Player Development Testing & Conclusion

[jcard]

Originally Posted by phenom

These boards are also loaded with people who won't give an ounce of aid unless you've supplied 149 years of detailed statistical analysis to the tenth decimal point with a 47-page slideshow and 16-page spreadsheet and 27 8x10 color glossy pictures with circles and arrows and a paragraph on the back of each one explaining what each one was highlighting every anomaly encountered.

Intentionally or not, there is definitely an apologies to the 1930’s Cardinals) a “Gaslight Gang” here.

[jcard]

Originally Posted by Rain King

Yeah, while I appreciate anyone doing extensive testing on this stuff. I think we continually see how difficult it is to test something well and produce useful conclusions.

I also think the "Standard Game" is a bad control. The players in that set are not created or developed via the OOTP engine, they are manipulated to try to get realistic 2025 results. It is essentially a custom save. I don't think it is reasonable to expect the game to level out in future seasons exactly at the levels that set was created.

Well, the game bills itself as a baseball simulation, and certainly highlights the MLB in its promotion. Moreover, the MLB is the most obvious target model. Unless there is something abnormal about the talent distribution in the 2025 MLB itself, why is it wrong to expect the player creation mechanism to produce talent that mirrors the initial player pool that has been “manipulated to try to get realistic 2025 results”?

[Rain King]

Originally Posted by jcard

Well, the game bills itself as a baseball simulation, and certainly highlights the MLB in its promotion. Moreover, the MLB is the most obvious target model. Unless there is something abnormal about the talent distribution in the 2025 MLB itself, why is it wrong to expect the player creation mechanism to produce talent that mirrors the initial player pool that has been “manipulated to try to get realistic 2025 results”?

I think you are underestimating how difficult/complicated matching that up would be considering how many manually created players we are talking about here (around 6,000 in the MLB game plus any other leagues such as the KBL).

Meanwhile, the player creation/development/aging algorithms have to take much more than the 2025 players into account. They have to look at years of statistical chaining, results of which aren't linear throughout the history of the game...and they have to determine where their "balance" is going to be and in addition to that make sure all of the other features work well with the base levels of each of those things (this is why you see new features take a while to get ironed out).

OOTP has always been a game that is pulling the development team in multiple directions and they only have all that much manpower to go around. Compromises have to be made along the way. Having all of these real players simply makes it harder to balance what the OP is highlighting here. They could create fewer "real" players or care less about how those players initially perform and this would probably be better. Is that what the majority would want though?

I think they do a great job considering, but if people are going to put things under this kind of microscope I just want it to be understood what we are looking at, which brings us back to the initial point. If the OP wants to truly look at how the game works/balances into future seasons, testing a fictional league that is created by the game code is a better "Control" for that.

[Pelican]

Whoa. I appreciate Dr. Satan's help and settings. I have used similar adjustments myself, in order to obtain a result more to my liking. I never looked at those changes are correcting or reflecting faults or weaknesses in the game. Rather, I appreciate the flexibility to manipulate those settings, if the "standard game" settings yield disappointing results.

Don't get me wrong. The "standard game" settings are important, certainly for first-time users. The rest of us have settings we have been using for years. I sure do. Through tinkering I have found adjustments to the default settings that work for me. I would encourage others to do the same.

[DrSatan]

Originally Posted by phenom

These boards are also loaded with people who won't give an ounce of aid unless you've supplied 149 years of detailed statistical analysis to the tenth decimal point with a 47-page slideshow and 16-page spreadsheet and 27 8x10 color glossy pictures with circles and arrows and a paragraph on the back of each one explaining what each one was highlighting every anomaly encountered.

Bhahahahaha! This gave me the giggles. These guys definitely exist, but to be fair, I've found people on this forum overwhelmingly helpful over the last 20+ years.

[DetroitStyle]

I can't really tell you what's realistic, but here's a snapshop of a random future year in one of my saves



This is the average OVR/POT for all MLB players who played at least 1 game in the pros during the current year followed by a count of how many players with a specific overall/potential.

What feels wrong is the sea of mediocrity. 65% of the league is at 40-45 overall. 70% is below 50/50. From my perspective it just gets really hard to distinguish a sea of "45" overall players from each other. Sure you can look for tools like maybe a 45 is a 65 power or maybe 70 defense.

But when 3 in every 4 players are just 45/40/45 or something it just becomes boring. Then you have the AI giving $10mil to one guy at 45/40/45 and $1mil to another at the same position and ratings. Like why? What's the difference? How do you even approach this in FA when nearly every player is (by the scouting numbers) the same.

