Feedback Requested: Tweaked Settings for Aging Curves

[JinAZ]

Originally Posted by Doughnuts?

Someone should sticky this mofo.

Seconded.

[Seviien]

If we do get a sticky, i ask that the "Feedback Requested" be dropped from the thread title. Not that I don't want feedback; I just think a changed title would make this look more like a finished product and less like a work in progress.

[endgame]

Great follow up and thanks for providing the new data. I hope this nears my final inquiry. If I elect NOT to adjust league totals yearly and/or import PCM for my fictional leagues, yet still utilize the majority (I'm not assuming I will mimic the minors exactly) of your modifications, would I be correct in thinking the likely result will be twofold: that the range of performance - both in the areas of incoming talent and field demonstrated results - would widen a flux over time AND ultimately impact the aging curve although not as drastically as having not used them at all?

Again, thanks for all of your efforts and your prompt responses.

[Seviien]

If you do not auto adjust PCM and league totals, then I would assume you wouldn't see much of a difference. Since the PCM and league totals changed over time but my curves largely didn't, I feel safe saying the aging modifiers are largely independant of PCM and league totals.

And you needn't copy the minors exactly, but I think there's a lot more potential to change the curve shapes with the minors and their MLEs than with the PCM and league totals. Just think your changes through and you should be all set.

[Curtis]

After all this hard work I hesitate to ask this Seviien, but with your modifiers how long is the development process? I'm looking for settings that will develop players to their potential in around four years, with six as kind of an upper level cutoff. (I mean with the first standard deviation falling within the fourth year, the second within years three and five, and the third within years two and six.) Is this possible? Do you have any idea how?

I'd also like to be able to set player aging so that the 'peak' performance years are roughly from 29 to 31. Is that possible?

I guess I should've added that I use all five tiers of minors, and have 'early' and 'late' seasons for Rookie and Short A (ending before and starting after the amateur draft). Early season players from both leagues get consolidated onto the 'late' Short A roster, and the new draftees go to the 'late' Rookie roster. (I have no idea if any of that would make a difference in your answer.)

[sfgiants88]

Sev, my main concern if I adjust the aging modifers is that I want accurate numbers of 32+ year olds and also accurate numbers of players younger than 26.

Your aging modifers seemed to have done an excellent job of flattening the peak, but as a result, the figures from 32+ were a bit low.

Since you are much more familiar with the aging modifiers than I am, do you have a ballpark figure of the aging modifiers that I could use as a start (and then tweak to my desire?) Again, my main concern is the number of mid-to late thirties players.

Thanks for all of your hard work!

[Seviien]

Curtis,
Taking a rigorous look at development rates is a bit too cumbersome. Keep in mind, though, that the AB and IP curves track MLB data pretty well through the early 20s. It would be difficult for that to occur if development rates were wildly different than in real life.

Annecdotally, while I was testing I did take a look at a number of players to see how long they had spent in the minors. I was pleasantly surprised. Most players took about 3 years, but there were a few "projects" that took about 5-6 years. If there is a weakness, I believe it's that the exceptionally strong players only spend a season or so in the minors. Fortunately, that's not too difficult to find plausible. Just look at Miguel Cabrera and Andruw Jones.

With respect to "peak" years... that's a bit subjective. What I can say is that in the test I ran from 1901 to 1979, the average age of batters with the highest VORP (80th to 95th percentile bracket) was 29.01. I think that would put the peak years in your 29-31 range.



Giants,
The accuracy of the curves are somewhat limited by the dynamics of the game. If I could have gotten curves that yeilded a more accurate representation of players in their mid to late 30's, I would have. The problem is that achieving better fit for the older players distorts fit for the younger ones. I don't think that there's any issue with the batters, but yeah, I'd like to see some older pitchers in there. I just couldn't get them without causing an unrealisticly small number of players in their early 20s.

The batting curve is pretty accurate until around 35, but keep in mind that only 8% of the ABs in MLB were made by players older than 35.

The IP curve begins to break from reality at about 33. Pitchers 33 and older log ~20% of the innings in MLB, so that is more of an issue. Still... it's not at all bad.

