Stats-Based CPU Sliders [MLB 13 Version]

[ralphieboy11]

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

I do all my stuff on All-Star. I know the word has always been that it doesn't matter on cpu vs. cpu but I've always found that sort of strange. For instance, when I play myself, it seems the cpu will throw more hittable pitches on veteran than upper levels. Wouldn't that be the same even if it was cpu vs. cpu? Not sure.

[ralphieboy11]

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

Here is another 75 game test run cpu vs cpu(Season mode). For this run I only bumped up fielder reaction to 7. I kept everything else the same from my previous tests.

Strike Frequency 3
Foul Frequency 6
Fielder Reaction 7
Fielder Speed 4
Arm Strength 2
Starter Stamina 6
Baserunner Speed 6
SB Frequency 4

Average .249
Walks 2.99
SO 7.47
2B 1.72
3B .17
HRs .88
Runs 4.30
SB .59
CS .14
HBP .37
WP .43
Errors .61

These numbers look pretty good. Batting average dropped 15 points with this set. It's tough to say how much fielder reaction had to do with it, considering it's still a small sample size. Plus I saw Stevie with it all the way at 10 and it wasn't doing anything crazy. You probably reach a point with that slider where it doesn't do much. Sometimes a hit is just a hit. I think I'm going to try fielder reaction at 6 for my next run.

Regarding everything else I'm pretty pleased. Home Runs could use a boost but my previous average was 0.96 so not too concerned there. HBPs and WPs are up a bit. This is the first I've really recorded those, so not sure if I should worry about it. Pitcher consistency seems to control that according to description but that's a slider I'm a little leery of.

From here I think I'm going to go with a few more minor tweaks that shouldn't affect the numbers too much.

Going to go with what Nomo has for these..
Manager Hook 4
Fielding Errors 6
Throwing Errors 4

Also going to bump reliever stamina up to 7. I had the issue with the tired bullpen that I previously reported in this thread. Maybe this will help slightly with that.

[steviegolfballs]

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

I'm nervous with dropping the hook, as you saw in my post it sure seems like it is making for some big innings, although the sample is small. I also noticed that there were bigger innings, in the five games played there were 4 innings that had 6+ runs, in my 50 games from before with hook at 5 I only had two innings that had 6+. I'm wondering if leaving the hook at 5 and adjusting the stamina might produce fewer pitching changes and still keep offense realistic. I am going to do more tests at hook 3 this afternoon but if I continue to see massive middle innings then I know something is up.

Here is another 75 game test run cpu vs cpu(Season mode). For this run I only bumped up fielder reaction to 7. I kept everything else the same from my previous tests.

Strike Frequency 3
Foul Frequency 6
Fielder Reaction 7
Fielder Speed 4
Arm Strength 2
Starter Stamina 6
Baserunner Speed 6
SB Frequency 4

Average .249
Walks 2.99
SO 7.47
2B 1.72
3B .17
HRs .88
Runs 4.30
SB .59
CS .14
HBP .37
WP .43
Errors .61

These numbers look pretty good. Batting average dropped 15 points with this set. It's tough to say how much fielder reaction had to do with it, considering it's still a small sample size. Plus I saw Stevie with it all the way at 10 and it wasn't doing anything crazy. You probably reach a point with that slider where it doesn't do much. Sometimes a hit is just a hit. I think I'm going to try fielder reaction at 6 for my next run.

Regarding everything else I'm pretty pleased. Home Runs could use a boost but my previous average was 0.96 so not too concerned there. HBPs and WPs are up a bit. This is the first I've really recorded those, so not sure if I should worry about it. Pitcher consistency seems to control that according to description but that's a slider I'm a little leery of.

From here I think I'm going to go with a few more minor tweaks that shouldn't affect the numbers too much.

Going to go with what Nomo has for these..
Manager Hook 4
Fielding Errors 6
Throwing Errors 4

Also going to bump reliever stamina up to 7. I had the issue with the tired bullpen that I previously reported in this thread. Maybe this will help slightly with that.

