Most REALISTIC Settings for MLB

[MizzouRah]

Originally Posted by Matt Arnold

The defaults I believe are quite good (I know a few years ago they seemed well off, but I know Markus did some tuning of them again this year too), but if you want to adjust them up or down, you're more than welcome to.

You can always try out with those values. Create a league, set the values what you want, and sim out 20 years and see if you have too many old guys sticking around, or guys are developing too quickly that they're not spending time in the minors, or whatever. Getting outside of the 0.65-1.5 range I would expect you might start seeing weird patterns develop, but if you're within that zone it should be fine.

What evaluation settings do you use?

Also, does anyone know the default MLB QS and fictional evaluation settings?

I've been using 50/30/15/5 ever since last years game and I'm seeing an oddity every so often, which I know is just part of a game being a game.. but I was thinking about making a change next season.

I figured a developer might have an insight to what settings would provide the best overall experience.

[PSUColonel]

He's going to say the default....I have to go to my boot camp class right now..but I will elaborate when I'm done. I too have been using 50/30/15/5 and have seen improved results. They key is to keep ratings high while maintaining a dissernable difference between current and past year, as well as last year and year teo

[PSUColonel]

Ok...the idea is to really stagger the numbers in the way you want. From what I have gathered (I could be wrong but this is what I have gathered) the numbers themselves aren't as important, as long as:

1. Ratings are the overwhelming factor in the equation
.
2. How you stagger the ratings becomes the next point. Currently, with 50/30/15/5 the current year is 20 points off from ratings, and 15 from last year. SO the question becomes not so much how much each year is valued, but how much more/less it's valued compared to other years.


The default is 65/20/10/5...

so let's say we start from here. let's start by going 60/20//15/5

The reason I started with last year, is I feel there is not enough difference in values between last year, and 2 years ago. But now, we are faced with the same dilemma between this year and last year, so now let's go:

55/25/15/5.....this to me might be the magic number.

This is not the default, but it takes stats into consideration by 10 percent more, while still staggering the numbers in a way where the AI should act logically.

[MizzouRah]

I like that last one.. what is default fictional? I think default MLB QS is just so someone doesn't get the game and complain about the AI shuffling players around... unless the AI sees true ratings, but I doubt it if you use normal scouting.

I am fighting for a playoff this year though.. so I'm a bit afraid to change them from 50/30/15/5.

My only issue right now is a player here and there being released with decent ratings and good current year stats and last year stats. (which might be something you just can't fully fix)

[adamj]

Originally Posted by MizzouRah

I like that last one.. what is default fictional? I think default MLB QS is just so someone doesn't get the game and complain about the AI shuffling players around...

Yes, the defaults for fictional leagues are different. They are: 30/50/15/5.

[drhay53]

I'm guessing that the player evaluation settings are effectively the weights on a weighted average. In that case you shouldn't think of '10% in between them' as the important number, but their ratio. So for instance, with the above 25/15/5 ... 25/15=1.67, so this year's stats is 67% higher weight than last year's stats. But 15/5=3.0 so last year's stats is 200% higher weight than 2 year's ago stats. Now, let's say you look at 20/10/5

20/10 = 2, 10/5 = 2, so basically, look at it like 1 year ago is twice as important as 2 years ago, and this year is twice as important as last year.

This so far is a bit more confusing-sounding than it really is. Just think of it like this: the number for ratings, set that to how important you want ratings to be. Do you want it to be 50%, 60% of the total? After that, break down the numbers so they add up to 100 in whatever way you like. If OOTP is doing it right it doesn't matter if they add up to 100, but whatever. So let's say you want 60 for ratings, and then to halve the weight for each previous year's stats. That would be approximately 60/23/12/6.

For the math inclined, the weighted average is just sum(weights * values) / sum(weights).

[PSUColonel]

Originally Posted by drhay53

I'm guessing that the player evaluation settings are effectively the weights on a weighted average. In that case you shouldn't think of '10% in between them' as the important number, but their ratio. So for instance, with the above 25/15/5 ... 25/15=1.67, so this year's stats is 67% higher weight than last year's stats. But 15/5=3.0 so last year's stats is 200% higher weight than 2 year's ago stats. Now, let's say you look at 20/10/5

20/10 = 2, 10/5 = 2, so basically, look at it like 1 year ago is twice as important as 2 years ago, and this year is twice as important as last year.

