Talent Changes

[Elendil]

Originally Posted by Erithtotl

It's my understanding that messing with the talent change frequency would make the problem worse, not better, because it does not effect severity of change, but rather just the frequency.

While ultimately I'd like to see a reworking of how development is handled, I think seperate sliders for talent change severity and injury severity would go a long way towards me being able to tweak the game the way I'd like. For example, in both cases I'd likely INCREASE frequency, but decrease severity, to create more short to medium term injuries (and fewer of the 8 month variety) and more subtle talent decreases and increases (and fewer massive, Rick Ankiel variety collapses).

I've heard that you can edit an injury text file to create injuries of the variety and length that you like, but I've never done it myself.

[injury log]

Originally Posted by Elendil

I've heard that you can edit an injury text file to create injuries of the variety and length that you like, but I've never done it myself.

You can indeed. The resulting game behaviour is slightly unexpected, however. My understanding of OOTP's injury implementation, which may well be wrong, is that the game first decides whether an injury happens, and determines its severity. It then scans the injury database for an appropriate diagnosis. If no appropriate diagnosis exists, then the injury simply doesn't happen. If you, for example, edit all the severe injuries to become mild ones, when in-game an injury happens that the game decides is severe, it won't find an appropriate injury in the database, and you get no injury at all. The upshot: if you change your severe injuries to mild ones, you will not only see no severe injuries, you'll also see fewer injuries overall.

If I'm totally off-base in my understanding of how this works, I'd be very interested in a clarification. I'm planning on running a few rigorous tests to confirm or refute what I'm saying, and will post the results eventually, in a separate thread.

[jbone]

Originally Posted by Erithtotl

Tatis is the closest thing I can come to an example of this, but Tatis also had a bunch of injuries after his two productive years (his year after his big year wasn't bad). And that emphasizes my other point. Tatis didn't just collapse. The year after his monster year he suffered through an injury prone year where he hit only .253, but with a .379 OBP and a .491 slugging, an excellent year for a 3rd baseman. His problem was that he had all those injuries and lacked the work ethic or the luck to overcome them. And Ankiel clearly had some kind of mental problem that would be be implemented as an injury. Again, if the massive talent collapses were associated with injury, it might make some kind of sense, but this guy was healthy as can be, and his talent and ability collapsed in one day, rather than over a period of several months.

I'm not complaining about losing the player. I'm complaining about the way it happens in the game is not realistic. You don't just wake up one day and forget what you could do fine yesterday. And, again, this is not an isolated occurance, it has happened to me 3 times in 6 seasons and I can only assume it happens for the AI at similar frequency, so that means 15 players a year or so, rather than 1 or 2 every 5 years as might be the case by the few examples we have been able to come up with.

Point well-taken.

[jbone]

Originally Posted by Erithtotl

Agree with you %100. It likely will never happen, but I always thought an interesting model would factor current athletic ability, athletic potential, work ethic, intelligence, injuries, and a big chunk of randomness to determine player development, with no 'potential ratings' in skill related categories. The game could adjust those 'randomness' factors by real world knowledge (like the idea that players with extensive 'old players skills' (high BB and HR, low BA) have early peaks, as do heavier players, while all around athletes tend to age better. So when you look at a prospect you'd see the following:

How good the guy is now at baseball stuff. How athletic he is. Your scouts estimate of how well he could 'grow into his body', so to speak. Your team estimates of how hard a worker he is and how smart he is. That's it. No knowledge of when he is going to peak or what his hard ceiling is.

To me that would be a truly innovative and challenging growth model. But it would take a rewrite of an engine that for the most part does a good job of meta-simulating real world talent development, even if it falls short in a few areas right now (as documented here).

Regardless of anything I said before, I have to agree that, from a gamer's perspective, this would be the ultimate. All those periphery factors are all that we should/could know about any prospect (other than college/hs stats), not that he has an "eye potential of 14", which doesn't really mean anything anyways. We could scout a hitter and say, "hm, he was a patient hitter in college, and he's intelligent, so he's probably more likely to be selective at the plate in the pros".

Ironically, at this point you would be modelling the game more like Football Manager than OOTP (those of who haven't may want to check out that game and how it functions).

[jbone]

first off, sorry for all the posts, but...

