Stats-Based CPU Sliders [MLB 15 Version]

[Heroesandvillains]

Re: Stats-Based CPU Sliders [MLB 15 Version]

1. IMO, using mayo on a burger is pure laziness.

2. IMO, french fries are so 1980s... c'mon, it's 2015 now. Time to go to Chipotle...

My impression so far is that SB- and fielding error-related sliders likely need similar treatments as in the past. It's too early to tell, and I dread the thought of re-discovering the BR Ability slider having its effect reversed, but then again who would buy The Show without that slider being reversed?

Heathen!

[manooly]

Re: Stats-Based CPU Sliders [MLB 15 Version]

Boy, based on your collected data so far the default settings with pitch speed @ 10 are looking pretty good.

I'm also assuming that you are using the default rosters that came with the game correct? Thanks!

[Ebpmd]

Re: Stats-Based CPU Sliders [MLB 15 Version]

Does it matter what hitting/pitching level (all star, pro, etc...) you use?

[kcsam]

Re: Stats-Based CPU Sliders [MLB 15 Version]

No. Nomo has said in the past that it's "thought" the system runs off of All-Star in CPU vs CPU games. But either way, it doesn't matter in CPU vs CPU games what setting you use.

I play on All-Star hitting to make things as realistic as possible when playing RTTS.

[jts88]

Re: Stats-Based CPU Sliders [MLB 15 Version]

I'm not much of a contributor but have been a huge fan of your sliders since I made the switch to MoM a few years back. Just wanted to say thank you for doing these again. Looking forward to your set. Thanks again

[ralphieboy11]

Re: Stats-Based CPU Sliders [MLB 15 Version]

Looking at the results so far, they are very similar to the ones I was getting on default.

One oddity is that fly balls are actually above their normal percentage compared to ground balls, yet we are still getting a lot of ground balls into double plays.

One observation from watching the games play out is that the ball seems to come off the bat with much more velocity this year. I'm seeing more hard hit balls put into play. Many of these hard hit grounders seem to make it easier to turn the double play, even against the speedier runners.

[nomo17k]

Re: Stats-Based CPU Sliders [MLB 15 Version]

Just when I thought offense is down, this happened:

[knich]

Re: Stats-Based CPU Sliders [MLB 15 Version]

Just when I thought offense is down, this happened:

keep playing against NY...the stats will get even better. :wink:

[nomo17k]

Re: Stats-Based CPU Sliders [MLB 15 Version]

I'm not much of a contributor but have been a huge fan of your sliders since I made the switch to MoM a few years back. Just wanted to say thank you for doing these again. Looking forward to your set. Thanks again

The thanks should go to the people who actually keep improving the game... I'm only re-affirming how some parts of the game is very interesting in realism department.

Boy, based on your collected data so far the default settings with pitch speed @ 10 are looking pretty good.

I'm also assuming that you are using the default rosters that came with the game correct? Thanks!

I do this with the same setting every year so I've gotten lazy and haven't thoroughly been explaining my procedure. I'll write it up soon.

Does it matter what hitting/pitching level (all star, pro, etc...) you use?

No. Nomo has said in the past that it's "thought" the system runs off of All-Star in CPU vs CPU games. But either way, it doesn't matter in CPU vs CPU games what setting you use.

I play on All-Star hitting to make things as realistic as possible when playing RTTS.

It is actually not clear if CPU vs. CPU games run on All-star difficulty level. All I know is that when I change difficulty levels in those games, I don't see any effects showing up that depend on the difficulty setting. At least it worked like that before... Exactly at what level CPU gets locked in in CPU vs. CPU is not clear... but it probably makes sense that the default difficulty level is set to All-star, since it's right in the middle.

Looking at the results so far, they are very similar to the ones I was getting on default.

One oddity is that fly balls are actually above their normal percentage compared to ground balls, yet we are still getting a lot of ground balls into double plays.

One observation from watching the games play out is that the ball seems to come off the bat with much more velocity this year. I'm seeing more hard hit balls put into play. Many of these hard hit grounders seem to make it easier to turn the double play, even against the speedier runners.

Since categorizing FB/LD/GB isn't really done exactly similarly with where I am getting the real-life MLB numbers (fangraphs), I think we cannot expect them to exactly line up. Given that BABIP is a bit low, however, it's possible that the game might be producing more fly balls than they should be... but then again the line between a fly ball and line drive isn't exactly well defined so some of the discrepancy might be due to some difference in methods of batted ball categorization....

