Stats-Based CPU Sliders [MLB 13 Version]

[ralphieboy11]

Re: Stats-Based CPU Sliders [MLB 13 Version]

To clarify, are you playing with Pitcher Control @ 3, Strike Frequency @ 3 and Pitcher Consistency @ 6?

I've been going with Pitcher Control 5, Strike Frequency 3, and Pitcher Consistency 6.

I just recently moved Pitcher Consistency from 5 to 6 to try and combat what I felt were a few too HBPs and WPs. Probably not enough of a change to see anything drastic, but they have come down slightly in the games so far.

[geisterhome]

Re: Stats-Based CPU Sliders [MLB 13 Version]

Would it make sense to use your Fielding and Throwing Error sliders also for HUM vs CPU sliders?

[nomo17k]

Re: Stats-Based CPU Sliders [MLB 13 Version]

Would it make sense to use your Fielding and Throwing Error sliders also for HUM vs CPU sliders?

I think so... if you are controlling fielders, your skills will matter more than ratings obviously, but the setting should work fine where dice roll elements come in... and some elements of the game always depend on dice roll.

[steviegolfballs]

Re: Stats-Based CPU Sliders [MLB 13 Version]

My latest results thru 60 games. This was done in franchise mode and takes me through the #4 starters. All sliders set to 5 unless otherwise noted:

Foul frequency 6
Solid hits 4
Starter stamina 6
Strike frequency 3
Pickoffs 0
Fielding errors 6
Throwing errors 4
Fielder run speed 3
Reaction 10
Wind 0

AB 33.74
R 4.67
H 8.52
RBI 4.37
BB 3.07
K 7.03
2B 1.55
3B .15
HR .96
SAC .21
SF .25
HBP .41
DP .74
E .66
SB .6
CS .19

Analysis:
Everything looks pretty good except I am producing too many runs. My next test is moving the pitcher consistency slider to 7 and see if I can reduce hit batters and wild pitches, thus reducing scoring opportunities without reducing hits. My numbers are low but they climb throughout each 15 game stretch as I move through pitching rotations. Now that I am getting to the #5 starters I would expect the see a slight rise in offense.

This test was more so to prove that lowering the solid hit slider would raise the HR total. From what I can gather, this slider changes the trajectory of the ball coming off the bat, turning line drives into a few extra choppers but also a few extra fly balls. This increase in fly balls has raised the HR total by about .1 per team game played. I am concerned that raising the consistency slider will hurt my walk total and could possibly produce more big hits if the pitches are more centered but I am encouraged that Ralph has raised his to 6 without a significant change in offensive totals. I moved it to 7 to see if it produces an extreme total, I am guessing that it will end up at 6.

Thanks to all the same cast of characters, as you can see my sliders are are built off the information shared by others.

Happy testing! Steve

[nomo17k]

Re: Stats-Based CPU Sliders [MLB 13 Version]

My latest results thru 60 games. This was done in franchise mode and takes me through the #4 starters. All sliders set to 5 unless otherwise noted:

Foul frequency 6
Solid hits 4
Starter stamina 6
Strike frequency 3
Pickoffs 0
Fielding errors 6
Throwing errors 4
Fielder run speed 3
Reaction 10
Wind 0

AB 33.74
R 4.67
H 8.52
RBI 4.37
BB 3.07
K 7.03
2B 1.55
3B .15
HR .96
SAC .21
SF .25
HBP .41
DP .74
E .66
SB .6
CS .19

Analysis:
Everything looks pretty good except I am producing too many runs. My next test is moving the pitcher consistency slider to 7 and see if I can reduce hit batters and wild pitches, thus reducing scoring opportunities without reducing hits. My numbers are low but they climb throughout each 15 game stretch as I move through pitching rotations. Now that I am getting to the #5 starters I would expect the see a slight rise in offense.

This test was more so to prove that lowering the solid hit slider would raise the HR total. From what I can gather, this slider changes the trajectory of the ball coming off the bat, turning line drives into a few extra choppers but also a few extra fly balls. This increase in fly balls has raised the HR total by about .1 per team game played. I am concerned that raising the consistency slider will hurt my walk total and could possibly produce more big hits if the pitches are more centered but I am encouraged that Ralph has raised his to 6 without a significant change in offensive totals. I moved it to 7 to see if it produces an extreme total, I am guessing that it will end up at 6.

