Player development in OOTP 9.2.2

[Childress]

Originally Posted by Russ

Take for instance the rate of change in pitchers' K/9 ratio. These guys may be ending their careers with realistic strikeout totals but the way they're getting there isn't realistic.

It's been years since reading the Almanac, but one of Bill James' findings was that strikeout pitchers tended to enjoy longer and more productive careers- with consistent K totals- than control specialists. Also, fast base runners held up better. It seems counter-intuitive be he backed it up.

[Russ]

Originally Posted by Childress

It's been years since reading the Almanac, but one of Bill James' findings was that strikeout pitchers tended to enjoy longer and more productive careers- with consistent K totals- than control specialists. Also, fast base runners held up better. It seems counter-intuitive be he backed it up.

Right. OOTP is way off on this.

[StyxNCa]

The way I see it is that managers do not use an old player the same way they would use that same player in his young days. You play the player in a way that makes use of his remaining strengths while avoiding his weaknesses as much as possible.

When simming a game it manages the player the exact same way forever, unless one goes through the extremely time consuming effort of going into every players strategy settings and adjusting those.

This is where human decisions can make a difference. I have players who are well past their prime but their stats haven't deteriorated in a fashion I feel is abnormal, but then I have changed the way I use them and tend to rest them more. On the other hand there are players, some young, who have great ratings but never perform up to that level. I really haven't seen this problem to the extent that some of you are saying is there. I've looked through retired players and checked their stats and I just don't see it. Yes a few dropped off badly, but not most. Some retired and I still haven't figured out why since they still looked good statistically their last year.

I think the problem is too many look at ratings and expect everyone to reach their potential. There are too many ratings, personality ratings and such, that can effect just how well they do.

I am working on a database and have seen many young players come along, have a very good year or two, then disappear. It's not as rare as some of you seem to think. There are also some who trudged along for a number of years then suddenly exploded for a good period of time (before the steroid era). Sandy Koufax wasn't much in his first 5 or 6 years then he took off.

This thing of saying that pointing out one player is not evidence. There have been almost 17,000 players come and go. Have you looked at every single one to support your argument that these things don't happen a lot more frequently than you think? I have looked at 14,500 so far and I don't see where the dropoff from players in my league is any worse than real life.

Most of this is just expectations of the person playing the game. You expect some kind of consistancy from every player and expect every player to perform exactly as someone comparable. If things were that way, just how boring would the game be?

Maybe I'm just not understanding the argument.

[mattshwink]

Originally Posted by Russ

I'm really trying to avoid this kind of discussion. If you're happy with it, I'm happy for you. I am not. To each their own. We all have a choice of what to do with our money.

But there are countless threads about MVP type seasons being followed by seasons that wouldn't even keep somebody in the big leagues without injury, studies about how the aging and development are wrong. One guy named Kevin Maas from twenty years ago does not make this 'right'.

The refusal to even acknowledge that there could be a problem with this has caused me to lose faith.

I don't think the problem is the refusal of the problem. I think there needs to be specific instances of what the problem actually is.

One of the things people have to realize is that this game is intended to simulate over 140 years of baseball history, as well as model the modern (and future) game. And then, there are the fictional leagues which may not follow any of the "established" systems.

Getting a player development model that works for all of those is not easy. So Markus stated above he took a holistic approach. That, to me, is perfectly acceptable. OOTP is so tweakable that if I don't like it, I can tweak it.

RonCo took the engine, ran some long term statistical simulations, and then analyzed the data and tweaked the settings to get 1980 era results (he wanted to make his baseball universe modern, but withholding the 90's and the "juiced" era. If you are trying to simulate another era of baseball they may not work for you. He clearly states that.

Markus also stated above that he did tweak the dropoff for established players that are not injured. So that should no longer be broken. Of course, we'll have to see some results....

[injury log]

Originally Posted by Russ

The things is, though, is that RonCo's work isn't opinion. He took hours upon hours to collect real-life data and OOTP data and compare the two and find factual differences. The differences he found aren't his biases or his interpretation. They can't be argued and they shouldn't be summarily dismissed.