As for GIDPs, I tend to agree with your impression... though I did have a chance to chat with B Ma about hit variety, and he did mention that it has been updated with new data, and the game does seem to produce fairly realistic hit variety now (though it wasn't that bad before). This may be in ball physics category, but I still feel especially "poor hit" variety (including ground balls) can still be greater. Also CPU fielders don't make as many mistakes as human players. That's another factor.

[nomo17k]

Re: Stats-Based CPU Sliders [MLB 15 Version]

Testing methodology

Here are the gameplay settings:

Spoiler

I did not take the screenshots, but I'm using a season mode with the default SCEA roster that comes with the game. It's on a 30-team control mode with disabling CPU trades, legends, etc. Basically the idea is test CPU vs. CPU game results in the default settings.

I'm playing every single game in the season in CPU vs. CPU mode. At each slider settings, I try to get 75 games in, in order to rotate all five pitchers in every team (15 x 5 = 75 games).

The initial slider set is all set at default (5, where min = 0 and max = 10). However, I set Pitch Speed to 10 for aesthetic reasons (want the pitch speed to be closest to real-life).

With each sliders set, I record all kinds of stats from box score, game log, and pitcher analysis screens. The numbers are recorded into spreadsheet:

https://docs.google.com/spreadsheets...it?usp=sharing

Each tab uses different slider settings, and slider settings for that tab can be viewed in the A-1 cell.

Mainly, what I look at are various "per X" statistics. In-game numbers are compared to recent MLB numbers (most of them are weighted by 50-25-25% for 014, 2013, 2012, just as attribute ratings are computed).

Based on my current understanding of what each slider does, I try to make adjustments to align in-game numbers more closely to MLB numbers, where there are "significant" deviations.

[ralphieboy11]

Re: Stats-Based CPU Sliders [MLB 15 Version]

Just a personal note on fielder reaction this year...

In the past I've personally used the fielder reaction slider in cpu vs. cpu games to tweak batting averages here and there. I found it a good way to increase batting average or extra base hits without messing with the batter/pitcher interaction. I ran an initial 75 game default test myself and batting averages and offense was a little lower than I wanted. In my next test I decided to drop fielder reaction down to 4. What I found is that fielder reaction is very volatile this year. Over a 50 game stretch I saw batting averages rise .025 points with that drop to 4. While this could have been an abnormality and you have to be careful about reading too much into small samples, I think I saw enough to deduce that fielder reaction was the root cause. I ran 10 more games with it back at 5 and changing nothing else. Sure enough the batting averages and scores dropped right back to what I was initially seeing and what Nomo has in his spreadsheet.

Previously I noted that the ball seems to be coming off the bat with more velocity on average this year. I believe the devs even mentioned in one of the pre-release streams that the trajectory of some batted balls had changed. I think if you combine those factors with the fielder reaction slider you can get some vastly different results. I'm going to stay away from that slider for now and leave it at default(5).

Anyway, just something I wanted to pass along. This thread is always one of my favorites every year.

[nomo17k]

Re: Stats-Based CPU Sliders [MLB 15 Version]

Nearing the end of the initial 75 game set with default sliders, and here are some fun stats.

When I compute BA, OBP, and OPS for the first 25 games (pitched mostly by aces and No. 2's) and the last 25 games (pitched mostly by No. 4's and 5's), they look like this:

                  BA  OBP  OPS
First 25 Games  .232 .300 .667
Last 25 Games   .267 .338 .755
All             .250 .317 .719

This is kind of interesting because if you take MLB starting pitchers from 2014 and take the top 30 and bottom 30 players in BA against, their median is about .231 and .265, respectively. Kinda shows the ability gap is fairly realistically reproduced so far.

This also shows that it is important to sufficiently rotate pitchers of varying abilities. Results can change this much based on who the starting pitchers are.

[nomo17k]

Re: Stats-Based CPU Sliders [MLB 15 Version]

Just a personal note on fielder reaction this year...

In the past I've personally used the fielder reaction slider in cpu vs. cpu games to tweak batting averages here and there. I found it a good way to increase batting average or extra base hits without messing with the batter/pitcher interaction. I ran an initial 75 game default test myself and batting averages and offense was a little lower than I wanted. In my next test I decided to drop fielder reaction down to 4. What I found is that fielder reaction is very volatile this year. Over a 50 game stretch I saw batting averages rise .025 points with that drop to 4. While this could have been an abnormality and you have to be careful about reading too much into small samples, I think I saw enough to deduce that fielder reaction was the root cause. I ran 10 more games with it back at 5 and changing nothing else. Sure enough the batting averages and scores dropped right back to what I was initially seeing and what Nomo has in his spreadsheet.