Originally Posted by RonCo

To get the idea of game design for talent development into the right frame of reference, you need to stop playing the part of "scout" and start playing the part of "God". I'm fine with ratings ebbing and flowing on a day-to-day basis based on whatever arbitrary model the designer wants to use...biorhythms, streaks, bolts of lightning, whatever...but in accordance with comments of Erithtotl and others, a model that has a _lot_ of guys waking up one day to find themselves fundamentally worse baseball players than they were yesterday is flawed in an elemental way.

It is flawed, I agree with that.

Okay, time to perhaps freak some of you people out, I apologize in advance. But...

It seems as if all these games are produced based on the concept that there is one universal truth or "God", mind you, about what is real. This thought process is based on our Western, monotheistic ways of viewing things. But real life is constantly changing, every millisecond, every atom on your keyboard even i believe. There are other systems of thought that believe that God is within ourselves, so to speak, ie. Buddhism. How the player views himself is paramount to what talent and performance is. Why not do player development from a Buddhist perspective? I don't understand why a computer game would need some concept of "ultimate truth" in order to function. After all, the players are only competing against each other, not against God. Players only need current abilities, not pre-ordained potentials, in order to play a game of baseball.

I am not arguing that players are miraculously able to grow six inches in one day. What I'm arguing is that ultimately it is up to the player to make the most of his genetics. But other than genetic limitations, it is an open-ended system which gives everyone a chance. In MLB, no player is ever drafted with the pre-ordained potential to max out at Class A ball - if so, there wouldn't be any point in drafting him. While a player may have certain genetic physical limitations, no player should ever be precluded from being a major league contributor, just look at regular-Joe types such as Greg Maddux, or the legally-blind Mike Bordick. These guys with very poor athleticism found a way to succeed. Maddux may have seemed like a weak prospect at draft-time, but he used his guile and work ethic to increase his abilities. It can be done.

Check out this prospect, drafted in the 17th round by the Red Sox in 2006:

William Redick, Middle Georgia College
Position: OF Bats: Left Vitals: 19 years old, 6-2; 180

Go find Redick in person. Are you going to go up to William Redick and tell him to his face that his pre-ordained potential by God as a baseball player is to be a AA outfielder with an 8/7/8? It would be nonsensical to say something like that, and you'd be laughed out of the stadium. If Redick didn't have a shot, he wouldn't be there. In spite of our perceptions regarding OOTP, the foundation of player development as creating an "ultimate truth under God" just doesn't make any sense.

[Syd Thrift]

As usual, I find myself in total disagreement with RonCo. Don't get me wrong; I think he is an unbelievably intelligent guy and the fact that he states his position so clearly that I can so vehemently disagree with him is a huge plus. Thanks to him, I think that scouting and development will greatly improve next year whether Markus takes his advice or the advice of his "opponents" (in quotes because in truth, I think we're all united behind the common goal of making the game better.

I'll go point by point with what I want to see:

- I think that adding a third "tools" rating in with actual ratings and potentials would not make the game more realistic; in fact, I'm thinking it would make it even less realistic than it is right now. If you want the game to generate how fast a guy can run the 40 or how many bench presses he can do, make the game generate those as cosmetic stats based on speed and homerun power and with a lot of randomness thrown in.

- I think that players not reaching their potentials ought to be a common, every day sort of thing - something that happens way more often than it does now. You want to recreate superstars? Give them high POTs, sure, but make their ACTs a very high percentage of those to start out with. A great player generally has a high percentage of his POT realized even when he's very, very young. A guy good enough to be one of the best players in history should probably be good enough to be a league-average player when he's 18. Or better. Look at how Ted Williams and Joltin' Joe DiMaggio tore up the PCL. With pitchers it's even more extreme, as historically speaking an *awful* lot of hurlers seem to have had better seasons in their early to mid 20s than in their early 30s. More on that in a second...

- That being said, I don't think it's unrealistic to occasionally see players lose potential when they are young. Should they actually drop below their current abilities? Eh, probably not except in the case of injury (but I could see that in extreme circumstances - see Bob Hamelin or Eric Yelding or Pat Listach), but there are absolutely tons of examples of young rookies who burst on the scene, looked awesome, but never became the superstars they were projected to be or even advanced beyond the players they were in their rookie seasons. Claudell Washington springs to mind. See also: Adrian Beltre (2005 notwithstanding). I sincerely doubt the game is going to model what caused Cesar Cedeno to slide from aWesome to very good from age 21 to age 27, so a random hit would be appropriate there too.

