Poor talent in recent Draft Years

[BeancheBlanco]

Originally Posted by The Wolf

I've played through fifteen years with one OOTP 14 league. The draft classes were pathetic, and their careers were pathetic.

the pitchers or the hitters?

[Lukas Berger]

Originally Posted by HH20xx convert

3-Continue the old 'flame-out' factor despite these changes. We don't need predictability in outcomes.

Was just reading through the thread again and noticed I never really responded to this. It's a very good point.

I'm open to correction from someone more knowledgeable on this but I'm pretty sure that none of the changes made to create lower rated draftable players have created any more predictability. At least not in the sense that late round picks will never turn into MLB players or that all first round picks will.

That's still controlled by the TCR ratings, which I don't believe have changed.

And even if you think things are too predictable for some reason then there's an easy solution to that. Raise the TCR's, as wisely suggested above by SirMichaelJordan.

So I'm pretty sure that this is no more (or less) an issue than it was (or wasn't) in 13.

[Walsh06]

I was unaware of a lot of this but I definitely prefer the new system. It seems more realistic and much less annoying. I'd much rather have the joy of a gem rather than annoyance of a bust. And I'd also like to see wolfs ratings 15 years in the future. He's so adamant about it I'd like to see.

[JMDurron]

Originally Posted by Walsh06

I was unaware of a lot of this but I definitely prefer the new system. It seems more realistic and much less annoying. I'd much rather have the joy of a gem rather than annoyance of a bust. And I'd also like to see wolfs ratings 15 years in the future. He's so adamant about it I'd like to see.

Although, if luksaberger is right, the ratings themselves won't tell us anything. We'd need to see if the statistical outcomes of his league are appropriate or not. If instead of the average player overall rating being 70, it might now be 40, but lukasberger's argument is that the statistical outcomes should still be roughly equivalent.

Interestingly, in my massive sample size of ONE DRAFT CLASS, the pitching talent still seemed to be present, it was the position player talent pool that was quite shallow.

[Cinnamon J. Scudworth]

Originally Posted by lukasberger

Was just reading through the thread again and noticed I never really responded to this. It's a very good point.

I'm open to correction from someone more knowledgeable on this but I'm pretty sure that none of the changes made to create lower rated draftable players have created any more predictability. At least not in the sense that late round picks will never turn into MLB players or that all first round picks will.

I think the predictability issue was a response to the statement that Righty Groove quoted from Markus earlier in this thread, which seems to imply more consistency in talent retention.

[Righty Groove]

Originally Posted by Cinnamon J. Scudworth

I think the predictability issue was a response to the statement that Righty Groove quoted from Markus earlier in this thread, which seems to imply more consistency in talent retention.

That is what I got from Markus comments aswell. That the top rated draft-elegible players were less likely to tale huge hits in their ratings and fail miserably.

So the results, i.e., the number of top/good/average players from each draft remained the same, just more predictable. The top players are more rare but retain their talent better.

IMO, it's total crap. And kills the possibility of "fixing" this by upping the PCM's since it will result in more talent getting into the league. Maybe, MAYBE changing the PCM's and also upping the TCR (I alwayd had it at 100, default) could work, but I don't feel like messing with my Dynasty, going since OOTP10.

So, back to 13 I went. Saddly.

[Fyrestorm3]

Forward: A lot of large images here, so I'm hotlinking them all instead of actually posting them. Most of them are boring. Sorry, but this was about documentation more than interesting results.

-----

Okay, I did some investigative reporting. It's taken me three whole days, but here's the results of a few tests I ran. First, I created a 24-team fictional league in OOTP13 (leaving all default settings intact), and simmed forward 10 years. I saved that league and imported it to OOTP14. Each league created its own new draft pool:

OOTP13's Pool
OOTP14's Pool

That is a grand total of 5 players rated above 3 stars in 14, and 31 rated higher than 3 stars in 13 (they don't even fit on one screenshot). It's plainly obvious that the perceived talent in the draft pools created by OOTP14 has been drastically reduced. But whether you like it or not, we've already established that. Now for the interesting part.

I saved the '13 game with the draft pool already revealed, and then imported THAT game into '14.

OOTP13's Pool in OOTP14

There are some significant variances from 13's version of this pool, but the perceived talent distribution is about the same (27 players over 3 stars).

Lastly, for comparison, I started a new game in OOTP14, simmed forward a couple of years, and then checked out THAT draft class. The same default settings are in effect; the only possible difference would be a variation in scouting (since accuracy is set to Normal by default):

OOTP14 "Pure" Draft Class

In terms of player potential, it's almost identical to the one created in the imported league, which flies in the face of many claims in this thread.

