Stats-Based CPU Sliders [MLB 12 Version]

[nomo17k]

Re: Stats-Based CPU Sliders [MLB 12 Version]

I'm impressed by the effort you've put into these. Kudos!

Is this set available for download from the Vault?

Yes, best to browse by name "Stats-Based ...." or by PSN name "nomo17k"

[mbarr403]

Re: Stats-Based CPU Sliders [MLB 12 Version]

Yes, best to browse by name "Stats-Based ...." or by PSN name "nomo17k"

Thank you sir.

[ralphieboy11]

Re: Stats-Based CPU Sliders [MLB 12 Version]

I tried a few games with Pitcher Control at 8 this afternoon and tonight. I even ran a couple while I went out and ran some errands to get some stats.

Obviously, it's tough to tell from a just a few games. From what I did see, I think you can really tell a difference by just moving Pitcher Control from 7 to 8.

Nomo, you are exactly right about it aleviating some of the issues at the start of the game and when relievers enter. They don't seem to be so rattled, and come in firing first pitch strikes in many instances.

While I do like that part, I think you were probably dead on about settling at 7 for Pitcher Control, in relation to the other sliders in this set. 8 is probably going to drive walks down too much and pump up strike outs.

I don't know why I bother with this stuff when you've done all the work. I guess I need to see with my own eyes. Ha.

I do wish they would just take this 'confidence' thing and throw it in the trash. Just let the pitchers perform to their ratings and stats. It will all even out in the end.

[nomo17k]

Re: Stats-Based CPU Sliders [MLB 12 Version]

I tried a few games with Pitcher Control at 8 this afternoon and tonight. I even ran a couple while I went out and ran some errands to get some stats.

Obviously, it's tough to tell from a just a few games. From what I did see, I think you can really tell a difference by just moving Pitcher Control from 7 to 8.

Nomo, you are exactly right about it aleviating some of the issues at the start of the game and when relievers enter. They don't seem to be so rattled, and come in firing first pitch strikes in many instances.

While I do like that part, I think you were probably dead on about settling at 7 for Pitcher Control, in relation to the other sliders in this set. 8 is probably going to drive walks down too much and pump up strike outs.

I don't know why I bother with this stuff when you've done all the work. I guess I need to see with my own eyes. Ha.

I do wish they would just take this 'confidence' thing and throw it in the trash. Just let the pitchers perform to their ratings and stats. It will all even out in the end.

I doubt just one click would change things so much, but I did set Pitcher Control = 7 to avoid issues you are concerned about... Too high, and you'll see less walks and slightly more strikeouts. Too low, pitchers have problem building confidence up. My main reason for using 7, however, was it's effect on the batting average. From 5 to 10, the batting average dropped by about .015. I wanted to reduce it by about .005 - .010, so I chose 7...


I actually like pitcher confidence system. Without it, I think you'll see too many consistent performances... but I think logic still needs some tweaking.

The issue is that it's a bit too sensitive right now, meaning it's a bit too easy to rattle a pitcher just coming in by taking a lot of pitches (since he usually has command issues). On the other hand, once he gets on his groove with high confidence, it's a bit too hard to bring him back to the ground.

I think increasing Pitcher Confidence can alleviate the first issue (don't know what to do with the latter), so that's a hint as to how it can be improved...

[ralphieboy11]

Re: Stats-Based CPU Sliders [MLB 12 Version]

I think the pitcher confidence system could be really good for the game if it was improved. The increasing of Pitcher Confidence that you mentioned would be a good start. The way it is now, it does need improvement.

I've played many table top and computer baseball games that have developed rules where a pitcher will sometimes come into a game with "good stuff" or "bad stuff". These rules were probably implemented to prevent too many consistent performances.

Some games just have systems where after a pitcher gives up a certain amount of runs or hits, it has been proven he has "bad stuff" for the day. His ratings go down after reaching a certain hit or run threshold, and at that point it is probably best to put in a new pitcher.

Many believe without using a system like that, you still will see varying performances for pitchers. So it's a debate that's been going on for years.

[nomo17k]

Re: Stats-Based CPU Sliders [MLB 12 Version]

I think the pitcher confidence system could be really good for the game if it was improved. The increasing of Pitcher Confidence that you mentioned would be a good start. The way it is now, it does need improvement.