Thanks to all the same cast of characters, as you can see my sliders are are built off the information shared by others.

Happy testing! Steve

The numbers look good enough to me, even runs. I actually doubt you 0.1 HR difference you see is entirely due to Solid Hit slider adjustment. A small number like 0.1 HR difference can easily happen due to random fluctuations, so I bet if you repeat the same run a few times (I'm not suggesting that you do though), you'll draw a different conclusion. Your overall adjustments are well done enough at this point now that I think you won't be seeing the effects you'd like to see. It's just the nature of small differences and the limited sample size.

[nomo17k]

Re: Stats-Based CPU Sliders [MLB 13 Version]

I'm playing around with Contact slider and seeing something interesting.

Those who know my insanity regarding CPU sliders know that I love base-on-balls so much that I actually start looking for ones in my local bar scene when I'm out drinking. I'd marry them if I could.

I have been exploring the ways to induce more walks by lowering Pitcher Control slider, which hasn't been particularly successful. Walks and strike % have changed, but not drastically so, even if I lower the slider down to 1, which is where I'm testing right now. Given this post by Brian@SCEA, it's not really surprising. After all, if a pitcher aims the black but still miss it within a circle, half the time it will be a strike and the other half a ball. That remains the same even if the inaccuracy diameter is increased. Similarly, if a pitcher aims at one of the corners of the strike zone, 1/4 of the time the pitch will land in the strike zone, 3/4 ball, regardless of the size of the circle of inaccuracy (assuming the radius don't go over the size of the strike zone, that is).

So I wanted some change in CPU hitters' approach, to see if it ends up inducing more walks... and so I lowered CPU Contact down to 1... very low just to magnify the effects, if any.

Expectedly, the batting average went down. Strikeouts increased (CPU is whiffing like crazy). But the walks also actually increased... which is surprising.

I'm still 20 games in with this Contact = 1 experiment, so I don't say my numbers are robust yet... since I'm seeing unexpected and interesting result, I'll run the experiment a bit longer to see what happens.

But if lowering Contact actually increases walks, isn't that rather surprising and interesting? I'm thinking of the reasons as to why this is happening. Why does reducing CPU discipline (by lowering Contact) increase walks? Shouldn't it be reverse?

By the way, O-Swing % isn't really increasing, so by lowering Contact CPU isn't necessarily chasing... (which is kinda disappointing)... so a lot of action is still happening within the strike zone... which I hope expands in the future iteration of The Show.

[capa]

Re: Stats-Based CPU Sliders [MLB 13 Version]

I have a TON of respect for what you are doing both this year and in year's past. I follow it very closely and I actually use your recommended sliders in both this year's and last year's game. Your analysis and discipline in tracking stats and tendencies is amazing.

I have always thought that trying to perfect the sliders to get realistic player stats and tendencies was only half the equation. I always have thought that the real way to get what we are all after is to standardize and tweak player ratings/attributes. For every game I have owned that has always seemed to be the key. I remember doing so in NFL2K5, many versions of Madden, High Heat, etc. The games always played better and more realistically when the attributes were analyzed and a players ratings were changed as required to get the right balance of look and feel on the field as well as stats comparable to current NFL or MLB stats.

I think the reason this approach is shunned is because of the work involved in making global edits without a global editor! You have to manually edit each player in dozens of categories potentially. If there was a pc editor that existed like the one developed for 2K5 or even last year's Madden, this approach would be a lot more feasible. I do think it is the correct approach in conjunction with tweaking sliders.

Just curious to your thoughts...

C

[Bobhead]

Re: Stats-Based CPU Sliders [MLB 13 Version]

I'm still 20 games in with this Contact = 1 experiment, so I don't say my numbers are robust yet... since I'm seeing unexpected and interesting result, I'll run the experiment a bit longer to see what happens.

But if lowering Contact actually increases walks, isn't that rather surprising and interesting? I'm thinking of the reasons as to why this is happening. Why does reducing CPU discipline (by lowering Contact) increase walks? Shouldn't it be reverse?