If you're looking for a good starting point, I'd say the settings as published are a good spot. There are 23 tests and ~30 hours behind them To tweak the pitching curve more to your liking, I would suggest first lowering the aging modifier to 0.001 or 0.003. Look at the results. If you're pitchers are too old, bump the dev modifier up by 0.050. If instead the pitchers are still to young after adjusting the aging modifier, then take the development modifier down by 0.100.

[Doughnuts?]

Hate to keep beating that dead equine, but why oh why was this not tested and implemented properly before release or even in one of the patches? To have to lower the pitcher aging modifier to .001 to get it to approximate reality is really messed up.

[Lee]

Originally Posted by Doughnuts?

Hate to keep beating that dead equine, but why oh why was this not tested and implemented properly before release or even in one of the patches? To have to lower the pitcher aging modifier to .001 to get it to approximate reality is really messed up.

Hopefully this will be looked at when Markus gets back. We should have a combination of accurate modifiers and solid AI at close to the default settings.

On that thought, I think the AI is the bigger issue here, as it is responsible for placing Pitchers in the bullpen far too frequently once they hit 28-30, and their talent drops just a bit. If this was tweaked to not favor younger Pitchers so much, we would probably not even be having these discussions on playing around with the aging modifiers so much.

BTW, great job Seviien, I have put around 30 hours into testing/playing with this as well with different settings and I know it can get frustrating.

[Carplos]

On that thought, I think the AI is the bigger issue here, as it is responsible for placing Pitchers in the bullpen far too frequently once they hit 28-30, and their talent drops just a bit. If this was tweaked to not favor younger Pitchers so much, we would probably not even be having these discussions on playing around with the aging modifiers so much.

Except that if the aging numbers were better, the AI wouldn't make those silly decisions because the 'older' pitchers would still be good enough to start and not heading the wrong way with ratings as fast.

There's a balance in there somewhere.

[endgame]

Originally Posted by Doughnuts?

Hate to keep beating that dead equine, but why oh why was this not tested and implemented properly before release or even in one of the patches? To have to lower the pitcher aging modifier to .001 to get it to approximate reality is really messed up.

Well, not to stand the horse upright, but how do uncover an answer to that question? And to what end if you did? That fact is, there are people 'testing' it now (or continuing their previous looks) and reporting findings that will, hopefully, be reviewed on Markus' return. Albeit, the people who are involved are skilled at data collection and the analysis of the same, it wouldn't hurt anything for others to do their own examinations and render judgments based on personal experience rather than assuming the easiest position on the rail of the current bandwagon.

If it needs 'fixed', I assume a fix will be put into place. If it needs modifying then work is already underway to provide that modification for those who desire to utilize it.

[His Own Bad Self]

Awesome work, Sev. I do have one comment/question. The real-life aging curve seems to be stretched out further to the right than it used to be - players are playing longer and later than they used to, thanks to modern medicine. Do you have any idea how the default and modded aging curves would compare to earlier eras? Can you research that? Maybe even the modded curves don't quite fit the modern era, but perhaps they fit earlier ones...

[Can of Corn]

this deserves to be bumped again

[Doughnuts?]

Originally Posted by endgame

Well, not to stand the horse upright, but how do uncover an answer to that question? And to what end if you did? That fact is, there are people 'testing' it now (or continuing their previous looks) and reporting findings that will, hopefully, be reviewed on Markus' return. Albeit, the people who are involved are skilled at data collection and the analysis of the same, it wouldn't hurt anything for others to do their own examinations and render judgments based on personal experience rather than assuming the easiest position on the rail of the current bandwagon.

If it needs 'fixed', I assume a fix will be put into place. If it needs modifying then work is already underway to provide that modification for those who desire to utilize it.

I am certainly grateful to the skilled data collection folks who put in all of the hours that the rest of us poor shnooks with kids and jobs that take up too much of our time are unable to put into the game. The point is, this is a very good game. But I shouldn't have to be doing the developers job 2 months after release on something so important.