I think DrSatans settings help expand that range as players get better sooner and fall off a cliff later. But I don't have a solution and again can't really tell you what a good or realistic distribution would look like.

[Rain King]

Originally Posted by DetroitStyle

IBut I don't have a solution and again can't really tell you what a good or realistic distribution would look like.

You are looking at one .

Seriously, that looks right about what I would expect given how the 20-80 scale is used in MLB.

Think about it this way. Assuming you are playing a 30-team league, that means this chart is accounting for 37+ players per team.

There are only 26 on a roster at a time. So, you are looking at 330(ish) players on your scale that shouldn't even be on a roster (if not for injuries and/or situational reasons).

That covers almost everyone 40 and below on your grid.

45's are your borderline MLB players. Backups, platoon players, non-back of the bullpen relievers, 5th starters, etc. Generally, guys who are major leaguers but their value is depth or situational.

50+ is where you get into the solid MLB players. 50 = average in terms of overall playing time, not in the number of players that are actually on a roster.

[kidd_05_u2]

Originally Posted by DetroitStyle

I can't really tell you what's realistic, but here's a snapshop of a random future year in one of my saves

Beta-like distribution. To me at least, it could not look any more perfect.

Originally Posted by DetroitStyle

What feels wrong is the sea of mediocrity. 65% of the league is at 40-45 overall.

65% of MLB batters had a WAR of 2.0 or less last season.
If you only look at WAR, you are not going to distinguish the majority of the players in the MLB.
If you only look at OVR, you are not going to distinguish the majority of the players in OOTP.

Originally Posted by DetroitStyle

But when 3 in every 4 players are just 45/40/45 or something it just becomes boring. Then you have the AI giving $10mil to one guy at 45/40/45 and $1mil to another at the same position and ratings. Like why? What's the difference? How do you even approach this in FA when nearly every player is (by the scouting numbers) the same.

Again, every player looks the same if you only look at OVR. If you play this game by evaluating players on their OVR only, yeah it's going to be boring because it's not the way you are supposed to be playing it.

[OutS|der]

Originally Posted by DetroitStyle

I can't really tell you what's realistic, but here's a snapshop of a random future year in one of my saves



This is the average OVR/POT for all MLB players who played at least 1 game in the pros during the current year followed by a count of how many players with a specific overall/potential.

What feels wrong is the sea of mediocrity. 65% of the league is at 40-45 overall. 70% is below 50/50. From my perspective it just gets really hard to distinguish a sea of "45" overall players from each other. Sure you can look for tools like maybe a 45 is a 65 power or maybe 70 defense.

But when 3 in every 4 players are just 45/40/45 or something it just becomes boring. Then you have the AI giving $10mil to one guy at 45/40/45 and $1mil to another at the same position and ratings. Like why? What's the difference? How do you even approach this in FA when nearly every player is (by the scouting numbers) the same.

I think DrSatans settings help expand that range as players get better sooner and fall off a cliff later. But I don't have a solution and again can't really tell you what a good or realistic distribution would look like.

This list is overpowered compared to RL. There should only at best 2 rated 80 overall, 70 is for the top 5. You have 20 players rated 70+
Same for potential, 19 players are projected to be the top 2 players in the league meaning 17 of them should never reach that potential.

So you're right it's wrong but not in the sense you think.

[DrSatan]

Originally Posted by OutS|der

This list is overpowered compared to RL. There should only at best 2 rated 80 overall, 70 is for the top 5. You have 20 players rated 70+
Same for potential, 19 players are projected to be the top 2 players in the league meaning 17 of them should never reach that potential.

So you're right it's wrong but not in the sense you think.

He didn't divide it up by pitchers/hitters, but I'm willing to bet a good chunk of the 70+ rated players are RP's.

[Rain King]

FYI, I simmed a fictional league to 2055 on the most recent patch and this is the distribution I get on Opening Day (includes players on the IL) after changing scouting to 100% Accurate. This is Current Overall.

80 - 13 (1 RP)
75 - 10 (2 RP)
70 - 18 (1 RP)
65 - 22 (2 RP)
60 - 49 (9 RP)
55 - 62 (9 RP)
50 - 156 (27 RP)
45 - 360 (83 RP)
40 - 246 (178 RP)
35 - 22 (11 RP)
30 - 3 (1 RP)
25 - 1 (0 RP)

[noooooo4499]

Isn't the 20-80 scale supposed to be normally distributed around the average starter? So 68 percent should be between 40-60, 16 percent each for 60+ and 40-, 2.5 percent each for 70+ and 30-, and 0.15% each for 20 and 80.

