantle is the only guy under 40), and for the top 10 pitchers in WAR it's 42.8 years (Mathewson is the only guy under 40). Among generated players, those averages are and 38.9 and 40.5 respectively.

2686 was the most games played by any generated batter of which ranks 38th all time. Only seven generated batters were able to break the 2500 game mark, as opposed to the 68 non generated players. This (along with the current MLB environment), leads to things like not a single generated player reaching 3000 hits (top 3 are 2840, 2835, 2803), 2000 RBI (top 3 are 1969, 1792, 1777), or 1800 runs (top 3 are 1755, 1673, 1664).

FINAL THOUGHTS

Obviously some of these discrepancies can be remedied by mucking around with the development and aging settings, but in limited testing, slowing aging and/or speeding up development significantly increases the talent level in the league. Players may improve their ratings faster, but it's unlikely they'll reach the majors anytime sooner since the requisite level for playing in the MLB will increase. The default settings (in my opinion) do a very good job of maintaining a similar distribution of talent to the initial settings and fiddling with the messing around with development / aging settings messes this up. Which settings you use really depends on what is more important to you.

One of the core issues which I hinted at earlier, is that there is far less variance in aging curves in OOTP than in real life. OOTP doesn't really ever produce players who enter the league at 20/21 only to never to get better (eg. Delmon Young), or players who don't really figure things out until their 30's (eg. Jose Bautista). Instead every player has a fairly typical aging curve which for an average hall of famer looks something like break into the league at 22-23, peak at 26-28, remaining highly productive (if healthy) into their early 30's before not being good enough to play sometime in their late 30's.

There are plenty of other interesting things I noticed in this long term sim, but I don't want to clutter this thread with a bunch of largely unrelated observations.

[liberty-ca]

Originally Posted by greenOak

There are plenty of other interesting things I noticed in this long term sim, but I don't want to clutter this thread with a bunch of largely unrelated observations.

It would be interesting to know what have noticed. Will you share your other observations, perhaps in a different thread?

[Markus Heinsohn]

One thing to keep in mind, one sim is a very small sample size. What we typically do is create a fully fictional MLB setup, then sim for 50 years, and then look at various metrics. In our tests, the latest update looks very good overall, except the following weaknesses, which we will still work on of course:
- We need a few more star prospects who develop really, really fast so they make a real impact in MLB at age 20, 21, 22.
- A few more standout players in general wouldn't hurt, sometimes it feels the number of true mega-stars is a bit low. But this varies from sim to sim.

One of the core issues which I hinted at earlier, is that there is far less variance in aging curves in OOTP than in real life.

I disagree here. The question is, what do you look at, raw ratings or stats? We look at stats output, and here the aging curves show a nice variety and in general feels close to real life.

Oh, and in general, let's keep in mind we're still in March, and as I am typing this, MLB didn't even have its proper Opening Day. Plenty of time to continue to work on this part of the game Overall, this is the best launch we ever had.

[Paul Reuschel's Mustache]

Originally Posted by greenOak

PROSPECT DEVELOPMENT

The number of 21-22 year olds seems reasonable, but the vast majority of them do not seem MLB ready. [...] there isn't a single 21 year old rated over 30, and only eight 22 year olds are rated 40+.

PLAYER AGING

With the default settings, players don't really have a chance at breaking a lot of records because they don't get to play enough games in their careers. [...] 2686 was the most games played by any generated batter of which ranks 38th all time.

FINAL THOUGHTS

[...] but in limited testing, slowing aging and/or speeding up development significantly increases the talent level in the league. Players may improve their ratings faster, but it's unlikely they'll reach the majors anytime sooner since the requisite level for playing in the MLB will increase.

Thank you for this excellent feedback!

I will soon be starting a long-term MLB (plus international leagues) save. More variance in aging curves and career development would be welcome. I am considering changing TCR from the default 100 to 150 or even 200 to generate more unexpected outcomes.