Next, it was time to draft. I took over the same team in both the OOTP13 league and the OOTP14 imported league, and drafted the same six players: four of them rated higher in '13 than in '14, one the inverse, and one rated identical. Interestingly, despite the differences in ratings, their demands were all similar - barring Gabriel Pagan, who for some reason cut his demand in half in OOTP14, despite being rated pretty much exactly the same.

Demands in '13
Demands in '14

Ratings Comparisons:
Angel Acosta
Jonathan Foley
David Jenkins
Gabriel Pagan
Mark Patrick
Jesus Velazquez

For some reason, they all behaved EXACTLY the same way when it came to signing - each one signed with me on the same day in both versions. Including Jenkins, who held out so long I was forced to edit him onto the team in both games. I then simmed until the next year, so the worldwide re-scout would take effect, and to give the players enough time in my system for the inevitable Ratings drop in OOTP13 to kick in. As expected, in OOTP14, the changes in ratings were very minimal, if there were any at all. In OOTP13, Acosta dropped down to 2.5 stars (despite still being rated highly in '14) and Patrick dropped half a star. Everyone else was still rated the same.

Finally, I simmed forward, stopping every couple of years to check the progress of my draftees. Five years later, and the difference in their ratings between the games was small enough to be chalked up to variations in scouting. By the by, the draft classes continued to be seemingly "weaker" in OOTP14, but that shouldn't be a surprise. After 15 years, I took a look back on their careers, and their numbers were remarkably similar across versions.

David Jenkins never made it to the bigs in either game. A total flame-out in OOTP13, but right in line with his ratings in '14.

Mark Patrick appeared in less than 35 games and was wholly unimpressive in both versions. Another OOTP13 disappointment.

Gabriel Pagan was the star of that draft class. In both games, he went on to become his team's #1 pitcher. This was totally in line with his potential rating in both games.

Jonathan Foley is the first major difference between the games. Despite being rated higher in OOTP13, in that game, he never saw any big-league action. Meanwhile, in OOTP14, where he was projected with a half-star potential, he became a serviceable bench player. So, yet another '13 flame-out, but a '14 surprise.

Jesus Velazquez did not make the cut in either game. That's a plus for OOTP13, which correctly predicted him as being a half-star player. He was only 1.5 stars in OOTP14, but that's still a drop-off.

Angel Acosta had one at-bat in OOTP13, and struggled to make his way in '14. That's a disappointment on both ends.

Okay, so in a small sample size, the players pretty much performed to their true abilities all around, regardless of potential ratings. But they WERE created in OOTP13, and as we've seen, this is pretty much an issue with player creation more than anything else. So, onto the next experiment. In both games, the year was 2037. I simmed forward to that year's draft, and made an Excel sheet documenting the potential ratings for every draftee in both games. I was actually surprised to see a relative abundance of talent in '14's Pool, though 17 players rated higher than 3 stars still pales in comparision to '13's 29.

Draft Excel Sheet - You can only see part of it, but you get the idea.

As a total side note, some highlights in the generated names were John Farrell, Pedro Martinez, Zander Wilberforce, Lorenzo Lorenzo, and Alberto & Norberto Delgado, who I decided had to be twins (sadly, neither made the bigs).

I also checked the pools with 100% accurate scouting turned on, and there were minimal differences. I opted to go with Normal scouting, since that's the default setting and the one I feel will apply to more people. I then shortlisted every player, and simmed forward 15 years once again. At that point, I catalogued every player's career and compared it to their previous Potential ratings. I gave players a "+" if I felt they exceeded their Potential ratings, and a "-" if I felt they were a disappointment.

For the record, there were four possible ratings I could give players. Either they never reached the majors, they were a Scrub (reached the bigs but rarely played), they became an average, everyday player, or they became an All-Star (this required at least 2 All-Star selections in their career).

Catalogued Results

Out of OOTP13's 156 draftees, 20 were a surprise (12.8%), and 26 failed to meet expectations (16.7%). The draft class saw 8 All-Stars (5.1%), and 31 decent players (19.9%). There were 89 who failed to reach the bigs (57.1%), with 28 Scrubs (18%). Only 2 of the overachievers were actually All-Stars, and both had been projected as 3.5 stars. Meanwhile, 14 players who were projected to be 4-stars or above turned out to be Scrubs or Washes.

Out of 144 draftees in OOTP14, 16 of them performed above expectations (11.1%), while 7 underperformed (4.8%). The draft produced 9 All-Stars (6.3%), and 20 average major leaguers (13.9%). 76 failed to reach the majors (52.8%), and the 39 remaining were Scrubs (27.1%). Six of the surprises became All-Stars, 3 of which had previously been rated 2-stars or below. And only 1 player was projected to be great and wound up a Scrub.