[bp_]

Originally Posted by Can of Corn

this deserves to be bumped again

needs to be stickied

[Seviien]

Originally Posted by jdawgsenior

Awesome work, Sev. I do have one comment/question. The real-life aging curve seems to be stretched out further to the right than it used to be - players are playing longer and later than they used to, thanks to modern medicine. Do you have any idea how the default and modded aging curves would compare to earlier eras? Can you research that? Maybe even the modded curves don't quite fit the modern era, but perhaps they fit earlier ones...

The curves have actually changed very little. I looked at data from as far back as the 1960s, and while there certainly are more players playing into their mid 30s today, it's not enough to qualify as a "curve shift." Off the top of my head, it's not larger than players 33+ constituting more than a few percentage points of the distribution.

[Seviien]

Originally Posted by Doughnuts?

I am certainly grateful to the skilled data collection folks who put in all of the hours that the rest of us poor shnooks with kids and jobs that take up too much of our time are unable to put into the game. The point is, this is a very good game. But I shouldn't have to be doing the developers job 2 months after release on something so important.

See... if *you* were the one who did the work, I could understand your concern I'm 100% fine with where this game is going. Yes, I would appreciate it if development and aging were revisited, mostly because after playing through my first game I think these modiiers yeild a few too many "phenoms" and the AI seems to regard anyone older than 28 as over the hill. And after all the work I put in, I think this is about as much improvement as can be achieved through the modifiers.

The thing is, it's not worth saying the game is broken or insinuating that this part of the game was done wrong. Like I said in the top post of this thread, the factors that affect aging and development are many. Whether people realize it or not, at least some of what people experience with respect to aging and development comes as a result of how they setup their universe.

[Markmeister]

Seviien - I've tried your modifiers on a different league setup. I may need to change the modifiers simply because the league structure is different (3 major leagues w/ minors, 2 independent minor leagues and 2 foreign leagues).

Aging seems to look right at a glance, but I am experiencing the phenom problem you alluded to in your post, though.

1910 through 1920 looks weird with several players hitting 20+ home runs. 1921 through 1930 does develop some sluggers hitting 50+ home runs, but I need to do something to tone back batting averages - league leaders are consistently hitting over .430 with 3 to 4 players reaching this plateau per year.

You're probably right that improvements need to be made to the AI front to get better age distribution. OOTP probably needs to look at the comparison of established players versus rookies differently. Rookies seem to be getting the nod over the established players once the rookies' ratings are better, but maybe not enough consideration is being given to the established player's prior year(s) stats. Make it more like real life baseball; when a starter who went 18-10 the prior year loses a couple miles of hour on his fast ball, a real-life manager wouldn't immediately replace him in the lineup with a 24 year old rookie. The real-life manager would at least give the established starter some starts before he decides to move him to the bullpen. Once the stats start dropping, then the decision to replace him with the rookie should be made.

[His Own Bad Self]

Originally Posted by Seviien

The curves have actually changed very little. I looked at data from as far back as the 1960s, and while there certainly are more players playing into their mid 30s today, it's not enough to qualify as a "curve shift." Off the top of my head, it's not larger than players 33+ constituting more than a few percentage points of the distribution.

OK. I was under the impression that the supposed curve shift was more pronounced. But if not, very well then.

[Doughnuts?]

Originally Posted by Seviien

See... if *you* were the one who did the work, I could understand your concern I'm 100% fine with where this game is going. Yes, I would appreciate it if development and aging were revisited, mostly because after playing through my first game I think these modiiers yeild a few too many "phenoms" and the AI seems to regard anyone older than 28 as over the hill. And after all the work I put in, I think this is about as much improvement as can be achieved through the modifiers.

The thing is, it's not worth saying the game is broken or insinuating that this part of the game was done wrong. Like I said in the top post of this thread, the factors that affect aging and development are many. Whether people realize it or not, at least some of what people experience with respect to aging and development comes as a result of how they setup their universe.

My statement was only in response to Endgame's suggestion that we all should be testing and providing feedback. I wouldn't presume to comment on your behalf.