Data is data, but interpretation is opinion. I won't dispute any of RonCo's data, but I have in the past disputed the conclusions drawn from that data. We had a very, very long discussion about this issue last year, that I don't want to rehash, but I articulated the issues I have with comparing Tango and OOTP data (in posts 279, 291, and others):

Does OOTP Player Development Needs to Be Fixed?


Now, I don't know which of RonCo's studies in particular you are referring to, but if you're discussing his comparisons between OOTP development and Tango curves, you really need to understand how Tango derived his curves to understand how they might be relevant to OOTP development. There are serious issues there.

[mattshwink]

Originally Posted by kq76

That's the thing, though, Markus. All methods are not as good as each other. I think the reason why so many of us trust RonCo is he clearly knows his stuff. Many of us might not be idiots when it comes to stats, but even those of us who understand the stuff most don't really know how to go about doing it and especially what not to do. The consensus is RonCo does. So if he's using one method and you're using another, I'm afraid you'll have to explain why you think yours is better for us to be on your side. Maybe he is wrong, maybe you're right, but you need to make your case like he has. So far you've just basically told us it's good enough and you don't really care to make it better. Well, for the perfectionists among us, that's not good enough.

Let's back up a second. RonCo's method was a type of regression. He knew the type of curve he wanted (modern era minus juiced). All he was trying to do was get the engine to give him the output he wanted. So he kept tweaking the settings (which is what a lot of us do) to get what he wants.

There are many people who use RonCo's suggestions, and further tweak them to what they want (so should Markus use RonCo's settings, or some consensus of tweaked RonCo settings?). So to say that RonCo knows what he is doing is certainly correct, he is definitely an expert. But to try to compare Markus's method to RonCo's isn't necessarily comparing apples to oranges. RonCo was looking at a very specific slice of baseball history through his own lens. Markus is looking at the entire baseball universe through its entire history. The same methods won't work for both.

Markus stated his method is to get the league totals and leaders to look realistic for their era while at the same time getting the overall curves correct. I think that is a good approach.

Getting individual curves (different player types) modeled would be a huge challenge. Not only because you would have to make sure the curve looked right but that it also wouldn't adversely affect the bigger picture. So to get one curve correct you would have to essentially get them all correct. Just thinking about how many problems that would cause makes my head hurt!

Originally Posted by kq76

I think this might be the problem in a nutshell. For people who don't really care about statistics, averages are good enough. Well, for anyone really interested in the topic matter, whatever it may be, averages are far from enough. The entire shape of the curve is what matters: where do they start, how fast do they go up, how fast do they go down, and where do they end. By you saying that averages are about all that matters to you, you're effectively saying that the curve of these things could be way steeper or flatter than the truth and you just don't care because at least the averages are bang on. What you said right there tells me the issue is possibly much worse than my cursory interest in the issue had lead me to believe. I hope you change your mind about that sometime soon.

The shape of the curve is extremely important. But I haven't found any big problem with the distributions using the normal settings for different eras (its not perfect, but I don't think it can be). But for what its worth, there is a fair amount of debate in baseball circles on many of these points. Heck, 30 years ago, 50 years ago, how much did splits matter? How about the growth of the specialized closer? Baseball has changed over time, getting a game to approximate all that (and get the stats even close) is a huge challenge. Fixing settings in one area may have large (unintended) effects in others.

[Childress]

Originally Posted by Russ

Right. OOTP is way off on this.

Yeah, James' Almanac should own a place on Markus' bedside table. If it isn't there already.

He always demonstrated that (IIRC) players tended to peak statistically at age 27.

[akw4572]

Originally Posted by Markus Heinsohn

No, since the default setting produce very realistic results. Our research showed the OOTP 9 development is as realistic as ever. RonCo's research was great too, but it had different methods and different results. That's the thing about research using real stats, you can get different results depending on the method you use, and comparing these to OOTP stats is another matter as well. So, again, I personally feel the development & aging in OOTP 9 is the most realistic ever, and far far superior to any other baseball sim out there.

The only thing that was tweaked in OOTP 9.2.2 is the frequency of non injury-related talent drops for established player has been reduced. So, if your prefer RonCo's settings you can continue using these...

Thanks! That's all I wanted to know, didn't mean to stir up a hornet's nest.

[Russ]

I'm not trying to minimize the task of simulating the entire history of baseball in one game. Its daunting, I'm sure.

