Player Development is not right?

[Trebor]

Can someone shed some light on why so many of my young players never develop? Am I missing something here?

Or maybe the player development in the game is not right, I see many young players in the free agent and waiver wire pool, that at one time where can;t miss prospects.

[Seviien]

Todd Van Poppel was "can't miss" too at one point

I can't say anything for the AI and how it decides to release players, but in aggregate the player development Model in v9 is the best it's ever been -- and the best in any baseball sim I've ever taken a look at. Some players are going to bust. It just happens. If anything, OOTP needs more superstars to come out of nowhere. If you run a league for any length of time, the top 10 all-time HR hitters will ALWAYS be taken within the top 10 picks of the draft -- usually within the top 5 picks.

FWIW...
Since OOTP 2006 I've plotted the stat ouputs per AB or per BF for players at different ages in fictional OOTP leagues that have been running for at least 50 years. Then I compared those results back to MLB for the same time period.

e.g. I'd let a fictional league run from 1900 to 2008, then take the stat outputs from the league for the 1980-2008, chop them up by player age, and then compare those results to simmilarly sliced MLB data from the same time period.

Those reports, coupled with Ronco's comparisons of ratings to Tango data helped Markus ground the development model during the beta process.

In 2006 we were way, way off. So far off it wasn't funny.
In 2007 / v8 we made huge improvements and things got close enough to "perfect" that 99% of people won't notice what could be better.
In v9, things got a tad better than in v8

On the whole, I'd say we're 90% or better at capturing how players really develop. "Stuff" is still lacking, or at least it was the last time I ran a test.

just my $0.02

[RonCo]

Originally Posted by Seviien

Todd Van Poppel was "can't miss" too at one point

I can't say anything for the AI and how it decides to release players, but in aggregate the player development Model in v9 is the best it's ever been -- and the best in any baseball sim I've ever taken a look at. Some players are going to bust. It just happens. If anything, OOTP needs more superstars to come out of nowhere. If you run a league for any length of time, the top 10 all-time HR hitters will ALWAYS be taken within the top 10 picks of the draft -- usually within the top 5 picks.

FWIW...
Since OOTP 2006 I've plotted the stat ouputs per AB or per BF for players at different ages in fictional OOTP leagues that have been running for at least 50 years. Then I compared those results back to MLB for the same time period.

e.g. I'd let a fictional league run from 1900 to 2008, then take the stat outputs from the league for the 1980-2008, chop them up by player age, and then compare those results to simmilarly sliced MLB data from the same time period.

Those reports, coupled with Ronco's comparisons of ratings to Tango data helped Markus ground the development model during the beta process.

In 2006 we were way, way off. So far off it wasn't funny.
In 2007 / v8 we made huge improvements and things got close enough to "perfect" that 99% of people won't notice what could be better.
In v9, things got a tad better than in v8

On the whole, I'd say we're 90% or better at capturing how players really develop. "Stuff" is still lacking, or at least it was the last time I ran a test.

just my $0.02

Since you bring my name into it ...

I'll say that overall the development curves in v9 appear to me to be a fairly big step backward, requiring changes in the aging and development modifiers to bring them closer into line with reality (not really necessary in v2007). They have also not improved as I would expect them to have--pitchers are still off, and hitters still hold their power too long.

Career lengths are still too short overall (possibly due to AI, but I don't think so, since I've been testing mostly at AI settings highly weighted to ratings). The numbers appear to be pretty much equal from v2007->v9.

Its one "improvement" is that Markus muted the number of big changes, thereby lowering the number of immediate and vocal complaints. Ths is a big improvement because it was no fun whatsoever to have all your great prospects killed by the v2007 game engine...I heartily recommend changing to v9 for this factor only (my online league will be doing so soon). However, all this has really done in the big picture is remove some of the variation and surprise from the draft. You see fewer stars coming from late rounds. There are some, of course. But not many...and HOF registers are still massively swayed to the 1st round, as you noted.

As I said, I view the dev curves in v9 to be quite a bit inferior to those in v2007. Quite honestly, and with nothing but proper reverence for the dev team and all Markus's hard work, I would appreciate it if my name weren't associated with them.

[SittingDuck]

Wowsers.

[DrSatan]

I have to disagree. I'm not sure what kind of settings your using, but I consistently see late round draft picks develop into quality players, some into stars. I think development has greatlly improved compared to previous versions. I don't have nearly as many busts, the problem with pitchers being Cy Young next year and crap the next year appears to be fixed, and there are a lot more positive talent changes with prospects in this version. I think the problem is people may not know how to promote properly in the minors causing players to bust. An easy way to see if late round prospects are getting talent upgrades is to shop a good player and look at all the 3+ star players that are 23yrs old and younger, and see what round they were drafted in. You can also see their talent changes.

[RonCo]

Originally Posted by DrSatan

An easy way to see if late round prospects are getting talent upgrades is to shop a good player and look at all the 3+ star players that are 23yrs old and younger, and see what round they were drafted in. You can also see their talent changes.

That may be an easy way. But the _easiest_ way to see if your players are getting talent/rating changes is to pull csv dumps and compare players from sim-to-sim and year-to-year, which is what I do. This leaves nothing to the imagination, and tells me exactly which players are changing and by how much.

Once you do this, the easiest way to compare versions is to run thousands of players created by each OOTP version through an automated routine that tracks careers and compiles reports on them.