Originally Posted by greenOak

There are plenty of other interesting things I noticed in this long term sim, but I don't want to clutter this thread with a bunch of largely unrelated observations.

Please share your findings and thoughts. Much appreciated!

[greenOak]

Originally Posted by Markus Heinsohn

One thing to keep in mind, one sim is a very small sample size. What we typically do is create a fully fictional MLB setup, then sim for 50 years, and then look at various metrics. In our tests, the latest update looks very good overall, except the following weaknesses, which we will still work on of course:
- We need a few more star prospects who develop really, really fast so they make a real impact in MLB at age 20, 21, 22.
- A few more standout players in general wouldn't hurt, sometimes it feels the number of true mega-stars is a bit low. But this varies from sim to sim.


I disagree here. The question is, what do you look at, raw ratings or stats? We look at stats output, and here the aging curves show a nice variety and in general feels close to real life.

Oh, and in general, let's keep in mind we're still in March, and as I am typing this, MLB didn't even have its proper Opening Day. Plenty of time to continue to work on this part of the game Overall, this is the best launch we ever had.

I looked at ratings rather than stats so maybe things look better if you look at stats.

Also, I don't get why the sample of players produced in single 200 year sim should be any different than the sample of players produced by doing four separate 50 year sims.

[Paul Reuschel's Mustache]

Originally Posted by Markus Heinsohn

Overall, this is the best launch we ever had.

Markus, you and the Team rock! I'm still relatively new to OOTP, but the level of direct interaction I'm seeing here between the developers and the players is exceptional.

[Lukas Berger]

One thing to keep in mind with development is that the default settings need to more or less work with each different league type and in different eras.

So imo it makes sense to be more conservative here, as things can always be adjusted one way or another using the dev and aging settings.

[Markus Heinsohn]

Originally Posted by Paul Reuschel's Mustache

Markus, you and the Team rock! I'm still relatively new to OOTP, but the level of direct interaction I'm seeing here between the developers and the players is exceptional.

Thanks for the kind words.

[gentlemanofleisure]

Originally Posted by Markus Heinsohn

One thing to keep in mind, one sim is a very small sample size. What we typically do is create a fully fictional MLB setup, then sim for 50 years, and then look at various metrics. In our tests, the latest update looks very good overall, except the following weaknesses, which we will still work on of course:
- We need a few more star prospects who develop really, really fast so they make a real impact in MLB at age 20, 21, 22.
- A few more standout players in general wouldn't hurt, sometimes it feels the number of true mega-stars is a bit low. But this varies from sim to sim.


I disagree here. The question is, what do you look at, raw ratings or stats? We look at stats output, and here the aging curves show a nice variety and in general feels close to real life.

Oh, and in general, let's keep in mind we're still in March, and as I am typing this, MLB didn't even have its proper Opening Day. Plenty of time to continue to work on this part of the game Overall, this is the best launch we ever had.

I agree. I did a 30 year sim last night with a Fictional league with Historical rookies from all eras being drafted (1960-1989). There were only 4 or 5 true generational talents, and 3 of them were international discoveries. The only 2 historical All Time Greats were an old time pitcher named Foghorn Bradley and Dave Parker. There were a lot of 3 or 4 year wonders, guys that were drafted as 4 or 5 star players, had incredible seasons at 22,23,24 years old, and then dropped like a stone. Ken Griffey Jr. , Chuck Klein, and a whole bunch of starting pitchers looked to be emerging All time greats and then were average by 27 years old. It is like after they hit 26, they become 3 star players at best.

I think just a minor tweak of historical rookies being drafted to keep their mojo beyond two or three years would do wonders for the game.