If anything, there seems to be to many players who are 75/80 judging by the data provided above. I think the devs just need to add a linear rating scale and default to that so people stop complaining about this and then only use this scale if the player toggles it on.

[jcard]

Originally Posted by noooooo4499

Isn't the 20-80 scale supposed to be normally distributed around the average starter? So 68 percent should be between 40-60, 16 percent each for 60+ and 40-, 2.5 percent each for 70+ and 30-, and 0.15% each for 20 and 80.

If anything, there seems to be to many players who are 75/80 judging by the data provided above. I think the devs just need to add a linear rating scale and default to that so people stop complaining about this and then only use this scale if the player toggles it on.

Thanks to people who posted their rating distributions. Two caveats:

1) Are these absolute or league-relative? This matters in comparing the distribution to those seen, for example, on MLB’s “Prospects” page (which are absolute-scale).

2) The flatter ratings highlight a tension between realism and gameplay. Perhaps dubiously, professional teams can at least feel that they are making personnel choices based on comparative criteria—statistics, physical tools, biometrics, psychological profile, etc. (with other factors such as balancing lineups for pitch-type and -zone preference in game-day roster decisions). Many of these types of data are actually or effectively absent in OOTP. In terms of user agency, making decisions between seas of anonymous 50/55/50/45/45 players can feel arbitrary.

[jcard]

Originally Posted by Rain King

I think you are underestimating how difficult/complicated matching that up would be considering how many manually created players we are talking about here (around 6,000 in the MLB game plus any other leagues such as the KBL).

Meanwhile, the player creation/development/aging algorithms have to take much more than the 2025 players into account. They have to look at years of statistical chaining, results of which aren't linear throughout the history of the game...and they have to determine where their "balance" is going to be and in addition to that make sure all of the other features work well with the base levels of each of those things (this is why you see new features take a while to get ironed out).

OOTP has always been a game that is pulling the development team in multiple directions and they only have all that much manpower to go around. Compromises have to be made along the way. Having all of these real players simply makes it harder to balance what the OP is highlighting here. They could create fewer "real" players or care less about how those players initially perform and this would probably be better. Is that what the majority would want though?

I think they do a great job considering, but if people are going to put things under this kind of microscope I just want it to be understood what we are looking at, which brings us back to the initial point. If the OP wants to truly look at how the game works/balances into future seasons, testing a fictional league that is created by the game code is a better "Control" for that.

I disagree, but nevertheless appreciate your considered words.

[LansdowneSt]

Originally Posted by jcard

2) The flatter ratings highlight a tension between realism and gameplay.

I think this is the crux of it. If you've been around long enough, you can recall plenty of threads with vocal complaints about the old 20-80 scale and calls to, if it is going to be in the game, to make it an accurate representation of the actual MLB scale. So, they did. But as a mechanic for a game played by people that are not necessarily intimately familiar with the scale, it does create a huge middle. That middle is MLB reality imo and its existence does compel one make decisions on other tools and clues the game gives you and that's all akin to real life... but for a gamer, it's much harder to make decisions.

I've always played on the 1-100 scale so the old discussions of how best to do the 20-80 scale in-game and strike a balance between realism and fun/ease never impacted me but having seen reactions as that scale has become more aligned with reality, it has been interesting. If I had a recommendation it would not be on the OOTPD side but the user side, just to try another scale and see if that works better for your personal enjoyment of the game. They put lots of scales in for lots of user types/play styles after all.

[kidd_05_u2]

Originally Posted by jcard

2) The flatter ratings highlight a tension between realism and gameplay. Perhaps dubiously, professional teams can at least feel that they are making personnel choices based on comparative criteria—statistics, physical tools, biometrics, psychological profile, etc. (with other factors such as balancing lineups for pitch-type and -zone preference in game-day roster decisions). Many of these types of data are actually or effectively absent in OOTP. In terms of user agency, making decisions between seas of anonymous 50/55/50/45/45 players can feel arbitrary.

This is why complaints about player evaluation being boring should not target the distribution of OVR ratings on the 20–80 scale, which works as intended, but rather the limited number of component ratings.

[DetroitStyle]

Originally Posted by jcard

In terms of user agency, making decisions between seas of anonymous 50/55/50/45/45 players can feel arbitrary.

This perfectly sums up how I feel.

[OutS|der]

Originally Posted by DetroitStyle

This perfectly sums up how I feel.

Welcome to reality. You need to stop looking at overall/potential and look at the individual ratings . Overall has never been a true representation of how good a player is, the formula for it changes every year. And yes I agree we could use with a lot more ratings to help distinguish players