But there are certain things that hold true no matter the era. Power pitchers is one thing that is just modeled all wrong.

There are a lot of players whose rookie season was there best season. Wally Joyner and Al Kaline come to mind off the top of my head. But the significant drop off that is found way too often in OOTP is not found at that rate in real baseball. There's a reason that twenty years later people are still talking about Kevin Maas. It just doesn't happen that often. Players don't go from 50 HRs one year to 6 another at the age of 24 without a significant injury.

RonCo's analysis was done from a desired result backwards. But some of the results he found are not era specific. Couple Ronco's differences with Markus's method at looking at things from a far away perspective and it is obvious that there are problems here.

The main issue I'm having here is Markus listing those six or seven criteria as his way at looking at his player progression model. You have to look at an individual's career path to see if its working. This is more insightful than looking at leader boards, league averages, or retirement age.

[StyxNCa]

Originally Posted by Russ

There are a lot of players whose rookie season was there best season. Wally Joyner and Al Kaline come to mind off the top of my head. But the significant drop off that is found way too often in OOTP is not found at that rate in real baseball. There's a reason that twenty years later people are still talking about Kevin Maas. It just doesn't happen that often. Players don't go from 50 HRs one year to 6 another at the age of 24 without a significant injury.

Maas never had a change like that either. He went 21, 23, 11, and 9 HR's in 4 years. Gary Gaetti hit 5 in 1984 after hitting 25 and 21 in his 2nd & 3rd seasons then followed up 1984 with 20, 34, 31 and 28 the next four seasons. It isn't that rare. Those two I remember off the top of my head but I can dig up more if need be.

What I don't seem to be understanding is how can this "study" be the be all and end all when it's the AI controlling everything in the game? That's it, it handles all players the exact same way from their first game to their last. I play out games and I have not seen this problem. Yes, on occassion a player has a bad season, and in some cases a couple bad ones but that seems like a natural thing to me. Some players, Dock Ellis for example, just lose it for some reason. Same with Rick Ankiel.

As far as power pitchers, most don't last that long anyway. Ryan was one exception. I'm still amazed at how long he lasted. Of the ones who are around for a long period of time, many become a different type of pitcher as they age and lose some of the speed on their pitches, at least among the starters.

Now I will admit I haven't read what Markus uses or RonCo for that matter so maybe that's where I am misunderstanding this entire topic, or it appears I am. Oddly my biggest complaint has been how long many players stay so good which seems to be the complete opposite of what the issue seems to be here.

But when it comes down to it, what % of the 17,000 players that have made it to the majors had long careers into their late 30's or early 40's. It's not a very big number.

Anyway if it seems I'm rambling, I probably am. I haven't slept in 28 hours. Not a good thing for a truck driver so I better get to bed.

Sorry if I'm mistaking what you're saying. I'll read it all again when I'm not brain dead.

[BruceM]

Originally Posted by Russ

I think kq hit it on the head. Its not enough that the big picture looks good, its the little things that make up the big picture that need to be accurate.

Take for instance the rate of change in pitchers' K/9 ratio. These guys may be ending their careers with realistic strikeout totals but the way they're getting there isn't realistic.

So when you say "the way they're getting there isn't realistic", define realistic. Is that realistic as in MLB?

OOTP is a baseball sim that has many options and settings individuals can tweak to set it up to the way they like it, including aging/development modifiers.

Not happy with the way the game does it? Tweak them until you find something you like.

A lot of people keep mentioning "realistic" in this thread. Well guess what, not everyone plays real MLB. Fictionals, if I remember correctly from some polls, are far more popular. If you're playing fictional, who's to say the aging/development isn't realistic then? Fictional isn't realistic.

[kq76]

Originally Posted by mattshwink

...

It's a very valid point that Markus is going for the entire history of the game while others might not be, but that's not really the issue as I see it. The issue, as I see it, is that different qualities of players (i.e. speed, power, eye, whatever) increase and decrease at different times and rates. I could be wrong, but I don't believe these things have really changed all that much when it comes to humans over the past 100 years or so. Yes, our sprinters have gotten faster, we've all gotten taller and I guess stronger, it sounds like we've gotten healthier, but has it really changed that much relatively speaking? And I'd say, if anything, those things have only gotten better. I doubt any have gotten worse. Players didn't have great eyes at an early age back in the day or get faster at an older age.