- You notice that I didn't ask why in those cases? I don't remember Claudell Washington or Pat Listach or Cesar Cedeno getting hurt (unless you count being heckled in every town for perhaps committing murder as a psychological injury). They just never cashed in on their potential, that's all. Or maybe their potential was false. Either way, IMO it is realistic to model this with POT hits.

- Some stats are ridiculous for scouts to be able to project at all. Eye and Control are the two that make me giggle (I know this has been stated elsewhere but still). I'm not saying you should remove potentials for these. They should be hidden whether you have scouts on or off.

- This seems unrelated, but trust me it is not: injuries ought to be even more plentiful than they are but at the same time there should be a lot fewer "OUT" injuries. Very rarely is a player so badly hurt that he is absolutely unable to play for a series, much less 2 weeks. Players even play through broken legs or hands (as long as they're not too badly broken) and then get surgery in the offseason. More often what happens is that a player complains that he's hurt (or tries to hide it) and the manager, GM, and coaching staff get together and decide that they'll be better off in the long run if they let him rest. This was close to the model that Baseball Pro '98 used. While injured, the player should be more prone to re-injuring himself, the injury should take longer to heal, and each day he's hurt there should be a (slight) chance of a ACT/POT hit to a different attribute depending on the extremity that has been damaged.

- Also, unless the player is hurt so badly that he can't play at all, we should have a lot less information on how badly he's likely to play. PureSim used a scale from Mild to Very Serious IIRC, which is a great way to set it up. Nobody should know that a guy has a "9%" injury; for one thing, a player with a sprained finger should have basically no hit whatsoever to his speed, and for another thing boiling it down to a number is silly and unrealistic. Sure, there should be a number in the back end but not even House could tell you a 9% injury is a 9% injury with a straight face. Well, maybe House but only because that man can tell brutal lies with a straight face.

- One effect of waaaay more injuries: more talent and/or POT hits. Maybe then we could get rid of the random drops and then RonCo would be happy! I am perfectly content to accept that, say, Listach suffered an injury the Brewers didn't tell anybody about following his rookie year and that's why he sucked donkey marbles afterwards.

- How many injuries is "waaay" more? I have an inkling that at any given time, roughly 1/3rd to 1/2 of all pitchers ought to be nursing some sort of injury. Maybe they don't affect them much except that they stand a greater chance to be hurt down the line if they're overused (although I think the majority of guys should see their abilities actually reduced) or whatever, but IMO a big part of being a real-life manager is figuring out which pitchers' complaints you can live with and which ones you can't. That would also help to solve another big issue with the game: the one where pitchers are far more consistent than in real life.

- I know (from talking to Markus) that PAPs play a factor there but IMO they should play an even larger one. A player who throws, say, 150 pitches in a game ought to have basically a 100% chance of getting injured. Now, if that guy is Randy Johnson in 1995 he's probably still going to be way effective even through a minor injury, but IMO the game right now does not sufficiently punish managers for behaving like Kevin Kennedy. Erik Hanson threw one - one - lots-o-pitch game and that essentially ruined him. The same thing set back Tim Wakefield for a couple years, even though he's a knuckleballer with a supposedly rubber arm. IIRC Bret Saberhagen was also essentially destroyed as an innings-muncher by one or two long outings.

- An alternate take on the above: while throwing a guy a lot of pitches in one game should hurt, what should also hurt is throwing a guy a lot more pitches than he's used to. If you train a guy to throw 100 pitches a start, you may be keeping him out of injury, but if he then throws 130 pitches he ought to have a far greater chance of getting hurt than a guy who averages 120 pitches a start. If we just model the game as above, the 120 averaging guy is probably going to be more prone because he'd have probably been injured several times before from overuse and his proneness rating went up. Should that trend be reversed? Hard to say... evened out maybe?

A lot of thoughts, I know. Coherent? I'll let you be the judge of that!

[magnet]

Syd is just hoping the beta team next year is chosen by # of words posted.

[Syd Thrift]

Originally Posted by magnet

Syd is just hoping the beta team next year is chosen by # of words posted.