I've played many table top and computer baseball games that have developed rules where a pitcher will sometimes come into a game with "good stuff" or "bad stuff". These rules were probably implemented to prevent too many consistent performances.

Some games just have systems where after a pitcher gives up a certain amount of runs or hits, it has been proven he has "bad stuff" for the day. His ratings go down after reaching a certain hit or run threshold, and at that point it is probably best to put in a new pitcher.

Many believe without using a system like that, you still will see varying performances for pitchers. So it's a debate that's been going on for years.

There would be natural variations even without some system like pitcher confidence as implemented in the Show... Just as a fair coin flipping can string together a series of 5 tails at times, even though at any given time the probability of getting either side is equal at 50%.

I guess conceptually I like (some sort of) confidence system that could change the tide in the game depending on the situations (say human elements if you will). Otherwise your best strategy always will be the one guided by ratings and probabilities. (Without thinking about energy) you'll always be better off staying with Verlander than using Dotel to relieve him.

And without confidence system, you can almost always count on Verlander to "regress to the mean" and perform normally later in the game, even after a very very shaky start in the 1st. So his "off" days cannot be simulated. For that reason I've never pulled my starters early in computer simulation games (which do not have some sort of confidence system). Because I know the starter was just very unlucky to start off, there's nothing wrong with his "stuff" and the game is actually pretty much ignorant of what happened previously when the new inning comes up.

That is, those events are independent... just as the coin wouldn't have any memory of what it resulted previously in coin flipping game. The game doesn't remember what happens without some sort of confidence system.

So that's the reason why I'm still in favor of confidence system, even though the logic could still see improvements.

[ralphieboy11]

Re: Stats-Based CPU Sliders [MLB 12 Version]

I agree with everything you wrote. When I was thinking about it before I read your post, I actually thought about Verlander and why would you ever pull him from the game(if he wasn't tired).

A confidence system would definitely be needed in that situation.

I'm still having lots of fun with the cpu vs cpu stuff. I started a game with the Nationals and the Cubs. I watched the Nats hit a leadoff home run to start the game. I kept the game running, and left the house for a bit. When I got back home the game was over. Cubs won 2-1. That leadoff home run was all the Nats could muster. Good stuff.

[steviegolfballs]

Re: Stats-Based CPU Sliders [MLB 12 Version]

Through 25 games (50 team games) of my franchise, mostly #1 and #2 pitchers so far using the following sliders:

conact 4
power 5
timing 6
foul fr. 4
solid hits 4
st. stam. 7
rp. stam. 5
control 5
consist. 4
str. frq. 5
mgr. hook 6
pickoffs 0
pit. speed 10
fld. error 7
thr. error 5
field run speed 3
field reaction 9
field arm str. 2
base run sp. 6
steal ability 4
steal freq. 8
wind 0
injury (off)

at bats 1684 33.7 (34.1)
runs 208 4.16 (4.39)
hits 427 8.54 (8.78)
RBI 199 3.98 (4.18)
BB 162 3.24 (3.21)
K's 338 6.76 (7.04)
doubles 88 1.76 (1.75)
triples 12 .24 (.19)
home runs 45 .90 (.96)
sac hits 9 .18 (.34)
sac fly 9 .18 (.27)
hit by pitch 8 .16 (.32)
double play 52 1.04 (.75)
errors 24 .48 (.62)
st. base 27 .54 (.64)
caught st. 15 .30 (.25)

POSITIVES: hits and runs are lower than avg. which they should be facing better pitchers, expected to level out as I move into the #3's and #4's

XBH ratio is excellent, triples got high only because in game #24 Boston had three at Comerica Park. Park dimensions play a huge role in all XBH

Walks are great!

wonderful variety of game scores, so far I have had a 1-0 game, an 11-9 game and many in between.

Two complete games so far, only one was a shutout. furthermore, there have been two other shutouts that were not complete games which I find excellent because many MLB shutouts still have the set-up man and closer involved. score of game plays the major role here.

NEGATIVES: No new news on DP's or HBP's

strikeouts started out great and even a bit high but have hit a lull, hopefully just a quirky stretch of games.

stolen bases are a bit concerning but not way out of whack given the sample size.

errors aren't too bad but they seem to come in bunches

GAME 26 NOW PLAYING!

[ralphieboy11]

Re: Stats-Based CPU Sliders [MLB 12 Version]

Thanks for sharing Stevie. Looking good.

That makes me feel better about my early walks issue.

