Stats-based sliders for CPU

[nomo17k]

Re: Stats-based sliders for CPU

I don't know if it means anything or not, but the park factors for Yankee Stadium for 2010 were: Batting - 113, Pitching - 111 according to baseball-reference.com. (www.baseball-reference.com/teams/NYY/2010.shtml) There were also 223 total home runs hit at Yankee Stadium in 2010, with 8 being the most in one game and 0 in several games. The number of home runs combined in one game that appeared the most was 4.

For 2009, it's first season, it was Batting - 97, Pitching - 96.

The Yankees is kinda tough because they are offense without park factor is way off. I definitely feel the Yankee Stadium is crazy power friendly in the game, but I don't know it's coded to be that way. One thing we could do is to compute the park factors in the game, but we might not have enough sample size though I've played a few hundred games by now (still in April).

[nomo17k]

Re: Stats-based sliders for CPU

For completeness I post updates on my slider adjustments.

I was still experimenting with possible ways to generate comparable amount of walks to real life. My test till this point was only using the strike frequency slider. But from this set, I decided to get help from the pitcher consistency slider.

Everything at default (5) except:

Power: 4
Foul Frequency: 4
Pitcher Consistency: 3
Strike Frequency: 3
Pitcher Hook: 0
Fielding Error: 10
Throwing Error: 5
BR Steal Ability: 10
BR Steal Frequency: 8
Wind: 3

The stats after 35 games are attached. This set didn't quite generate the amount of walks I wanted.

[nomo17k]

Re: Stats-based sliders for CPU

Since the previous set didn't generate enough walks, I clicked down the pitcher consistency one more notch.

Everything at default (5) except:

Power: 4
Foul Frequency: 4
Pitcher Consistency: 2
Strike Frequency: 3
Pitcher Hook: 0
Fielding Error: 10
Throwing Error: 5
BR Steal Ability: 7
BR Steal Frequency: 8
Wind: 3

The stats with this set are attached. I also set BR Steal Ability to 7, hoping that will get SB success % closer to real-life (I've tested the number at 5 and 10, and computed what matches best the real life number there).

I like the number of walks generated. However, I didn't like the early count strike %, and also the overall strike % slightly below real life numbers. It's subtle but I felt the tiny difference gave me the feeling that the pitchers are pitching from behind too much.

[nomo17k]

Re: Stats-based sliders for CPU

In order not to force pitchers to pitch from behind too often, I increased the strike % slider by one, hoping the walk number stays roughly the same.

Everything at default (5) except:

Power: 4
Foul Frequency: 4
Solid Hit: 6
Pitcher Consistency: 2
Strike Frequency: 4
Pitcher Hook: 0
Fielding Error: 10
Throwing Error: 5
BR Steal Ability: 7
BR Steal Frequency: 8
Wind: 3

For this set I also edited all catchers "fielding ability" to 99, hoping this affects their ability to hold on to the pitch in the dirt more. DaiYoung in the CPU thread tested with editing reaction and blocking and that appears to work to reduce the number of WPs. But I think the root of the WP problem is that catchers drop the balls too often, and my thinking was that ability might be controlled by the fielding ability rating. After all, if somebody craps on your carpet in your last home party, you'd better remove the crap itself than covering it up with your towel or something.

Unfortunately maxing out catchers' fielding ability doesn't appear to reduce the number of WPs.

The stats with this set are attached. Fortunately, the walks are still there! The overall strike % is slightly down (0.6% compared to real life), but I like how the early count strike % is where I like it to be, about 0.59 - 0.60, which is my guess for real life number.

So I'm pretty happy with the combo of strike % at 4 and pitcher consistency at 2. I think what I did is to let more wildness take care of walks instead of forcing pitchers to pitch from behind more. Maybe that's a more natural way of how walks happen.

(I also remember I increased solid hit slider to 6 for this set; the reason is described in the next post.)

[nomo17k]

Re: Stats-based sliders for CPU

So as far as matching stats to real life, I should feel generally happy with the slider set (except for pitcher hook which I've been setting to zero just to see how it affects the game).

One improvement that I could make is the ground ball / fly ball ratio. The game generates a few more % more ground balls compared to real life, which leads to slightly inflated GDIP.

Seeing how the solid hit slider set to zero reduced the number of line drives substantially, I thought increasing that slider might convert some ground balls to line drives. So I increased the solid hit slider.