On the other hand, it seems to me you're talking about the ebbs and flows of speed and power and such overall in a league (i.e. this era had lots of SBs, that era had lots of HRs). I, and I think others, are talking about changes in individual humans. Changes in league averages do happen and they can be pretty significant, but I think that's more due to non-human factors like juiced balls, raised mounds, etc. It's a different issue and I don't think player dev is supposed to effect that stuff, league totals are.

Player dev, to my understanding, is what is supposed to effect those changes in player qualities like speed, power and the like. And it's also my understanding that that is what people are taking issue with (e.g. when does speed decline? how fast does it decline? when does eye get better? how fast does it get better?). If a player peaks overall at 27 (or whatever the experts say it is now) and Markus has speed or eye peak at that age too (I'm not saying that's the case, just a possible example) when studies pretty clearly show speed peaks earlier and eye gets better with age, well, then there's a problem. It's got nothing to do with averages or leaderboards over 100 years; the issue is when do the different qualities of players in general peak and how fast do those qualities rise and decline.

Now if you want to argue that those rates have gotten better over time, well, you're most likely right, player dev probably should change over the eras, but we're not at that problem yet. I believe we're still struggling over the basic problem of this should change faster than that.

Perhaps the real problem is we're limited by the overall controls Markus has supplied us with. Sure, we can speed up or slow down batter or pitcher dev or aging, but we can't really speed up or slow down speed and do the opposite with eye or speed up K rates and slow down HR rates. So then when you speed up or slow down batters or pitchers in general to fix one aspect of them, you could very well be throwing off another aspect. If Markus gave us the controls to change each quality then maybe we wouldn't be complaining about it as much. It'd still probably be nice to have defaults that most people agree on and maybe even have it change over eras like we have for league totals, but even if we weren't given all that at least we could change the individual quality controls ourselves and move on.

I should stay out of this debate as others are more interested and therefore more well-versed in it than I am, but I believe that's the problem.

[Russ]

Originally Posted by BruceM

So when you say "the way they're getting there isn't realistic", define realistic. Is that realistic as in MLB?

OOTP is a baseball sim that has many options and settings individuals can tweak to set it up to the way they like it, including aging/development modifiers.

Not happy with the way the game does it? Tweak them until you find something you like.

A lot of people keep mentioning "realistic" in this thread. Well guess what, not everyone plays real MLB. Fictionals, if I remember correctly from some polls, are far more popular. If you're playing fictional, who's to say the aging/development isn't realistic then? Fictional isn't realistic.

That's what recalc is for. Even in fictional leagues you want player progression that is realistic. Unless you're playing with hobbits. That's an unknown right now.

The average career curve in OOTP does not match the average career curve in real life.

I tweak my settings now but there's still some things that can't be replicated because its the design of the game.

Maybe this is going back to the old argument of whether the settings should be defaulted to MLB type leagues. But using the factors markus uses to determine the accuracy will not lead to this being correct.

[BruceM]

Originally Posted by Russ

That's what recalc is for. Even in fictional leagues you want player progression that is realistic. Unless you're playing with hobbits. That's an unknown right now.

The average career curve in OOTP does not match the average career curve in real life.

I tweak my settings now but there's still some things that can't be replicated because its the design of the game.

Maybe this is going back to the old argument of whether the settings should be defaulted to MLB type leagues. But using the factors markus uses to determine the accuracy will not lead to this being correct.

Maybe you do, but I like very much what I'm getting now. I DO NOT want my fictional leagues to model realistic as you define it. I'm extremely pleased with the way it works at default settings. And no, I don't play with hobbits, whatever they are.

[Russ]

Originally Posted by BruceM

Maybe you do, but I like very much what I'm getting now. I DO NOT want my fictional leagues to model realistic as you define it. I'm extremely pleased with the way it works at default settings. And no, I don't play with hobbits, whatever they are.

The hobbits thing was a joke. I wasn't insulting you.

I'm glad you're happy. But I think that player progression is something that encompasses all types of leagues and should be based of reality. If MLB is the only place to see that kind of data, maybe that's too bad, but its all we got. I'm sure if studies were done with other leagues the aging process would be very similar.