This so far is a bit more confusing-sounding than it really is. Just think of it like this: the number for ratings, set that to how important you want ratings to be. Do you want it to be 50%, 60% of the total? After that, break down the numbers so they add up to 100 in whatever way you like. If OOTP is doing it right it doesn't matter if they add up to 100, but whatever. So let's say you want 60 for ratings, and then to halve the weight for each previous year's stats. That would be approximately 60/23/12/6.

For the math inclined, the weighted average is just sum(weights * values) / sum(weights).

Right. Which means, theoretically, the default of 65/20/10/5 is almost the same as 55/25/15/5...the difference being:

1. Ratings are not quite relied on as heavily (maybe they should be as in the default)
2. Last year's stats are considered more important ratio wise compared to two years ago.


So 60/25/10/5 could be another possibility

[MizzouRah]

Interesting default fictional is 30/50/15/5

I use 50/30/15/5 - so the default fictional flips the first two numbers around and places more emphasis on current year stats than it does ratings. I might go with those fictional defaults next season.

Kind of makes sense.. as with scouting on.. you would want to rely on stats a bit more than just ratings. Although I do look at OVR/POT ratings first, then stats when I'm looking for a trade.

[drhay53]

Originally Posted by PSUColonel

Right. Which means, theoretically, the default of 65/20/10/5 is almost the same as 55/25/15/5...the difference being:

1. Ratings are not quite relied on as heavily (maybe they should be as in the default)
2. Last year's stats are considered more important ratio wise compared to two years ago.


So 60/25/10/5 could be another possibility

I kind of like anything with ratings above 60, personally. I will probably go with 60/25/12/3 (descending ratios between stats evaluations).

[PSUColonel]

I feel you can get some odd results with something in the line of 30/50/15/5

I just feel you really need more emphasis on ratings. I've been playing around with these a lot...and right now I feel 55/25/15/5 is the sweet spot. The trade AI has really been tightened up overt he last couple of updates, and I honestly feel you can get the most realistic and challenging experience using mostly defaults.

The only two settings I've changed is:

1. Switching from 65/20/10/5 to 55/25/15/5

2. Putting the injury setting at modern day as opposed to classic.

I am also running every league available, and have minor league service time limits enabled, along with roster size limits.

My minors are as follows:

AAA: 25 man roster; no service time limit
AA: 25 man roster; no service time limit
A+: 25 man roster; 6 years max service time
A: 25 man roster; 5 years max service time
A- : 30 man roster 4 years max service time
DSL : 35 man roster ; 4 years max service time
R: 35 man roster; 3 years max service time

I have seen very few issues with these settings. And I really think OOTP has turned the corner when it comes to AI roster management. I do have one suggestion for the development team to help even further though..which would be:

Players assigned to rehab assignments in the minors not be charged as a roster slot.

[PSUColonel]

I also set inflation at 3-5%

[NoOne]

as you weight stats more you increase % error due to sample sizes - even with "Low" scouting accuracy. i have not used a lower setting for scouting accuracy.

how you weight the years is mostly arbitrary except that it is probably best that more recent seasons are weighted more heavily for obvious reasons. current year stats are a smaller sample size, therefore we are less confident in what they tell us compared to a full season. i would argue that current year stats should be weighted heavily, but also roughly the same as last year's stats.

some players do take a total nosefive in a very short period of time... this is a very small percentage of players in a league. i'd focus more on how this ai evaluation setting affects a larger number of players, first.


the only thing i'd say about the minors is that they are a bit different than the RL counterparts, so you might want to deviate a little bit from what they do in RL, too.

e.g. R, DSL, Ariz are all the same level of competition in the game as far as development is concerned (stats can wildly vary game to game). Same with A / A+. in the game there is no difference between A+ and A.

*note* -- there is a difference between A and short A.

i chose to use the larger RL service time limits for all the rookie level leagues and A/A+ leagues. i also used 5yrs instead of 4 years for Short A. i did slap a 8or9yr limit on AA, would only change it if i saw problems with rehabbing players.

realistic inflation? based on avg salary in last 10 years, it's a bit shy of 4.5%. are the players going to a get a larger or smaller chunk of revenues going forward? you'd have to make a storyline for your leauge and fill in the blanks.

[BigRed75]

Originally Posted by PSUColonel

Players assigned to rehab assignments in the minors not be charged as a roster slot.

This is vitally important, and in my mind a top priority.

Basically, until this is done, I cannot play with service time limits. Which sucks, because I really want to. But I simply cannot risk having major league players released because they were too old/experienced for the league that they were sent on rehab.