Everything at default (5) except:

Power: 4
Foul Frequency: 4
Solid Hit: 7
Pitcher Consistency: 2
Strike Frequency: 4
Pitcher Hook: 0
Fielding Error: 10
Throwing Error: 5
BR Steal Ability: 7
BR Steal Frequency: 8
Wind: 3

(All catchers' fielding abilities are 99.)

With this slider, the number of line drives became comparable to real life. However, it didn't appear to happen at the expense of ground balls. Unfortunately it was compensated by the reduced number of fly balls.

[Jpowser]

Re: Stats-based sliders for CPU

Just stumbled onto this great work!! One question for this though is are the sliders for both cpu and user or just cpu?

[nomo17k]

Re: Stats-based sliders for CPU

Just stumbled onto this great work!! One question for this though is are the sliders for both cpu and user or just cpu?

You should also check out:

http://www.operationsports.com/forum...d.php?t=472764

Several people have worked on CPU vs CPU sliders this year and got very decent results so far.

Yes, these CPU vs CPU sliders are tuned to yield realistic baseball stats in CPU vs CPU games. So these are great if you are looking for a slider set for running your franchise in the manage only mode or mostly just watching the games as a GM or something.

I personally think you can use the same set for user vs CPU games. I'd just use exactly the same set for the user slider setting. The reason is, *if* your play style is exactly the same as CPU, then your play should be influenced similar ways as CPUs do to get similar results.

But of course people aren't really CPU, so that's why people enjoy adjusting sliders for users playing at different levels. Also we don't exactly understand how the difficulty level changes gameplay, so one slider set probably isn't appropriate for people playing at different levels.

[Jpowser]

Re: Stats-based sliders for CPU

You should also check out:

http://www.operationsports.com/forum...d.php?t=472764

Several people have worked on CPU vs CPU sliders this year and got very decent results so far.

Yes, these CPU vs CPU sliders are tuned to yield realistic baseball stats in CPU vs CPU games. So these are great if you are looking for a slider set for running your franchise in the manage only mode or mostly just watching the games as a GM or something.

I personally think you can use the same set for user vs CPU games. I'd just use exactly the same set for the user slider setting. The reason is, *if* your play style is exactly the same as CPU, then your play should be influenced similar ways as CPUs do to get similar results.

But of course people aren't really CPU, so that's why people enjoy adjusting sliders for users playing at different levels. Also we don't exactly understand how the difficulty level changes gameplay, so one slider set probably isn't appropriate for people playing at different levels.

Just read that! You can obviously tell my every move! haha. I'm actually working on a personal slider set where i've gotten everything to my skill except USER and CPU walks, this is the only thing I can't get down, I've read the raise the consistency and control to raise more walks and lower one and keep the other higher yadayadaya, but I'm still searching through all the different sliders on the forum to find something that works! Sorry, rant over.

[nomo17k]

Re: Stats-based sliders for CPU

Just read that! You can obviously tell my every move! haha. I'm actually working on a personal slider set where i've gotten everything to my skill except USER and CPU walks, this is the only thing I can't get down, I've read the raise the consistency and control to raise more walks and lower one and keep the other higher yadayadaya, but I'm still searching through all the different sliders on the forum to find something that works! Sorry, rant over.

For walks, my personal advise is to use pitcher consistency 2 and strike frequency 4 for CPU at least. Of course after reading all the info you see there is no one way that is correct, but in my testing this combination has given me the most satisfaction in terms of the stats I've been keeping track.

[Jpowser]

Re: Stats-based sliders for CPU

For walks, my personal advise is to use pitcher consistency 2 and strike frequency 4 for CPU at least. Of course after reading all the info you see there is no one way that is correct, but in my testing this combination has given me the most satisfaction in terms of the stats I've been keeping track.

Haven't tried that yet, i'll give it a shot though thanks man.

[nomo17k]

Re: Stats-based sliders for CPU

I got a little curious about how the solid hit slider affects the stats I'm keeping track of. So I run 30 games with that slider set to 10.

Everything at default (5) except:

Power: 4
Foul Frequency: 4
Solid Hit: 10
Pitcher Consistency: 2
Strike Frequency: 4
Pitcher Hook: 0
Fielding Error: 10
Throwing Error: 5
BR Steal Ability: 7
BR Steal Frequency: 8
Wind: 3

(All catchers' fielding abilities are 99.)

The stats for this set is attached. The effect appears to be what we expect, increasing the slider produces more solid, line-drive hits. So with this, I'm producing about 1.5 more hits per game per team, which obviously affect batting average.