[ralphieboy11]

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

I'm nervous with dropping the hook, as you saw in my post it sure seems like it is making for some big innings, although the sample is small. I also noticed that there were bigger innings, in the five games played there were 4 innings that had 6+ runs, in my 50 games from before with hook at 5 I only had two innings that had 6+. I'm wondering if leaving the hook at 5 and adjusting the stamina might produce fewer pitching changes and still keep offense realistic. I am going to do more tests at hook 3 this afternoon but if I continue to see massive middle innings then I know something is up.

I don't think the manager hook slider would be the reason for those spikes in certain innings. I don't think the slider is volatile enough for that. Nomo has pointed out that it does have some effect but it is my belief that the majority of manager bullpen decisions are hard coded into the AI. I wish the slider produced more extremes. For instance, I'd like to be able to set it at 0 and almost never see the CPU change pitchers. But I have put it at 0 in MoM games and others in the past, and the AI will still pull starters sort of early in a close ballgame. It also pulls the starter that is behind even if he's not tired. Sometimes that bothers me. If you have Verlander or Darvish pitching you are not going to pull them after 80 something pitches just because you are losing. Managers are going to extend them as long as possible because the bullpen option is usually a lesser arm. That might be a little of what you are seeing here. The pitchers that are brought in for middle relief are usually not the stars of the team. Those are usually starters or closers. If you couple that with the confidence system that is built into the game, middle relievers can really be susceptible to a poor outing from time to time.

Don't take my words as gospel. I'm sort of piggybacking on a lot of the stuff I've read from Nomo and others over the past couple years. I know he's done a lot of testing regarding runs scored per inning in the past.

[nomo17k]

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

Thanks, but just an input, if it's anything like last year then there is a difference. I've played many games so far on 13 but haven't changed it from All-Star yet. I know they said that last year too but I found a big difference between AS and Legend. Games were more spot on Legend than in AS. A lot more inconsistent in AS mode. Obviously this is my humble opinion.

Has anyone else tried these at different difficulties?

I'm setting it to All-stat, but the word from SCEA is that the difficulty levels don't matter in CPU vs. CPU games.

Last year, there were a couple guys adamantly insisting on how they really do matter despite this, so I was actually running a large portion of my testing on different levels for a while... my record shows that I was at one point using Rookie for both pitching and hitting for a while (a few dozen games), and then switched to Legend/Rookie, just to see if I see any difference.

I did not see any differences... at least in the stats that I was tracking.

I do all my stuff on All-Star. I know the word has always been that it doesn't matter on cpu vs. cpu but I've always found that sort of strange. For instance, when I play myself, it seems the cpu will throw more hittable pitches on veteran than upper levels. Wouldn't that be the same even if it was cpu vs. cpu? Not sure.

I don't find it strange... when you think about it, if you are using Rookie difficulty for hitting and Legend for pitching, how does CPU decides which difficulty to play the game on? We cannot set those difficulties separately for each team. CPU on the pitching side is suppose to be easy on the CPU hitting (because of Rookie hitting) by perhaps throwing more strikes, but on the flip side if CPU were to pitch on Legend AI, he should not be constantly throwing strikes. It doesn't make sense to me...


However, I actually observed something interesting last year... I don't know if I posted the result, but I did a 75 test set at the AA levels, playing minor league games. I was using the same slider set that I tuned for the MLB games, so walks should be about 3.3 per game per team... but down in AA, the walks were actually down (about 2.5), and the overall strike % was about a couple % higher as well (63% vs. 65%).

I wondered why then, since AA pitchers probably are wilder with low BB/9 and all (though AA batters may be less disciplined as well). Then I heard that minor league games are played at lower difficulty levels (at least in RTTS... not so sure if it's played in other modes).... and at lower difficulty levels, the game is tuned to yield about 2% more strikes overall.... coincidence?? I am not sure.... but in certain situations, difficulty levels could be propagated to CPU vs. CPU games as well.... but I don't see it does at the MLB level.