Also, every draft (32 team league), the first starting pitcher off the board is in the late 1st/ early 2nd round. Every single draft. 30 drafts total. Even guys like Walter Johnson that come into the game as 5 star players (5 star potential) dont get picked until the 1st or 2nd round. Not surprisingly, Johnson went from 5 stars to 2 and 1/2 by the age of 24, and was a waiver pick up at 25 years old. Out of the game entirely by 28. Sure, not every real life "ALL TIME GREAT" will become great in a fictional league, which is the Awesome part of OOTP, but some of these guys should. Not just Walter Johnson, but Noodles Hahn, Roy Oswalt, Carl Hubble, Robin Roberts. They all gets drafted early 2nd round as 5 star potential players and by 25 years old are 2 and 1/2 stars. Only guys that began late in real life like Teddy Higuera stay top of the rotation starters into their 30's.

[Lukas Berger]

I think it's important to realize that there is not one magic development and aging setting that will be able to fully recreate the entirety of baseball history in a given long-term sim.

If you look at baseball history, there were probably something like 6-7, if not more, fully different eras where aging and development were radically different.

The default setting can't have both dozens of batters playing to their early to mid-40's with no diminishment of their skills like in the steroid era, and an unusual amount of batters peaking and fully developed in their early 20's while mostly being toast by their mid-30's like the modern era.

Or SP's pitching long enough to get 300 to 350 wins like those in the 1970's and earlier, but also have guys not able to even get 200 for the most part like modern day.

To do this, you'd really need to frequently switch development and financial and other settings to recreate specific eras.

Basically the default settings are trying to get an environment that's relatively neutral, but shaded toward modern day. I think they're doing a pretty good job of that right now.

That's not to say we won't still make some tweaks here. There are definitely a few things still on our lists to adjust in the next few patches, especially in regard to outliers (creating more of them).

[Lukas Berger]

Originally Posted by gentlemanofleisure

Also, every draft (32 team league), the first starting pitcher off the board is in the late 1st/ early 2nd round.

This is two entirely different things, not really related to development. One, the AI is somewhat too much favoring hitters early in drafts now. This more reflects some modern trends of ranking hitters higher on prospect lists etc, but the effect is a bit too strong right now, and are planning to to make some adjustments here to get the AI to value pitchers more.

The second thing is that in some game modes historical players, especially pitchers, are importing with somewhat low potential ratings. This is not an issue in regular historical play where you use recalc, but in a historical gameplay where you use development, it does mean that pitchers will be somewhat undervalued. Also something we're working on adjusting.

[gentlemanofleisure]

Originally Posted by Lukas Berger

This is two entirely different things, not really related to development. One, the AI is somewhat too much favoring hitters early in drafts now. This more reflects some modern trends of ranking hitters higher on prospect lists etc, but the effect is a bit too strong right now, and are planning to to make some adjustments here to get the AI to value pitchers more.

The second thing is that in some game modes historical players, especially pitchers, are importing with somewhat low potential ratings. This is not an issue in regular historical play where you use recalc, but in a historical gameplay where you use development, it does mean that pitchers will be somewhat undervalued. Also something we're working on adjusting.

It is funny because I looked at the pitchers being undervalued in the draft as maybe a trend more than an OOTP issue that needs fixing. Similar to the NFL no longer valuing running backs in the first round of the NFL draft. Maybe the AI controlled teams are programmed to devalue starting pitchers as 1st round picks because they develop so randomly and get injured so easily.

But still, if you have a generational talent like Walter Johnson with a 5 star potential and he throws 98 MPH, he should be a top pick. There should be some exceptions to the rule.

Either way, love the game and look forward to playing it every weekend when I have free time!

[Lukas Berger]

Originally Posted by gentlemanofleisure

It is funny because I looked at the pitchers being undervalued in the draft as maybe a trend more than an OOTP issue that needs fixing. Similar to the NFL no longer valuing running backs in the first round of the NFL draft. Maybe the AI controlled teams are programmed to devalue starting pitchers.

But still, if you have a generational talent like Walter Johnson with a 5 star potential and he throws 98 MPH, he should be a top pick. There should be some exceptions to the rule.

Yeah, that's definitely more the modern trend and kind of what we are trying to recreate here, but the effect is probably a bit too strong now. Agreed on a generation guy like Johnson being the exception.