Previously I noted that the ball seems to be coming off the bat with more velocity on average this year. I believe the devs even mentioned in one of the pre-release streams that the trajectory of some batted balls had changed. I think if you combine those factors with the fielder reaction slider you can get some vastly different results. I'm going to stay away from that slider for now and leave it at default(5).

Anyway, just something I wanted to pass along. This thread is always one of my favorites every year.

Thanks for the warning. Dropping Fielder Reaction is exactly what I am thinking of doing next, given that it is probably the most sensible way to simply increase singles and reduce GIDPs...

I'm not sure how you are rotating starting pitchers, but I just posted a quick analysis on how batting average can vary depending on who starts the game. Depending on starting pitchers, it is possible that much of the .025 difference in batting average could be caused by pitching match-ups.

To be honest, I haven't really touched Fielder Reaction sliders before, as how it affects the game seems obvious but how much isn't really clear.

--

Anyways, the numbers look fairly good at default sliders... but if I could I want to reduce GIDPs and increase BABIP a bit (by increasing singles mainly).

To my surprise, for the first time, the walk rate actually exceeds the real-life MLB number!! Not sure if I bother try to make it come down a bit...

[ralphieboy11]

Re: Stats-Based CPU Sliders [MLB 15 Version]

Thanks for the warning. Dropping Fielder Reaction is exactly what I am thinking of doing next, given that it is probably the most sensible way to simply increase singles and reduce GIDPs...

I'm not sure how you are rotating starting pitchers, but I just posted a quick analysis on how batting average can vary depending on who starts the game. Depending on starting pitchers, it is possible that much of the .025 difference in batting average could be caused by pitching match-ups.

To be honest, I haven't really touched Fielder Reaction sliders before, as how it affects the game seems obvious but how much isn't really clear.

--

Anyways, the numbers look fairly good at default sliders... but if I could I want to reduce GIDPs and increase BABIP a bit (by increasing singles mainly).

To my surprise, for the first time, the walk rate actually exceeds the real-life MLB number!! Not sure if I bother try to make it come down a bit...

Considering what you've seen between top starters and bottom starters maybe you will see something different with fielder reaction than I did. In the first couple weeks after release I had access to two PS4s and two versions of the game. The problem with some of that initial testing is that I started one in season mode and the other in franchise mode. By the time I reached 75 games on default I had not really reached the bottom starters yet. As expected, the offensive numbers were a little low. Since then, I've tried to do a better job of mixing things up. While I think it was a little more balanced, my fielder reaction test was only over 50 games, so it's possible something else could have affected it(starting pitchers).

While I'm really enjoying the changes this year, right now I'm sort of at a loss as to where I want to go next. I wouldn't mind seeing the walks go down a tad and the strike outs go up. Not really sure how to achieve that when offense is sort of down already.

Just from an eye test I am going to bump infielder arm strength down. I was hoping that might mean a few more infield hits or less double plays. However, from initial tests it doesn't seem to make much difference.

[nomo17k]

Re: Stats-Based CPU Sliders [MLB 15 Version]

Considering what you've seen between top starters and bottom starters maybe you will see something different with fielder reaction than I did. In the first couple weeks after release I had access to two PS4s and two versions of the game. The problem with some of that initial testing is that I started one in season mode and the other in franchise mode. By the time I reached 75 games on default I had not really reached the bottom starters yet. As expected, the offensive numbers were a little low. Since then, I've tried to do a better job of mixing things up. While I think it was a little more balanced, my fielder reaction test was only over 50 games, so it's possible something else could have affected it(starting pitchers).

While I'm really enjoying the changes this year, right now I'm sort of at a loss as to where I want to go next. I wouldn't mind seeing the walks go down a tad and the strike outs go up. Not really sure how to achieve that when offense is sort of down already.

Just from an eye test I am going to bump infielder arm strength down. I was hoping that might mean a few more infield hits or less double plays. However, from initial tests it doesn't seem to make much difference.

Funny I was actually thinking of getting a second PS4 just for testing purpose, haha... it will accelerate the testing. With two consoles, you could still rotate pitchers properly if you played games from different days on the same schedules.

I think I'm basically making the same adjustments as you are trying. With default settings playing pretty well, I don't really see too many different ways to change things that are necessary to fix more annoying issues (e.g., GIDPs).