By the way, O-Swing % isn't really increasing, so by lowering Contact CPU isn't necessarily chasing... (which is kinda disappointing)... so a lot of action is still happening within the strike zone... which I hope expands in the future iteration of The Show.

I've seen the same thing with regards to walks. I think the decrease in Plate Vision causes the CPU to just swing at less pitches (I guess because it's more "confused"?). But it's not something I've really documented

I would strongly disagree with the notion that O-Swing doesn't increase when Contact is decreased though. I haven't specifically isolated it but my O-Swing has gone up every year after lowering CPU Contact, and it has done the same this year.

[nomo17k]

Re: Stats-Based CPU Sliders [MLB 13 Version]

I've seen the same thing with regards to walks. I think the decrease in Plate Vision causes the CPU to just swing at less pitches (I guess because it's more "confused"?). But it's not something I've really documented

By far the most visible change when Contact is reduced is that CPU whiffs more. Swing & miss % goes from about 20% to 27%, and the foul / strike % goes down as well. So lowering Contact is indeed effectively lowering Plate Vision.

There used to be separate Vision slider which got combined into Contact, so that's not surprising.... but I'm still not sure Vision is the attribute that affects AI strategy for swinging... because wouldn't it make more sense to think Plate Discipline controls the swing %?

I would strongly disagree with the notion that O-Swing doesn't increase when Contact is decreased though. I haven't specifically isolated it but my O-Swing has gone up every year after lowering CPU Contact, and it has done the same this year.

That's the thing though. I'm just reporting the numbers, so it's not really my opinion. It's still early so I'm not concluding anything, but after 20 games or so, if there's a big change, it's at a point I should notice. Swing & miss % change is pretty obvious at this point, but O-Swing %, not really (22.8% -> 20.9% when Contact is lowered... so it has actually lowered, but it's still within the margin of uncertainty, so I'm not making a big deal out of it...). Do you actually have numbers?

I'm just finding this very peculiar... I was thinking of changing Contact to affect AI discipline, and correct other unwanted changes by using Timing and Foul Frequency sliders, but the way Contact is behaving kinda funny I'm not sure if that's a sensible approach.

[nomo17k]

Re: Stats-Based CPU Sliders [MLB 13 Version]

I have a TON of respect for what you are doing both this year and in year's past. I follow it very closely and I actually use your recommended sliders in both this year's and last year's game. Your analysis and discipline in tracking stats and tendencies is amazing.

I have always thought that trying to perfect the sliders to get realistic player stats and tendencies was only half the equation. I always have thought that the real way to get what we are all after is to standardize and tweak player ratings/attributes. For every game I have owned that has always seemed to be the key. I remember doing so in NFL2K5, many versions of Madden, High Heat, etc. The games always played better and more realistically when the attributes were analyzed and a players ratings were changed as required to get the right balance of look and feel on the field as well as stats comparable to current NFL or MLB stats.

I think the reason this approach is shunned is because of the work involved in making global edits without a global editor! You have to manually edit each player in dozens of categories potentially. If there was a pc editor that existed like the one developed for 2K5 or even last year's Madden, this approach would be a lot more feasible. I do think it is the correct approach in conjunction with tweaking sliders.

Just curious to your thoughts...

C

I think nobody's shunning the approach, but it's simply too time consuming to test the effects of player attributes in gameplay. There are many attributes, and take too long to iterate over all the necessary combinations/permutations. When I do that myself, I have only done it within the "simmed" game context... I've done a little bit last year though... and though I'm not sure how sliders work in other games, for The Show sliders really are meant to neutralize player habits, and not to affect player attributes, I think. That's why sliders don't have a very wide dynamic range, and even maxing or minimizing one or more sliders don't have as strong as effect we expect... it's rather difficult to "break" the game by doing funny thing with the sliders in this game.



... and I actually trust SCEA ratings, at least more than what non-dev people generally do. I know people bitch about SCEA ratings all the time, but they actually have a very standardized way of rating those players, so internally what SCEA is doing (mostly) should be making sense. The con of the standardized approach (via canned algorithm using computer) is that it doesn't necessarily respond well to individual cases where the player has changed drastically in the last year or so.... like if a pitcher changed his approach drastically, or dropped velocity within the last year, taking weighted 3-year average doesn't work well. But it is still very consistent if we understand the compromise that it's just 3-year average.