[MizzouRah]

I'm sticking with 50/30/15/5 for now.. the few players I see released are being signed by other teams.. quickly.

[RubeBaker]

Hi everyone,

Thank you for the comments. Ever since I posted this way back in the spring, I've done a lot of random testing since then and I've come to the conclusion that default settings for AI Evaluation are the most realistic. I don't mess with the Player Development Ratings although I might tweak the Trade Settings ("Very Hard"/"Highly Favor Prospects" or "Hard"/"Neutral").

Playing Stats Only sounds like fun and something I really want to dive into sometime in the future, but my preference is for MLB instead of fictional. I definitely can see the appeal of Stats Only and The Wolf has made some excellent suggestions and points about it along with many others here.

I'm just bummed because I had completed the 2016 MLB season using default settings and it was so realistic, but a new patch came out and I started a new game. Ever since then, I've played with the settings and I haven't come close to the realism I had with the default so I'm sticking with default and QuickStart going forward.

[bly08]

One thing regarding trade settings is that if you place a lot of weight on ratings, 50+, trading on hard difficulty is essentially the same if not more difficult than very hard with more weight on stats.

[PSUColonel]

Originally Posted by bly08

One thing regarding trade settings is that if you place a lot of weight on ratings, 50+, trading on hard difficulty is essentially the same if not more difficult than very hard with more weight on stats.

Right...which is why in another thread, I am trying to explain that the default trade settings are very good right now. They think it's even too hard...but I think there is a perfect balance in using average difficulty with 55/25/15/5 AI evaluation.

[Klew1986]

Originally Posted by PSUColonel

I feel you can get some odd results with something in the line of 30/50/15/5

I just feel you really need more emphasis on ratings. I've been playing around with these a lot...and right now I feel 55/25/15/5 is the sweet spot. The trade AI has really been tightened up overt he last couple of updates, and I honestly feel you can get the most realistic and challenging experience using mostly defaults.

The only two settings I've changed is:

1. Switching from 65/20/10/5 to 55/25/15/5

2. Putting the injury setting at modern day as opposed to classic.

I am also running every league available, and have minor league service time limits enabled, along with roster size limits.

My minors are as follows:

AAA: 25 man roster; no service time limit
AA: 25 man roster; no service time limit
A+: 25 man roster; 6 years max service time
A: 25 man roster; 5 years max service time
A- : 30 man roster 4 years max service time
DSL : 35 man roster ; 4 years max service time
R: 35 man roster; 3 years max service time

I have seen very few issues with these settings. And I really think OOTP has turned the corner when it comes to AI roster management. I do have one suggestion for the development team to help even further though..which would be:

Players assigned to rehab assignments in the minors not be charged as a roster slot.

Very interesting idea with the minor league roster sizes and service time limits.

[PSUColonel]

I have also changed the prospect list to dynamic...it may be updated on the fly, but it works better by not including ranked prospects from International Leagues who are signed by a MLB team.

[bly08]

Originally Posted by PSUColonel

Right...which is why in another thread, I am trying to explain that the default trade settings are very good right now. They think it's even too hard...but I think there is a perfect balance in using average difficulty with 55/25/15/5 AI evaluation.

The way I see a heavy ratings weight is that the AI will essentially be judging the players the same way as you do. This applies even if you play with stats only, unless you ignore all scouting reports. For example, I play with ratings, and there's no case where a player's previous and 2 year old stats could have anywhere near the same influence on my judgement than his ratings and history of rating changes. I'd say my own human settings would be something like 70/25/5/0.

I just started a new game with the Marlins using your 55/25/15/5 with hard/neutral, and it's virtually impossible to fleece the AI even with scouting turned off (I use average). This is perhaps the most balanced trade setting combination I've used since starting OOTP, if not a bit on the difficult side. The AI will emphasize ratings regression in the same way and ignore short term trends. As to whether this is the most realistic, I would think not, but it's the most fair.

On realism, one of the biggest contracts on the Marlins is Chris Johnson, who finished second in the batting title race a few years back with the Braves then got paid. He then managed to produce negative WAR every year since. There are two OOTP interpretations for this, either he did indeed have good ratings but experienced an immediate dip via talent change, or the Braves placed much more weight on current year stats than what we do as OOTP players and normally use for AI eval settings. I would think the latter is the better explanation. So the most realistic setting would probably emphasize stats more, but then you can't allow yourself access to rating changes as comprehensive and immediate as the OOTP system, but then stats only doesn't really work for MLB.