I don't think they'll be bringing me back. If the work RonCo did is what's needed out of us, I think I maybe put in 1/10th of 1% of what was required. Still, it was a great deal of fun and I was glad to have been a part of it, even if my main "part" was getting under RonCo's skin.

[Brewer77]

Originally Posted by Syd Thrift

- An alternate take on the above: while throwing a guy a lot of pitches in one game should hurt, what should also hurt is throwing a guy a lot more pitches than he's used to. If you train a guy to throw 100 pitches a start, you may be keeping him out of injury, but if he then throws 130 pitches he ought to have a far greater chance of getting hurt than a guy who averages 120 pitches a start. If we just model the game as above, the 120 averaging guy is probably going to be more prone because he'd have probably been injured several times before from overuse and his proneness rating went up. Should that trend be reversed? Hard to say... evened out maybe?

Syd, good post, good points, however I strongly disagree with you on this. Pitch counts really train pitchers' arms so they can't throw more. Were the pitchers of the 40's and 50's more prone to injury because they consistently threw 150 or so pitches a game?

I may catch a little flack for saying that but I disagree with pitch counts.

[Syd Thrift]

Originally Posted by Brewer77

Syd, good post, good points, however I strongly disagree with you on this. Pitch counts really train pitchers' arms so they can't throw more. Were the pitchers of the 40's and 50's more prone to injury because they consistently threw 150 or so pitches a game?

I may catch a little flack for saying that but I disagree with pitch counts.

It's as close to a rock-solid fact that if you put a pitcher in long enough, he'll get hurt. Maybe 140 pitches is too low; maybe the bar should be 160. Also, this would be something that would vary by era; of course a deadball era guy should be able to toss both halves of a double-header in theory without wearing his arm out too much. So by the 50s, perhaps the cut-off point could be 180 tosses.

See also my later idea about training arms to tire out after a given number of pitches. Call it the NiMH theory of endurance.

[RonCo]

Originally Posted by Syd Thrift

As usual, I find myself in total disagreement with RonCo. Don't get me wrong; I think he is an unbelievably intelligent guy and the fact that he states his position so clearly that I can so vehemently disagree with him is a huge plus. Thanks to him, I think that scouting and development will greatly improve next year whether Markus takes his advice or the advice of his "opponents" (in quotes because in truth, I think we're all united behind the common goal of making the game better.

Nothing wrong with a good, healthy debate.

- I think that adding a third "tools" rating in with actual ratings and potentials would not make the game more realistic; in fact, I'm thinking it would make it even less realistic than it is right now. If you want the game to generate how fast a guy can run the 40 or how many bench presses he can do, make the game generate those as cosmetic stats based on speed and homerun power and with a lot of randomness thrown in.

Improvement would depend on how well the change was done, of course.

- I think that players not reaching their potentials ought to be a common, every day sort of thing - something that happens way more often than it does now. You want to recreate superstars? Give them high POTs, sure, but make their ACTs a very high percentage of those to start out with. A great player generally has a high percentage of his POT realized even when he's very, very young. A guy good enough to be one of the best players in history should probably be good enough to be a league-average player when he's 18. Or better. Look at how Ted Williams and Joltin' Joe DiMaggio tore up the PCL. With pitchers it's even more extreme, as historically speaking an *awful* lot of hurlers seem to have had better seasons in their early to mid 20s than in their early 30s. More on that in a second...

I agree with all that.

Using a more stable potential platform would probably make it easier to write a good development engine.

- That being said, I don't think it's unrealistic to occasionally see players lose potential when they are young. Should they actually drop below their current abilities? Eh, probably not except in the case of injury (but I could see that in extreme circumstances - see Bob Hamelin or Eric Yelding or Pat Listach), but there are absolutely tons of examples of young rookies who burst on the scene, looked awesome, but never became the superstars they were projected to be or even advanced beyond the players they were in their rookie seasons. Claudell Washington springs to mind. See also: Adrian Beltre (2005 notwithstanding). I sincerely doubt the game is going to model what caused Cesar Cedeno to slide from aWesome to very good from age 21 to age 27, so a random hit would be appropriate there too.