[StyxNCa]

Originally Posted by kq76

It's a very valid point that Markus is going for the entire history of the game while others might not be, but that's not really the issue as I see it. The issue, as I see it, is that different qualities of players (i.e. speed, power, eye, whatever) increase and decrease at different times and rates. I could be wrong, but I don't believe these things have really changed all that much when it comes to humans over the past 100 years or so. Yes, our sprinters have gotten faster, we've all gotten taller and I guess stronger, it sounds like we've gotten healthier, but has it really changed that much relatively speaking? And I'd say, if anything, those things have only gotten better. I doubt any have gotten worse. Players didn't have great eyes at an early age back in the day or get faster at an older age.

On the other hand, it seems to me you're talking about the ebbs and flows of speed and power and such overall in a league (i.e. this era had lots of SBs, that era had lots of HRs). I, and I think others, are talking about changes in individual humans. Changes in league averages do happen and they can be pretty significant, but I think that's more due to non-human factors like juiced balls, raised mounds, etc. It's a different issue and I don't think player dev is supposed to effect that stuff, league totals are.

Player dev, to my understanding, is what is supposed to effect those changes in player qualities like speed, power and the like. And it's also my understanding that that is what people are taking issue with (e.g. when does speed decline? how fast does it decline? when does eye get better? how fast does it get better?). If a player peaks overall at 27 (or whatever the experts say it is now) and Markus has speed or eye peak at that age too (I'm not saying that's the case, just a possible example) when studies pretty clearly show speed peaks earlier and eye gets better with age, well, then there's a problem. It's got nothing to do with averages or leaderboards over 100 years; the issue is when do the different qualities of players in general peak and how fast do those qualities rise and decline.

Now if you want to argue that those rates have gotten better over time, well, you're most likely right, player dev probably should change over the eras, but we're not at that problem yet. I believe we're still struggling over the basic problem of this should change faster than that.

Perhaps the real problem is we're limited by the overall controls Markus has supplied us with. Sure, we can speed up or slow down batter or pitcher dev or aging, but we can't really speed up or slow down speed and do the opposite with eye or speed up K rates and slow down HR rates. So then when you speed up or slow down batters or pitchers in general to fix one aspect of them, you could very well be throwing off another aspect. If Markus gave us the controls to change each quality then maybe we wouldn't be complaining about it as much. It'd still probably be nice to have defaults that most people agree on and maybe even have it change over eras like we have for league totals, but even if we weren't given all that at least we could change the individual quality controls ourselves and move on.

I should stay out of this debate as others are more interested and therefore more well-versed in it than I am, but I believe that's the problem.

This makes it clearer to me as far as what the issue is.

The part I highlighted is exactly why I don't like making adjustments. Since I'm not sure how any setting interacts with others I'm just very leary to make any changes from the default. I want to be able to play the game, not spend days and weeks and possibly months adjusting settings, simming a season, then adjusting settings again if things don't turn out right.

[Russ]

What kq said.

[Tycobbler]

Either A. play and enjoy the game or B. don't

With each passing year some want more, and those that don't get their way ie. RonCo, leave.

Those in beta know. We read RonCo's posts. Like his, and everyone else's they were taken into consideration and out of it OOTP 9 was born.

Could the game be better? I think the question should be will the game get better? The latter pretty much answers itself.

[kcroyalgm]

Originally Posted by Tycobbler

Either
A. play and enjoy the game or
B. don't
With each passing year some want more, and those that don't get their way ie. RonCo, leave.
Those in beta know. We read RonCo's posts. Like his, and everyone else's they were taken into consideration and out of it OOTP 9 was born.
Could the game be better? I think the question should be will the game get better? The latter pretty much answers itself.

Well said.

[BruceM]

Originally Posted by Russ

The hobbits thing was a joke. I wasn't insulting you.

I'm glad you're happy. But I think that player progression is something that encompasses all types of leagues and should be based of reality. If MLB is the only place to see that kind of data, maybe that's too bad, but its all we got. I'm sure if studies were done with other leagues the aging process would be very similar.

And I for one, disagree. I like the way fictional unfolds and gives me something else besides "reality". I only play 2 games, OOTP & Links Golf. In both I have an opportunity to get away from "reality" and everyday things.

To each his own....