-----

So what did I learn through these exercises? A couple of things. First, I saw no difference in the draft classes created in OOTP14 between leagues that have been imported from OOTP13 and those that haven't. This is notably contrary to what many people have posted here (with evidence), and I find that intriguing.

Second, it's not an issue with scouting (since talent distribution remained the same in the same draft between the games, even if some players were rated differently). It's an issue with player creation. This was already brought up in this thread, but now it's backed up.

Third, there is an obvious difference in how prospects develop between the two versions. There is far more consistency between Potential ratings and ultimate careers in OOTP14. That should sound obvious, because it's what everyone's been discussing, but again, now there's hard evidence.

Fourth and finally, there does not, in my opinion, seem to be an issue with overall career arcs, as Wolf has so vehemently argued. About 25% of the draftees in OOTP13 went on to become serviceable big leaguers, and 20% did in OOTP14. I feel like 5% is an insignificant variance, small enough to be chalked up to a slightly better or worse draft class by year.

I am making no judgment on any of this. All I am doing is documenting a few experiments to provide some examples in a debate that is sorely lacking them. Yeah, there's more that should be done here. I only sampled six players and one draft class; it would be nice to do this multiple times to ensure the results aren't a fluke, but it was exhaustive just getting this much done. If someone else wants to pick up where I left off, be my guest.

[Righty Groove]

Great work, Fyrestorm3. Thank you for getting this done and sharing with us.

I would have taken scouting off on an experiment like this, though. Scouting at normal accuracy (default) is usually not very accurate. So it's a variable easily dismissed in a controlled enviroment study like this.

Either way, thanks and great work. It seems like that markus quote is holding true after all.

[Fyrestorm3]

Originally Posted by Righty Groove

I would have taken scouting off on an experiment like this, though. Scouting at normal accuracy (default) is usually not very accurate. So it's a variable easily dismissed in a controlled enviroment study like this.

I was extremely tempted to take scouting off, and went back and forth on that a bunch of times. Ultimately, I decided to leave it on because I was attempting to reproduce the effects that people have been debating about, and I would assume that most of them play with scouting on.

In hindsight, though, I should have at least done one separate test without scouting, to ensure that it wasn't a variable.

[Lukas Berger]

Originally Posted by Fyrestorm3

Third, there is an obvious difference in how prospects develop between the two versions. There is far more consistency between Potential ratings and ultimate careers in OOTP14. That should sound obvious, because it's what everyone's been discussing, but again, now there's hard evidence.

Great stuff Fyrestorm. Just fantastic.

This is the biggest point imo. As much as I've defended the new system, gotta say that I don't like this.

Pretty sure this will be discussed on the beta forums quite soon

At least the fix is easy. Turn up the TCR ratings.

[cblacker]

Originally Posted by lukasberger

Great stuff Fyrestorm. Just fantastic.

This is the biggest point imo. As much as I've defended the new system, gotta say that I don't like this.

Pretty sure this will be discussed on the beta forums quite soon

At least the fix is easy. Turn up the TCR ratings.

Don't mean to argue semantics here, but I would call tuning the TCR more of a 'work-around' than a fix since it's going to change the development of all players, not just players in the draft.

I would like to see a bit of a compromise between the OOTP 13 and OOTP 14 draft class generation methods. Reducing the number of 5 star players that your scout sees in the draft is fine, but there should be more 2.5 to 3.5 star players IMO. When you get to the 5th round of the draft, for example, your chances of finding a ML level player are small as you pointed out, but in real life do teams picking in the 5th round actually look at all the players left and only see scrubs? This is pure speculation, but it would seem to me that scouts would see at least a couple of guys that have 3 star potential even if the chances are slim that they will actual fulfill that potential.

With a guy like Pujols, obviously no team saw a future superstar or even a decent ML player, otherwise he would have been drafted much higher, but somebody somewhere saw something and made enough noise to get him drafted. I'm dreaming a bit here, but how about a bit of an overhaul of the scouting system, at least for the draft, where you have more scouts (area scouts, cross-checkers, etc.) so that when you get to the lower rounds, maybe you still only see a bunch of half star and one star scrubs, but you could get some little message about a player saying something like "our area scout really likes this guy, but other scouts disagree."

[VanillaGorilla]

Originally Posted by Fyrestorm3

Third, there is an obvious difference in how prospects develop between the two versions. There is far more consistency between Potential ratings and ultimate careers in OOTP14. That should sound obvious, because it's what everyone's been discussing, but again, now there's hard evidence.