[nomo17k]

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

I'm nervous with dropping the hook, as you saw in my post it sure seems like it is making for some big innings, although the sample is small. I also noticed that there were bigger innings, in the five games played there were 4 innings that had 6+ runs, in my 50 games from before with hook at 5 I only had two innings that had 6+. I'm wondering if leaving the hook at 5 and adjusting the stamina might produce fewer pitching changes and still keep offense realistic. I am going to do more tests at hook 3 this afternoon but if I continue to see massive middle innings then I know something is up.

I don't think the manager hook slider would be the reason for those spikes in certain innings. I don't think the slider is volatile enough for that. Nomo has pointed out that it does have some effect but it is my belief that the majority of manager bullpen decisions are hard coded into the AI. I wish the slider produced more extremes. For instance, I'd like to be able to set it at 0 and almost never see the CPU change pitchers. But I have put it at 0 in MoM games and others in the past, and the AI will still pull starters sort of early in a close ballgame. It also pulls the starter that is behind even if he's not tired. Sometimes that bothers me. If you have Verlander or Darvish pitching you are not going to pull them after 80 something pitches just because you are losing. Managers are going to extend them as long as possible because the bullpen option is usually a lesser arm. That might be a little of what you are seeing here. The pitchers that are brought in for middle relief are usually not the stars of the team. Those are usually starters or closers. If you couple that with the confidence system that is built into the game, middle relievers can really be susceptible to a poor outing from time to time.

Don't take my words as gospel. I'm sort of piggybacking on a lot of the stuff I've read from Nomo and others over the past couple years. I know he's done a lot of testing regarding runs scored per inning in the past.

Yeah 5 games isn't enough to draw any conclusions... by far the most testing that I've done with Manager Hook was with MLB 11, but I don't think it had such a huge effect on reliever effectiveness, most of the time I was lowering it, even to 0...

The thing does adjust reliever usage, as well as how long starters stay in the game. If relievers seem ineffective, it's just a bad luck or maybe they are in a stretch where many of the guys are coming in with less than average energy, because of recent extra inning games and such.

[steviegolfballs]

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

Nomo,

Do you have any data on what effect the power slider actually has on the number of homeruns hit? I seem to have everything pretty well nailed down but I'm not producing quite enough homeruns compared to a league average.

Thanks, Steve

[nomo17k]

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

Nomo,

Do you have any data on what effect the power slider actually has on the number of homeruns hit? I seem to have everything pretty well nailed down but I'm not producing quite enough homeruns compared to a league average.

Thanks, Steve

I think I lowered Power to 4 in MLB 11... and that's because HRs were inflated back then. Wind was also very strong. 1.64 HRs per game per team. With Power 4 and Wind 3, it went down to about 1.00... which is quite a change.

If the HR number stays low, I might also consider increasing Power to 6... it might now be a bad idea given XBHs aren't coming as much.. the slider pretty much only seems to affect how fast the batted ball travels, so it probably has few unwanted side effects.

[steviegolfballs]

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

I've run 15 games with the same sliders I listed before except with power at 6 and it doubled my home runs and increased my xbh, neither of which are acceptable. I currently test with no wind, do you find that playing with wind in the game affects the home run output one way or the other or do the wind out games balance the wind in games? If I could get another .1 HR's per game without blowing up my other numbers I'd be one happy guy!

I think I lowered Power to 4 in MLB 11... and that's because HRs were inflated back then. Wind was also very strong. 1.64 HRs per game per team. With Power 4 and Wind 3, it went down to about 1.00... which is quite a change.

If the HR number stays low, I might also consider increasing Power to 6... it might now be a bad idea given XBHs aren't coming as much.. the slider pretty much only seems to affect how fast the batted ball travels, so it probably has few unwanted side effects.

[nomo17k]

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

I've run 15 games with the same sliders I listed before except with power at 6 and it doubled my home runs and increased my xbh, neither of which are acceptable. I currently test with no wind, do you find that playing with wind in the game affects the home run output one way or the other or do the wind out games balance the wind in games? If I could get another .1 HR's per game without blowing up my other numbers I'd be one happy guy!

For XBHs, rather than per game per team, can you look at 2B/H, 3B/H, and HR/H if you have been tracking those?