The occasional mysterious loss of top-end potential would be fine. That said, I don't see any of your examples actually showing that in their stats:

Bob Hamelin - Had one great year at 26 yo. However, two warning signs...he was a 25 yo rookie (old), and his great season included only 31 AB (terrible sample size).
Eric Yelding - One of my favorites...I kept using him on my fantasy team because he was so speedy. Unfortunately, he was never a good baseball player, never registering an OPS+ of greater than 69...Yikes...
Pat Listach - Wow, what a bad player. Great example of hype/scouting being way off. His ROY season consisted of 579 AB of less than league-average performance (99 OPS+). Beyond that year, he never had another that even came close to league average. Given his age (24) and that so much of his value was due to his speed (which is known to be falling heavily at that age), I find is unsurprising that his performance quickly fell to Yelding levels and below.
Claudell Washington - Another great example of hype. His actual performance his first two seasons show no signs of greatness (except for his age). But he was very fast, and some people liked that. So, despite an OPS+ that was never much better than just over league average, the scouts all raved about his potential. Taking the scouts out of it, perhaps he did have great potential but just didn't grow into it. Dunno. Washington was never knows as a great worker as I recall.
Cesar Cedeno - The best argument you've got, due to his 21-22 yo seasons being so outstanding. If you do the mathematical probability and statistics on it, though, I'll be willing to bet a lunch that the chances of a guy who routinely puts up 130 OPS+ numbers like Cedeno did throughout his career, putting up a couple 150s by random chance is pretty high ( my pure guess would be he had a 10-15% chance of having those numbers just by random chance assuming his later performance was a "true" baseline of his skills.

- You notice that I didn't ask why in those cases? I don't remember Claudell Washington or Pat Listach or Cesar Cedeno getting hurt (unless you count being heckled in every town for perhaps committing murder as a psychological injury). They just never cashed in on their potential, that's all. Or maybe their potential was false. Either way, IMO it is realistic to model this with POT hits.

For the most part they "didn't reach their potential" because they never had that potential to begin with...it was all scouting/marketing, or an inability to understand the story that the stats they actually put up were telling about them. Cedeno is the only example you've given that his even close to a mystery to me.

Regardless, each of these could be equally or better modeled by the system I've proposed.

- Some stats are ridiculous for scouts to be able to project at all. Eye and Control are the two that make me giggle (I know this has been stated elsewhere but still). I'm not saying you should remove potentials for these. They should be hidden whether you have scouts on or off.

Eye = BB/PA
Control = BB/BF

I agree they are difficult for a human being to project. Your argument just bolsters my side of the equation, I think.

- This seems unrelated, but trust me it is not: injuries ought to be even more plentiful than they are but at the same time there should be a lot fewer "OUT" injuries. Very rarely is a player so badly hurt that he is absolutely unable to play for a series, much less 2 weeks. Players even play through broken legs or hands (as long as they're not too badly broken) and then get surgery in the offseason. More often what happens is that a player complains that he's hurt (or tries to hide it) and the manager, GM, and coaching staff get together and decide that they'll be better off in the long run if they let him rest. This was close to the model that Baseball Pro '98 used. While injured, the player should be more prone to re-injuring himself, the injury should take longer to heal, and each day he's hurt there should be a (slight) chance of a ACT/POT hit to a different attribute depending on the extremity that has been damaged.

- Also, unless the player is hurt so badly that he can't play at all, we should have a lot less information on how badly he's likely to play. PureSim used a scale from Mild to Very Serious IIRC, which is a great way to set it up. Nobody should know that a guy has a "9%" injury; for one thing, a player with a sprained finger should have basically no hit whatsoever to his speed, and for another thing boiling it down to a number is silly and unrealistic. Sure, there should be a number in the back end but not even House could tell you a 9% injury is a 9% injury with a straight face. Well, maybe House but only because that man can tell brutal lies with a straight face.

The injury system seems greatly improved to me, but I'm sure it can always be done better. I really haven't studied it closely.

- One effect of waaaay more injuries: more talent and/or POT hits. Maybe then we could get rid of the random drops and then RonCo would be happy! I am perfectly content to accept that, say, Listach suffered an injury the Brewers didn't tell anybody about following his rookie year and that's why he sucked donkey marbles afterwards.

Per above, Listach's problem was he was not as good as some folks thought he was to begin with.