Originally Posted by gentlemanofleisure

Either way, love the game and look forward to playing it every weekend when I have free time!

Thanks for the kind words, it's always really nice to hear this!

[Syd Thrift]

Yeah, I personally have both aging and development flattened out a lot and then TCR goosed way up (I think I have it at 150 now) in my league because I think that approximates real life better, especially with ratings off (that people get better and worse all the time for non-age reasons and then we tend to assign reasons post hoc), but I also think the average person doesn’t like that level of randomness (see also the draft, which in OOTP has been a lot more like the NFL draft than the real life MLB one for a while) and so we get aging curves instead.

[SirMichaelJordan]

Originally Posted by greenOak

I looked at ratings rather than stats so maybe things look better if you look at stats.

Also, I don't get why the sample of players produced in single 200 year sim should be any different than the sample of players produced by doing four separate 50 year sims.

A 50-year simulation on 4 different saves is more reliable than a 200-year simulation on 1 save because it allows for more variance and provides more informed sampling. Your test is based on probability, so a 50-year simulation on 4 different saves is more reliable since its more random.

[uruguru]

Originally Posted by Syd Thrift

Yeah, I personally have both aging and development flattened out a lot and then TCR goosed way up (I think I have it at 150 now) in my league because I think that approximates real life better, especially with ratings off (that people get better and worse all the time for non-age reasons and then we tend to assign reasons post hoc), but I also think the average person doesn’t like that level of randomness (see also the draft, which in OOTP has been a lot more like the NFL draft than the real life MLB one for a while) and so we get aging curves instead.

I agree with everything you said in this post, however I don't understand what you mean by aging and development "flattened out".


Can you elaborate on your settings for that, and why?

edit: one other "realism" blurb I do is that, for every rookie class, I find all of the knuckleball pitchers and set their propensity to get arm injuries as low as possible. OOTP has them almost just right!

[Syd Thrift]

Aging is set below 1, development is above 1. I think I’ve got them somewhere around 0.75 and 1.25 respectively.

[uruguru]

Originally Posted by Syd Thrift

Aging is set below 1, development is above 1. I think I’ve got them somewhere around 0.75 and 1.25 respectively.

Yeah, I think with a high TCR it's a good idea to do both of those things. Thanks for clarifying!

[Syd Thrift]

Originally Posted by uruguru

Yeah, I think with a high TCR it's a good idea to do both of those things. Thanks for clarifying!

Yeah, the flip side of this is that I think you wind up with a lot more guys between, say, 23 and 35 who are all major-league / AAAA quality and I don't think the AI does such a great job of cutting guys for purely age-related reasons. IRL if you have two guys with similar skillsets and one is 33 and one is 27, even, you stick with the 27 year old because they're more likely to stay at that level for the next, say, 5 years. In the minors it's waaaaay more brutal than that - like 24 year olds who are better than 19 year olds get cut loose from A or even AA all the time - but again, the AI is way too "nice" to these guys.

Actually, come to think of it, one thing I do to mitigate that is I put age limits on all of my minor leagues below AAA. You do still end up with 30+ year olds in AAA (which, I play out all my league's games from the 7th on and also handle major transactions but I do allow the AI to sign/promote/demote/cut minor leaguers so I do see this effect) but I don't know, to a great extent this is as it should be. I also have, currently, two international leagues in Mexico and redactedland to pull some of those AAAA types off the top. I've definitely seen those leagues sign guys in their mid to late 30s who'd been cut from their MLB team so that seems to be working as it ought to be.

[uruguru]

Originally Posted by Syd Thrift

Actually, come to think of it, one thing I do to mitigate that is I put age limits on all of my minor leagues below AAA.

I do the same. These are my max ages:

BC - 22
A - 26
AA - 30
AAA - no limit

I don't really have a good feel for what proper limits would be, though. Just that someone over 22 shouldn't be in BC and anyone over 30 should be AAA or bust.