But in general I'd like to understand what attributes do in gameplay more, particularly the ones that affect AI strategy.

[Bobhead]

Re: Stats-Based CPU Sliders [MLB 13 Version]

2% is around what I've always seen also.
In 16 CPU games that I've ran so far this year, 24.19%

26.01% In my franchise against the CPU after about 20 games

26.10% In 6 random exhibition games against the CPU, where my team was random and starter was rotated evenly

Thats all I have right now.

But 2% is a lot. This is out of pitches out of the zone... only like 45-70 pitches we're talking about here. A 2% increase is like 2 or 3 more chases per game.

If you and I both get a 2% increase consistently... I think that's a big change.

[bubs3141]

Re: Stats-Based CPU Sliders [MLB 13 Version]

are the current suggested slider the latest?

[ralphieboy11]

Re: Stats-Based CPU Sliders [MLB 13 Version]

My latest results thru 60 games. This was done in franchise mode and takes me through the #4 starters. All sliders set to 5 unless otherwise noted:

Foul frequency 6
Solid hits 4
Starter stamina 6
Strike frequency 3
Pickoffs 0
Fielding errors 6
Throwing errors 4
Fielder run speed 3
Reaction 10
Wind 0

AB 33.74
R 4.67
H 8.52
RBI 4.37
BB 3.07
K 7.03
2B 1.55
3B .15
HR .96
SAC .21
SF .25
HBP .41
DP .74
E .66
SB .6
CS .19

Analysis:
Everything looks pretty good except I am producing too many runs. My next test is moving the pitcher consistency slider to 7 and see if I can reduce hit batters and wild pitches, thus reducing scoring opportunities without reducing hits. My numbers are low but they climb throughout each 15 game stretch as I move through pitching rotations. Now that I am getting to the #5 starters I would expect the see a slight rise in offense.

This test was more so to prove that lowering the solid hit slider would raise the HR total. From what I can gather, this slider changes the trajectory of the ball coming off the bat, turning line drives into a few extra choppers but also a few extra fly balls. This increase in fly balls has raised the HR total by about .1 per team game played. I am concerned that raising the consistency slider will hurt my walk total and could possibly produce more big hits if the pitches are more centered but I am encouraged that Ralph has raised his to 6 without a significant change in offensive totals. I moved it to 7 to see if it produces an extreme total, I am guessing that it will end up at 6.

Thanks to all the same cast of characters, as you can see my sliders are are built off the information shared by others.

Happy testing! Steve

These look really good.

The note on the solid hit slider is interesting. I too would like to find a way with my set to get those home run numbers up a bit. I would be wary of reading too much into 60 games. I've been guilty of that before.

I've been playing around with pitcher consistency myself. I even had it at 10 for about 25 ballgames. This all started because of what I thought were too many HBPs and WPs. I wanted to see what that slider would do. It did bring down HBPs and WPS, but it also brings down offense as a whole. Slightly lower batting average, slightly lower BBs, slightly more SOs.

Maybe that's what you want. I don't think you'll see anything too drastic at 6 or 7. My worry is that to my eye it might create too many extremes. It's kind of like pitcher confidence on steroids(Bad baseball word).

If a pitcher can get through those first few innings unscathed and with a few strike outs, look out! He's going to be even tougher to hit throughout the ballgame. That even goes for some of your lower rated pitchers.

[Bobhead]

Re: Stats-Based CPU Sliders [MLB 13 Version]

Increasing Solid Hits will increase Home Runs imo. It doesn't just increase line drives, it just increases the authority with which balls are hit.

"Deep Fly Balls"... the class of fly balls that generally result in home runs are also greatly increased by the Solid Hits slider.

I have Solid Hits jacked up in my game and I do not get the 80% line drive rate one would expect. I get a lot of fly balls still. Same for the CPU (so human element is not obscuring my observations...). My CPU Solid Hits has been at 7/8 or so for a while, and I give up plenty of homers, lol actually as I type this I just gave one up to Pedro Alvarez.

[bubs3141]

Re: Stats-Based CPU Sliders [MLB 13 Version]

is this thread still going? Have not seen anyone post in awhile.