Fourth and finally, there does not, in my opinion, seem to be an issue with overall career arcs, as Wolf has so vehemently argued. About 25% of the draftees in OOTP13 went on to become serviceable big leaguers, and 20% did in OOTP14. I feel like 5% is an insignificant variance, small enough to be chalked up to a slightly better or worse draft class by year.

First off, GREAT stuff. I haven't gone over your actual data (and it is soooo cool you put all of this together for review!), just your abstract and summary post, here.

------

These two quoted points seem to be contradictory.

If "Player Development" is altered, then so should the career arc, but in the fourth point you say the career arcs are the same.

In point three you are classifying "Player Development" in relation to scouted potential, but (if I am reading this correctly) the players are actually coming out the same, so they are developing precisely the same way. All that has changed is the expectancy of how an individual player might develop based on on the v 13 potential ratings vs the v 14 potential ratings.

Correct me if I have misread your post, but if the players wind up with similar performance between 13 and 14, then that points to the player development being constant.

The signing behavior being identical is a huge tell, imo. Even though the ratings are different from v13 to v14, the player behavior remains constant, which indicates that the underlying data of the draftees relative to one another is the same. If the difference was in player development, the players would have had dissimilar signing expectations (at least ONCE!) and there would have been more variance in their individual career arcs as v14 aged/developed their careers.

All that has changed here is that the scouts are more conservative and only giving Dave Winfield/Bo Jackson scouting reports as often as a Winfield or Jackson enters the draft, not every single season 10 times over.

[Lukas Berger]

Originally Posted by cblacker

Don't mean to argue semantics here, but I would call tuning the TCR more of a 'work-around' than a fix since it's going to change the development of all players, not just players in the draft.

Pretty sure that the TCR doesn't affect players when they're in the draft, the talent changes don't start till they're in the minors. So if the players from the draft are developing more consistently in accord with their ratings then all the players in the minors will be too. So changing the TCR does actually "fix" the "problem", since it's a problem that wouldn't simply be confined to recently drafted players.

[Lukas Berger]

Originally Posted by cblacker

I would like to see a bit of a compromise between the OOTP 13 and OOTP 14 draft class generation methods. Reducing the number of 5 star players that your scout sees in the draft is fine, but there should be more 2.5 to 3.5 star players IMO. When you get to the 5th round of the draft, for example, your chances of finding a ML level player are small as you pointed out, but in real life do teams picking in the 5th round actually look at all the players left and only see scrubs? This is pure speculation, but it would seem to me that scouts would see at least a couple of guys that have 3 star potential even if the chances are slim that they will actual fulfill that potential.

With a guy like Pujols, obviously no team saw a future superstar or even a decent ML player, otherwise he would have been drafted much higher, but somebody somewhere saw something and made enough noise to get him drafted. I'm dreaming a bit here, but how about a bit of an overhaul of the scouting system, at least for the draft, where you have more scouts (area scouts, cross-checkers, etc.) so that when you get to the lower rounds, maybe you still only see a bunch of half star and one star scrubs, but you could get some little message about a player saying something like "our area scout really likes this guy, but other scouts disagree."

I agree with a lot of this, you make some good points.

To enlarge on some of your points:

What folks are saying about lower round players looking like scrubs is exactly right if you only look at the star ratings or the overall ratings.

If you look at the real ratings in the editor, the guys that are being generated for the 5th round and even lower right now aren't really scrubs. Some of them actually have really nice potentials.

So I honestly think the problem here that's causing concern has more to do with the star ratings not working real well.

They tend to do a really bad job with players that project to be decent MLB backups or borderline MLB players that could be useful or even good if they get some small potential boosts. Right now these guys just seem to get 1* or '20' ratings most of the time.

The overall/star ratings don't differentiate between these guys and true no hopers at all right now.

So I'd suggest that maybe the star and overall ratings should be fine tuned a bit. I don't think there's any problem with the actual ratings though.

[cblacker]

Originally Posted by lukasberger

Pretty sure that the TCR doesn't affect players when they're in the draft, the talent changes don't start till they're in the minors. So if the players from the draft are developing more consistently in accord with their ratings then all the players in the minors will be too. So changing the TCR does actually "fix" the "problem", since it's a problem that wouldn't simply be confined to recently drafted players.

I guess I'm not quite understanding, then. What I think you are saying is that, as of right now, there's not enough unpredictability in OOTP 14 draft classes. Players develop more or less along the lines your scout thinks that they will (as opposed to OOTP 13 where there were many more 'busts'). But if I increase the TCR, then, sure my draft classes will be more unpredictable, but the development of my 30+ year old superstars, for example, will also be more unpredictable. Will they go on to have 4 or 5 more solid seasons, or are they going to fall off a ratings cliff? And maybe I don't want that.