I'll touch on pitch counts and pitcher injuries in another post.

[RonCo]

Originally Posted by Brewer77

Syd, good post, good points, however I strongly disagree with you on this. Pitch counts really train pitchers' arms so they can't throw more. Were the pitchers of the 40's and 50's more prone to injury because they consistently threw 150 or so pitches a game?

I may catch a little flack for saying that but I disagree with pitch counts.

I suspect that the truth will eventually be revealed in further data studies, but if I were guessing, based on things I've seen in the past year, we'll find that there's a middle ground here....

• Pitch counts are fairly obviously valid, throwing a lot of pitches when a pitcher is tired is an almost certain cause of injury.
• A pitcher can be "trained" to throw more pitches.
• A pitcher who throws 110 or so pitches can probably throw every 4th day rather than every fifth.

So I'm guessing that eventually we'll see teams sticking with pitch counts, but going to 4-man rotations again.

Regardless, the PAP viewpoint seems to be strongly supported by data. I support its inclusion in OOTP.

[Syd Thrift]

I remembered Yelding having some huge minor league season but looking back on things it looks like it was just a speed thing. He did have a really good year in A ball in 1987... at age 22, after moving down half a level. His minor league stats look just plain awful, and yet I remember actual statheadish publications screaming about how his arrival would be so awesome. I'll stand corrected on him.

Hamelin I know was an older rookie, but that's an argument for a descent into mediocrity after a solid first season. Hamelin completely fell apart in 1996. He did bounce back but was never anywhere near as good as he was his rookie campaign.

As for Claudell Washington, I think work ethic was cited more than bad hype. A player good enough to start in the major leagues at age 19 is, generally speaking, a very, very good player. A guy who can post OPS+es of 108 and 119 has a good shot at being stupendous. I saw A-Rod at a similar age and what Washington did was a lot more than what Rodriguez did. No, he never turned into that player, but to put it in an OOTP model, when every scout in the land thinks a guy has that kind of potential, he probably does have that kind of potential. He just didn't reach it, is all. To model a player like Washington correctly, it's not enough to rely on scouting error.

[RonCo]

Originally Posted by Syd Thrift

I remembered Yelding having some huge minor league season but looking back on things it looks like it was just a speed thing. He did have a really good year in A ball in 1987... at age 22, after moving down half a level. His minor league stats look just plain awful, and yet I remember actual statheadish publications screaming about how his arrival would be so awesome. I'll stand corrected on him.

Hamelin I know was an older rookie, but that's an argument for a descent into mediocrity after a solid first season. Hamelin completely fell apart in 1996. He did bounce back but was never anywhere near as good as he was his rookie campaign.

As for Claudell Washington, I think work ethic was cited more than bad hype. A player good enough to start in the major leagues at age 19 is, generally speaking, a very, very good player. A guy who can post OPS+es of 108 and 119 has a good shot at being stupendous. I saw A-Rod at a similar age and what Washington did was a lot more than what Rodriguez did. No, he never turned into that player, but to put it in an OOTP model, when every scout in the land thinks a guy has that kind of potential, he probably does have that kind of potential. He just didn't reach it, is all. To model a player like Washington correctly, it's not enough to rely on scouting error.

Comparing Washington and Rodriquez at that age is way out of line. By 20 yo, A-Rod had grown 35+ HR power, whereas Washington's value was totally built on speed (yet he still had a fairly low 2B/3B rate) and BABIP. A-Rod's 19 yo season projects to close to 20 HR if he got a full slate of AB...Washington hit 10 HR in 590 AB as a 19 yo. I know scouts were huge on Washington, but that data was screaming "a little above average." A-Rod's walk rate was about double Washington's showing an advanced plate discipline, too. I put the fault of Washington's failure to develop into a superstar fully at the feet of those human beings who proclaimed his potential to be greater than it seems to have been.

My opinion here is that anyone who looked at these two players and projected Claudell Washington to be as good or better than A-Rod was about as blind to what performance means as is possible to be.

[StyxNCa]

Well, I would like to know how my head scout just came back from scouting my player when he is supposed to be in China scouting their players.

[Syd Thrift]

Originally Posted by StyxNCa

Well, I would like to know how my head scout just came back from scouting my player when he is supposed to be in China scouting their players.