I'm 64 games in, so 128 game sample.

Walks at 3.62, so they came down a bit for me after a previous post. I had a lot at the start.

I'm seeing the same things you are regarding stolen bases. 0.55 stolen bases and 0.20 caught stealing for me. That's with the same settings you have.

I may need to eventually bump up Base Steal frequency to 9 like Nomo has. It kind of annoys me when everyone runs on the first pitch though.

[steviegolfballs]

Re: Stats-Based CPU Sliders [MLB 12 Version]

Same thing for me on the steals. Putting that slider at nine give me too many attempts at third base too.

[nomo17k]

Re: Stats-Based CPU Sliders [MLB 12 Version]

Instead of increasing Steal Attempt, perhaps you might decrease Steal Ability by one more click? That'll improve steal success rate slightly. Not sure if AI is coded that way but if he knows there's a greater chance for successful steal, he might attempt more.... not confirmed though, and that's not likely what will happen.

I say this because I see down in AA, there are a lot fewer steal attempts, even though the success rate is very high (above 80%). I'm using the latest slide set (Apr. 15). I think the best way to adjust steal attempts really is to do rating edit, and probably to increase BR Aggressiveness a bit more for those who should attempt steal more often.

But in general I'm not in favor of editing ratings, since it's just too tedious for the amount of improvement you'd get.

[steviegolfballs]

Re: Stats-Based CPU Sliders [MLB 12 Version]

When you bumped pitcher control to 7 did you notice an increase in singles and doubles?

[nomo17k]

Re: Stats-Based CPU Sliders [MLB 12 Version]

When you bumped pitcher control to 7 did you notice an increase in singles and doubles?

Not that I can see... I think Pitcher Control increases weak hits that tend to end up in outs but once pitches are solidly hit, the game produces the same amount of long balls in right proportions.

[nomo17k]

Re: Stats-Based CPU Sliders [MLB 12 Version]

This isn't of any interest for those who just play MLB games, but I ran 75 games at the AA level to see how the stats come out.

This is with the Apr. 15 set.

Things of note are:

-- significantly less HRs, probably because Power ratings for AA players are too low to generate HRs.

-- significantly less steal attempts. Maybe the Speed/BR ratings are too low for AI managers to give them green light?

-- significantly more fielding errors. Understandable...

-- significantly less walks, slightly more strikeouts. This perhaps is partly due to lower vision/discipline ratings? I see AA hitters chase slightly more than MLB hitters. But I also notice AA pitchers throw a lot more strikes compared to MLB pitchers. Are they actually programmed to throw more strikes? Or AA hitters swing at more pitches, making them strikes? It's hard to tell.

[seanjeezy]

Re: Stats-Based CPU Sliders [MLB 12 Version]

This isn't of any interest for those who just play MLB games, but I ran 75 games at the AA level to see how the stats come out.

This is with the Apr. 15 set.

Things of note are:

-- significantly less HRs, probably because Power ratings for AA players are too low to generate HRs.

-- significantly less steal attempts. Maybe the Speed/BR ratings are too low for AI managers to give them green light?

-- significantly more fielding errors. Understandable...

-- significantly less walks, slightly more strikeouts. This perhaps is partly due to lower vision/discipline ratings? I see AA hitters chase slightly more than MLB hitters. But I also notice AA pitchers throw a lot more strikes compared to MLB pitchers. Are they actually programmed to throw more strikes? Or AA hitters swing at more pitches, making them strikes? It's hard to tell.

Yeah, the problem lies within the ratings. Just took a look at the OSFM 1.5 set that was put out and it has Billy Hamilton with 92 speed and 48/48 steal ability/frequency. I know that this is done to prevent him from being too good too soon, but if a player has superior tools he shouldn't be penalized for them at the minor league level, they should be more pronounced IMO. Players don't mysteriously gain 20 points in power or speed over their careers, its more likely that they lose it as they grow older...

Maybe you should come up with a set of MiLB sliders Nomo, it won't alleviate the low numbers in franchise sims, but it would help those of us who actually like to play minor league games more than major league ones...

Edit:

Are you using OSFM for these tests Nomo? The high strike percentages are probably a result of individual pitch control being set at MLB levels since those ratings hardly progress. Don't know if this affects the CPU in any way, but the pitch repertoires are all out of whack as well, too much mph separation between fastballs and offspeed pitches.