I'm thinking specifically of a long thread form OOTP 13 forums discussing the aging and development modifiers that would most closely represent real life MLB career arcs and, as I remember, TCR was part of that. So if I change my TCR, then my entire aging and developmental curve is now potentially a little out of whack. So I guess I just don't see it as a good way to get more unpredictability in the draft class.

[Fyrestorm3]

Originally Posted by VanillaGorilla

These two quoted points seem to be contradictory.

If "Player Development" is altered, then so should the career arc, but in the fourth point you say the career arcs are the same.

In point three you are classifying "Player Development" in relation to scouted potential, but (if I am reading this correctly) the players are actually coming out the same, so they are developing precisely the same way. All that has changed is the expectancy of how an individual player might develop based on on the v 13 potential ratings vs the v 14 potential ratings.

Correct me if I have misread your post, but if the players wind up with similar performance between 13 and 14, then that points to the player development being constant.

When I talked about player development differences, I was really referring to HOW the players developed. There was consistency across the board in terms of overall careers (~20% that become serviceable major leaguers), but how they arrived there was different. In OOTP13, almost 30% of the draftees failed to develop according to their potential ratings. They either overperformed or underperformed, while the other 70% lived up to their billing. In OOTP14, that percentage dropped to 16%, pretty much half of what it was in '13.

That's what I was referring to in point #3. Point #4 was about the draft class as a whole. I get where you're coming from, though; when you look at the big picture, both drafts developed the same. But on an individual basis, development was drastically different.

[VanillaGorilla]

OK, I read you correctly.

Originally Posted by Fyrestorm3

They either overperformed or underperformed, while the other 70% lived up to their billing. In OOTP14, that percentage dropped to 16%, pretty much half of what it was in '13.

I come back to the point that the players are the same, just the scouting reports are different. You wouldn't take a set of players, in either v13 or v14, scout them with 'very low' and scout a cloned pool 'very high' and then say that the "player development" was different because there was more variance in the group scouted 'very low' than the one scouted 'very high' and perhaps to "fix" this we need to mess around with TCR, correct?

The difference is not in the players, but merely in how the players are presented to the user, be it between using very low vs very high scouting, or be it between v13 and v14.

[Fyrestorm3]

Originally Posted by VanillaGorilla

The difference is not in the players, but merely in how the players are presented to the user, be it between using very low vs very high scouting, or be it between v13 and v14.

Well yeah, isn't that what's been debated here? That's what I was testing, anyway.

EDIT: So yeah, okay, maybe "player development" wasn't the right term to use.

[RchW]

Originally Posted by cblacker

I'm thinking specifically of a long thread form OOTP 13 forums discussing the aging and development modifiers that would most closely represent real life MLB career arcs and, as I remember, TCR was part of that. So if I change my TCR, then my entire aging and developmental curve is now potentially a little out of whack. So I guess I just don't see it as a good way to get more unpredictability in the draft class.

I'm quite confident that TCR does nothing to the global aging and development outcomes. It's a zero sum setting. TCR determines the randomness of player talent changes only; not current ratings. The only way I can describe it is to draw or imagine a smooth line representing any individual players development and aging vs time, then superimpose on that curve a jagged line. The frequency of that jagged line would represent TCR. I guess that the amplitude is the amount of talent change since the manual only says "significant". We know that not every talent change causes a current ratings change but when TCR is normal or high you may get lucky and have a perfectly reasonable/normal current ratings boost happen at a time when talent is randomly higher. The inverse is also true. When TCR is low players will develop and age like the original smooth line ie with little variability. Markus obviously has data that allows him to make this feature available. On that basis we should assume that the default (100) setting is something close to the real life variability found in current baseball.

Bottom line no matter what the TCR setting, a given setting for aging and player development should provide similar populations of players by age.

I could be wrong.

[JMDurron]

Originally Posted by VanillaGorilla

The difference is not in the players, but merely in how the players are presented to the user, be it between using very low vs very high scouting, or be it between v13 and v14.

Ok, my sample size is only one fresh class in OOTP 14, but I don't use scouting when I'm checking out the draft class. The draft class generated by OOTP 14 featured significantly worse, as defined by in-editor potential ratings, talent compared to any class I've seen in OOTP 13. The difference IS in the players, and I had thought that lukasberger had admitted as much earlier in the thread. The difference was much more significant with the positional players than with the pitchers, for what that's worth.