I think this will give us a better idea as to the ability of the game to produce XBHs. I'm just totally guessing, but I bet in your 15 game set the offense was quite a bit inflated overall, which likely means you saw more hits, which leads to more XBHs per game, etc., etc. By looking at /H numbers, we can see what fraction of hits are XBHs

As for wind effect, wing blowing in and out seems to have a fairly strong effect on fly balls. Those number was from MLB 12, and there's no guarantee the wind works the same way this year... I have been tracking wind direction so at some point I'll revisit this. But I remember you totally turn of Wind for your games, correct?

I think the performance variation in this year's game really has been widened, partly due to player attribute adjustments (i.e., player rating differences lead to wider performance differences than before)... and that may be a reason why we are seeing quite a few short-term performance fluctuations. 15 games isn't such a large sample size.

[ralphieboy11]

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

Home runs have been down for me as well. Averaging .91 over all the games I've run. I've been playing with wind at default. I have not tried to up power but I would also feel like that would throw things off too much.

I will probably just live with it. Beyond slight player ratings tweaks, I'm not sure what else can be done. I guess we were thinking less control means more 'meatballs', but it doesn't seem to be translating into more home runs overall.

I guess you could slightly up the wind effect when the wind blows out and slightly reduce it when the wind blows in. While I have no statistical evidence it does seem the wind effect had been lowered this year. For instance, you don't see the big curves in ball flight on pop ups in foul ground like we have in the past.

[nomo17k]

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

Home runs have been down for me as well. Averaging .91 over all the games I've run. I've been playing with wind at default. I have not tried to up power but I would also feel like that would throw things off too much.

I will probably just live with it. Beyond slight player ratings tweaks, I'm not sure what else can be done. I guess we were thinking less control means more 'meatballs', but it doesn't seem to be translating into more home runs overall.

I guess you could slightly up the wind effect when the wind blows out and slightly reduce it when the wind blows in. While I have no statistical evidence it does seem the wind effect had been lowered this year. For instance, you don't see the big curves in ball flight on pop ups in foul ground like we have in the past.

Yeah, HRs are slightly down, but not to the extent that bothers me that much...

I think the "big curves" you are referring was mostly caused by too much backspin that the ball physics caused to happen in MLB 12. We see less "big curves" as well as less ball hang time in air this year, likely due to the adjustment made in this area... either that or SCEA stopped receiving consulting from Lane Bryant.

[steviegolfballs]

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

Here's what I am currently working on, It is in franchise mode and this is through 15 games so every team has played once and used their number one starter. I will update This again when I have my next 15 games done and I am through the number two starters but so far so good!

Sliders at default (5) unless specified:
Power 6
Foul frequency 6
Starter stamina 6
Strike frequency 4
Pickoff 0
Fielding errors 6
Throwing errors 4
Fielder run speed 3
Fielder reaction 10
Fielder arm strength 0
Wind 0

AB 34.27
R 4.23
H 8.53
RBI 4.13
BB 3.33
K 7.40
2B 1.60
3B .20
HR .97
SH .30
SF .33
HP .20
DP .47
E .53
SB .73
CS .17

Eye test....very pleased, one shutout (Chris Sale vs. KC) but also some nice offensive games (Pit 8 Cubs 6) most games with 1-2 HRs, Toronto hit 4 off Cleveland. I haven't been keeping specific pitch counts but I do check from time to time and pitchers seem to be right around 60% strikes and counts seem to be right around 3.8-4 pitches per batter. One game (Seattle at Oakland) went 13 innings, before that my ABs were closer to about 33.8

[steviegolfballs]

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

Forget the last post, offense exploded as I moved through the twos and threes. Back to the drawing board.

[Bobhead]

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

Forget the last post, offense exploded as I moved through the twos and threes. Back to the drawing board.

I recommend you do immediate rotations, instead of 15 games at a time. That way your stats are always informative, and you don't have to wait for them to "balance out"

eg: game 1, #1 starter, game 2, #2 starter... all the way to 15 games, instead of 15 games with the same spot #.