When you do this with v2007 on default settings you find that, while the hitter curves are a little messed up, they peak at about the right places. And that pitcher curves are really kinda messed up all over the place.

When you do this with v9 on default settings you find that hitters and pitchers both peak too young and that pitchers are still kinda all over the place. Markus seems to have improved the aging a little in the patches, but it's still off (and "worse" than v2007)...and pitchers stuff actually has this weird hump in it that it never had before.

Here's the HoF run from my last 100 year run--with the latest patch, and my best adjustment for aging...which I think still needs work based on the last patch.

Round 1: 67
Round 2: 10
Round 3: 3
Round 4: 2
Round 5: 2
Round 6: 3
Round 7: 4
Round 8: 0
Round 9: 1
Round 10: 5

Over two-thirds of the HOF comes from the first round, which is:

1) Too much based on real life
2) More than routinely came from v2007

I've been doing this for about 4 real-time years now with versions 5 and on, and over probably 500 leagues over that time. I use the same basic code (with different interfaces due to the changes Markus makes in the csv structure). Perhaps I'm missing something with v9, but I'm not inclined to think so.

[SittingDuck]

Yah, that first-rounder thing is an issue.

[sterjs]

RonCo: Is your test with scouting on or off?

With scouting off, that kind of distribution isn't very surprising and the distribution shouldn't be expected to mirror MLB.

With scouting on, I'd expect a wider dispersal across the rounds.

[neojonas]

You've been saying there's some improvement in the patches with development. What dev/aging settings are you finding to work best with patch 4 now out the door?

[RonCo]

Originally Posted by sterjs

RonCo: Is your test with scouting on or off?

With scouting off, that kind of distribution isn't very surprising and the distribution shouldn't be expected to mirror MLB.

With scouting on, I'd expect a wider dispersal across the rounds.

Scouting on or off doesn't make a difference. I wrote about this a bit ago when someone was complaining that scouting wasn't accurate enough. My supposition is the opposite--scouting is far too accurate by this measure.

[RonCo]

Originally Posted by neojonas

You've been saying there's some improvement in the patches with development. What dev/aging settings are you finding to work best with patch 4 now out the door?

I'm still playing with it. SD's thread has my rough guess right now, but I reserve the right to change my mind.

[emcgman]

Are you guys saying that the AI `decides' to develop players because of what round they were originally drafted in, as opposed to the player's actual potential that existed at the time of player creation?

Or, is it because the AI `knows' who has real potential, and drafts based on that knowledge?

Just a noob looking for clarification on this deep issue.

Thanks

[conception]

This is a bit concerning, as since my laptop is (hopefully temporarily) down I am using my desktop to form a new fictional league using Ronco's settings, which were radically different from the one's I used to use. No matter what I was planning on numerous test sims to test the right levels.

[RonCo]

No one really knows anything except Markus.

I _assume_ based on some element of experience that the AI drafts based on current talents.

I _think_ talent hits and bumps are random, though there exists some data to show that later round picks are a little more likely to get talent bumps than early picks. However, if you reduce the size of talent changes, you are, by definition, making it harder for a guy with lower talents (drafted in later rounds) to ever be very good.

Of course, we really shouldn't use the word "talent" or "potential" for this parameter, despite the fact that OOTP uses it as its label. That parameter, when compared to a player's current rating is better thought of as "today's growth direction and rate" rather than "talent/potential." Blah, blah, blah...

[neojonas]

Originally Posted by RonCo

I'm still playing with it. SD's thread has my rough guess right now, but I reserve the right to change my mind.

Are you talking about the post by goldenglove where he said:
Batter Aging: 1.750
Batter Development: .500
Pitcher Aging: 1.500
Pitcher Development: .500

I read other places like post #91 in that thread that you were thinking dev would need to be tweaked less, but goldenglove shows that the aging should be tweaked less.

Thanks for all you do with this part of the game. I always enjoy reading your posts and using your research in my games.

[silentrob]

When you veteran gamers come to a consensus on proper aging and development modifiers, will you post it here or another thread? I'm new to this version of the game (last one I owned was 6) and all the new options are a lot to tackle. I too appreciate all the research people on this board do to help dudes like me.

[SittingDuck]

Originally Posted by neojonas

I read other places like post #91 in that thread that you were thinking dev would need to be tweaked less, but goldenglove shows that the aging should be tweaked less.

For me, those aging parameters work well, but I have set development at 0.75 as I felt 0.50 was too slow.

[neojonas]

I've been playing around with these settings and finding them to work nicely. Basically I took RonCo's suggestions from when 9.0 came out and brought them all .25 closer to the default 1.0 values since I've been finding that the latest patch does seem to improve the development model. Any thoughts?

Batter Aging: 1.750
Batter Development: .750
Pitcher Aging: 1.500
Pitcher Development: .750

[The Wolf]

Sigh, and I was so hoping that developement was finally fixed.

Thanks, RonCo, for all your work.

[Greatgrampa]

Originally Posted by neojonas

I've been playing around with these settings and finding them to work nicely. Basically I took RonCo's suggestions from when 9.0 came out and brought them all .25 closer to the default 1.0 values since I've been finding that the latest patch does seem to improve the development model. Any thoughts?

Batter Aging: 1.750
Batter Development: .750
Pitcher Aging: 1.500
Pitcher Development: .750

Those figures are what I set mine at too. I think they are a good compromise.