How does this affect extra base hits? Both 2B and 3B hits are up, but HRs are actually reduced. In terms of fractions, with solid hit at default, about 20.2 % of all hits are doubles, 2.7 % triples, and 11.1 % are HRs. With solid hit at 10, 21.2 % of all hits are doubles, 2.4 % triples, and 8.3 % are HRs. So the change is most significant in HRs.

The line drive % went up. The fly ball % and ground ball % went down. Now the ground ball % is pretty close to real life. However there are too many line drives and not enough fly balls. Maybe fly ball HRs are being converted into line drive hits.

I like the ground ball to fly ball ratio, but not at the cost of making other stats off balance.

Again, only 30 games, and I should point out that the identification of fly balls, line drives, and ground balls are sometimes dubious in the game, so these results may not be particularly accurate. But the trends seem reasonable.

[nomo17k]

Re: Stats-based sliders for CPU

On the effect of the pitcher hook slider.

So I've been running games with the pitcher hook at 5 default for roughly half the time, and at zero for the rest. I wanted to see how this affect pitcher usage.

For each game, I recorded three stats (1) the number of innings pitched by a starter, (2) the number of pitchers used, and (3) whether CG was pitched.

I didn't know an easy way to get the MLB ave for (1) during 2007 - 2010 seasons but I got numbers for the rest, on ave 3.92 pitchers are used in a game, and CG thrown per game (I know...) is 2.9 % (i.e., 2.9 % of the games results in CGs).

This is how the number compares. Numbers are in order: MLB ave, pitcher hook at 5, pitcher hook at 0.

Starter innings: ?, 5.78, 6.13
Pitchers used: 3.92, 4.06, 3.76
CG per game: 2.9%, 2.0%, 2.9%

So the slider does have some effects. By reducing pitcher hook, starters are kept in the game longer, could pitch more CGs, and less number of pitchers are used in general.

But the effect is kinda subtle, and also adjusting this slider doesn't prevent CPU from making weird moves (such as pulling closer after allowing just one run), so the mileage is limited.

[nomo17k]

Re: Stats-based sliders for CPU

Considering all the data gathered so far in the thread, the first "official" release of the slider set is this

This is the April 16 version.

Everything at default (5) except:

Power: 4
Foul Frequency: 4
Pitcher Consistency: 2
Strike Frequency: 4
Pitcher Hook: 2
Fielding Error: 10
BR Steal Ability: 8
BR Steal Frequency: 8
Wind: 3

Comments: The game is a bit HR friendly, so Power needed to be lowered. Foul Frequency was lowered to make batters swing through more (seen in swing & miss %). Pitcher Consistency was lowered to make pitchers a bit wider to make them issue more walks. Strike Frequency was lowered to slightly lower the overall and early count strike %. Pitcher Hook was lowered to make CPU use slightly less number of relievers and also to let starters pitch slightly longer. Fielding Error is maxed out in an attempt to match the fielding error number IRL but still not quite enough. BR Steal Ability is at 8 to reduce the SB success % which is a bit high at default. BR Steal Frequency at 8 makes steals happen at realistic rate. Wind is set at 3 to roughly match the overall distribution of windy days in games to real life number (game data taken just for early April), slightly less windy days because the wind effect appears a bit exaggerated.

All these are done to match the MLB ave during 2007 - 2010.

[nomo17k]

Re: Stats-based sliders for CPU

I just realized there are so many typos and grammar mistakes in my posts. Sorry about that. I just wanted to post all the updates and thoughts for completeness and not waste time writing poetic proses.

[jripper09]

Re: Stats-based sliders for CPU

Nomo i'm impressed beyond belief and who cares about the typos i'm sure i'll have a few. finding the right pitcher sliders for cpu is very difficult tried youre sliders last night me (cards) cpu rsox) becket was tough as usual couldnt draw a walk but due to his high confidence with alot of strikeouts early. but even then he was still missing alot not robo pitcher! i havent tried adjusting the fouls and solid hits slider at 5 however will do this later this evening. also have not tried lowering the pitcher hook will do so as well. hopefully these will work for user vs cpu i'm encouraged though.

what sliders do you think i should use for user at hof level pitcher to make it similar to your'e cpu adjustments? i have mine human control 4, consistancy 3. thanks

if in franchise mode which pregame setting do i use to get cpu vs cpu where i can manage the game? i think there were 2 modes can't remember which ones though. like i said earlier i'm very impressed with all the stats and slider adjustments you have made and thank you for sharing these.