Really strong binoculars?

[RonCo]

Here's an example of the "problem"--which is only a problem if you consider the result a problem, of course. OOTP does a really great job of populating the majors overall, but occasionally you'll hear someone complain that they can't find any really young stars. Here's an example of why.

I took a test league of 8 teams and ran 60 seasons. I then took all players drafted in the first 40 seasons and grabbed their talents on Jan 1 of each season (so note, this data "misses" multiple changes in a single year). I then tabulated the changes in talent seen by players of each age from year-to-year, and present them here in "probability field" form.

I'll show only contact here. Contact, remember, is not a rating unto itself, but is instead a conglomeration of BABIP, AvoidK, and Power...so it works as an amalgam that says if a hitter's overall value is increasing or declining (not counting Eye, Gap, and Speed). To read this chart, you can say that 22% of 17yo hitters will see their Contact Talent raise by the time they are 18. Similarly, 42% will see if fall. There is no differentiation in this chart between a 1 point change and a 45 point change.

Here's the data:

% Change to Improve									
Contact	Age	Y1	Y2	Y3	Y4	Y5	Y6	Y7	Y8
	17	0.22	0.40	0.54	0.64	0.70	0.75	0.78	0.80
	18		0.23	0.41	0.53	0.62	0.68	0.72	0.75
	19			0.23	0.39	0.51	0.59	0.64	0.67
	20				0.22	0.36	0.47	0.53	0.58
	21					0.19	0.32	0.36	0.40
	22						0.16	0.27	0.34
	23							0.12	0.21
	24								0.09

The chart above says that a hitter who enters the league at 17yo has an 80% chance of receiving a Contact Talent increase by the time they are 24. Similarly, a guy who enters the universe ay 20 has a 58% chance of receiving a talent increase (remember again, most of these increases are small). Overall, this sounds pretty good again. At least reasonable.

But now let's look on the downside...talent/potential hits:

% Change to Decline									
Contact	Age	Y1	Y2	Y3	Y4	Y5	Y6	Y7	Y8
	17	0.42	0.63	0.75	0.83	0.88	0.91	0.94	0.96
	18		0.37	0.58	0.71	0.79	0.85	0.89	0.93
	19			0.33	0.54	0.67	0.76	0.83	0.88
	20				0.31	0.50	0.64	0.74	0.82
	21					0.28	0.48	0.63	0.75
	22						0.27	0.48	0.64
	23							0.29	0.51
	24								0.31

Now we see how harsh this developmant model is. A kid entering the universe at 19yo has a 80% chance of receiving an increase by the time he is 24, true enough...but he also has a 96% chance of receiving a talent/potential decrease. A kid who starts his pro career at 18yo has a 93% chance to receive a talent decrease by 24. And a kid who starts at 19 has an 88% chance of receiving a talent decrease by the time he's 24. In other words, if you draft a teenaged hitter, it is neaerly a mathematical lock that he will receive a Contact Talent decrease before he's 24. If he doesn't it's a mathematical oddity...an example of pure random luck of high order.

There are other nuances, of course. For example, the data shows that about 30% of players will receive both positive and negative talent movement before they are 24, so taking a hit does not guarantee a player won't recover. In addition, every skill has a slightly different probability field. I should note that some 25% of players took multiple hits before age 24. Finally, once again, over 90% of talent changes are less than 20 points, bu tthe chance of taking a negative hit > 20 points is 4 x higher than the probability of receiving a positive movmement of that magnitude. This means a highly gifted player may have to take several hits prior to becoming useless. But in the end, this gets the basic story across for the average player who enters the OOTP universe at any particular age.

[RonCo]

Hmmm...working too fast. I've got some obvious math wrong in the last charts, which I'll come back and cleanup...but the message is the same. Just a moment.

[RonCo]

Cleaned it up. Sorry about that.

[andymac]

I have always actually though the most realistic model would be to not have any potential ratings whatsoever. A player has his base rating and the development model should give him a certain chance of improving/losing those ratings based on his age & other factors (i.e. work ethic, coaching). Scouts would then be projecting future ability by only knowing current ability (which is the way it works IRL). It would also make it more possible for every scout to be differnt as they could have their own ratings which factor in the way they project (and any errors in current evaluation would also factor in).