CPU plate discipline logic has been re-written

[tessl]

Re: CPU plate discipline logic has been re-written

If you are really curious, I document all the changes I made to sliders and also all the data I gather in order to make decisions, and when people ask why some changes have been made, and I elaborate on why I made that decision personally:

http://www.operationsports.com/forum...4-version.html

It has the chase % (O-Swing %) from MLB and in the game. If anything, CPU hitters still don't chase nearly as much as MLB hitters do (they swing at about 27% vs. 30% of pitches outside the strike zone).

All the data I have gathered indicate that your suspicion about the perceived negative effect by the plate discipline change is unfounded. The fact that the game was a bit unbalanced at default is not attributable to the CPU discipline improvement. My guess is that it's more likely due to how the game now produce a lot more poor hits that go in between fielders, whereas these hits were very hard to generated in previous iterations of The Show.

I'm familiar with o-swing% from fangraphs. Maybe with more games it will work itself out but it seems like no batter is immune from it. Guys like Matt Carpenter with good contact, vision and discipline chase pitches and instead of missing they make contact which results in an out.

At the moment I have a small sample size but it looks like a change away from realism compared to last year. I'll have a better idea when I have played a full season.

[nomo17k]

Re: CPU plate discipline logic has been re-written

I'm familiar with o-swing% from fangraphs. Maybe with more games it will work itself out but it seems like no batter is immune from it. Guys like Matt Carpenter with good contact, vision and discipline chase pitches and instead of missing they make contact which results in an out.

At the moment I have a small sample size but it looks like a change away from realism compared to last year. I'll have a better idea when I have played a full season.

Personally, I think you'll be wasting a lot of time doing your own testing, if you already have a conclusion that you want to reach (i.e., the fact that a random guy like myself pushed for an "improvement" destroyed the game...). I've already run a couple seasons worth of essentially similar testing that you are trying to do, and I disagree with what you are seeing. If you want to do meaningful analysis, you should start from being neutral and just let the data speak for themselves before you even start forming your own hypothesis.

What you described with a guy like Matt Carpenter is likely an entirely different issue. I still like to see a wider separation in players' disciplines. That I agree, but that's just an additional improvement we'd like to see on top of the improvements we are fortunately now seeing (i.e., that the CPU needed to be much less disciplined).

[tessl]

Re: CPU plate discipline logic has been re-written

Personally, I think you'll be wasting a lot of time doing your own testing, if you already have a conclusion that you want to reach (i.e., the fact that a random guy like myself pushed for an "improvement" destroyed the game...). I've already run a couple seasons worth of essentially similar testing that you are trying to do, and I disagree with what you are seeing. If you want to do meaningful analysis, you should start from being neutral and just let the data speak for themselves before you even start forming your own hypothesis.

What you described with a guy like Matt Carpenter is likely an entirely different issue. I still like to see a wider separation in players' disciplines. That I agree, but that's just an additional improvement we'd like to see on top of the improvements we are fortunately now seeing (i.e., that the CPU needed to be much less disciplined).

Thanks for the reply. I do prefer to do my own testing and I'm tracking chased pitches per game, outs per game on chased pitches, o-swing% as well as runs per game and hits per game. Zero simming, all games played in MoM with default sliders and umpires off.

I've looked at you data previously and I have a question. Are you playing or simming and are you using default sliders with umpires off?

[nomo17k]

Re: CPU plate discipline logic has been re-written

Thanks for the reply. I do prefer to do my own testing and I'm tracking chased pitches per game, outs per game on chased pitches, o-swing% as well as runs per game and hits per game. Zero simming, all games played in MoM with default sliders and umpires off.

I've looked at you data previously and I have a question. Are you playing or simming and are you using default sliders with umpires off?

The fact that you ask a question like this shows that you have not looked at the thread or the data... and hence my worry that you will just be wasting time accumulating data that I have already collected and even been making public for everyone to access and use for a while by now.

CPU vs. CPU games are by definition all played out. No simming.

I put variable umpires on because I sometimes want to analyze the differences among the umpires (as in strike %). But they are recorded for each game, so it is not too difficult to neutralize the effect of variable umpires if necessary.

I always do an initial set of 75 games (= 150 games for two teams per game) with default sliders. Information without the baseline like that is useless.

[tessl]

Re: CPU plate discipline logic has been re-written

The fact that you ask a question like this shows that you have not looked at the thread or the data... and hence my worry that you will just be wasting time accumulating data that I have already collected and even been making public for everyone to access and use for a while by now.

CPU vs. CPU games are by definition all played out. No simming.

I put variable umpires on because I sometimes want to analyze the differences among the umpires (as in strike %). But they are recorded for each game, so it is not too difficult to neutralize the effect of variable umpires if necessary.

I always do an initial set of 75 games (= 150 games for two teams per game) with default sliders. Information without the baseline like that is useless.

Interesting reply. I asked looked at your data and asked about it here in reply to your post in March...
I wouldn't say I only care what Brian says but he is the lead AI programmer for the game and the only person in this thread who has both played and has the ability to adjust mlbts 14 so his comments do carry significant weight. His posts in the past regarding which attributes most impact the game are the most informative posts I've seen on this forum.

I commend you for your research. I looked at your chart and I have a question. Are those only for games with default sliders?


You answered my question, you are using variable umpires. That may give some information regarding umpire tendencies but I don't see how it can be accurately used to determine o-swing%. It seems to me in order to attain accurate information umpires should be off. No insult intended, you are obviously passionate about your research but I don't believe it is accurate.

When I get to 100 games played I will post results with umpires off in manage mode on default sliders. At the moment I'm seeing low run and hit totals compared to real life MLB. I'm also tracking outs per game on chased pitches. My early results indicate they didn't adequately account for the additional outs from more chased pitches and have stunted the offense.

[nomo17k]

Re: CPU plate discipline logic has been re-written

...


You answered my question, you are using variable umpires. That may give some information regarding umpire tendencies but I don't see how it can be accurately used to determine o-swing%. It seems to me in order to attain accurate information umpires should be off. No insult intended, you are obviously passionate about your research but I don't believe it is accurate.

When I get to 100 games played I will post results with umpires off in manage mode on default sliders. At the moment I'm seeing low run and hit totals compared to real life MLB. I'm also tracking outs per game on chased pitches. My early results indicate they didn't adequately account for the additional outs from more chased pitches and have stunted the offense.

I think you don't even seem to realize how little an effect variable umpires have on the overall game balance. I've looked at this a little on MLB 12, and over a longer run it really doesn't affect the game results in any discernible way (to me, unfortunately... I want to see a bigger effect).

And you don't seem to realize that you cannot actually compute O-Swing % from the stats you said that you will be collecting right now. Please go ahead with your own experiment and draw any conclusions you like. It will not affect my opinion (nor that of the devs I can say) at all.

I'm sorry but I can only go SMH for how some people here in this forum can be so arrogant while being totally ignorant of what they are doing.

Please just do not waste my time by mentioning me. Good luck.

[tessl]

Re: CPU plate discipline logic has been re-written

I think you don't even seem to realize how little an effect variable umpires have on the overall game balance. I've looked at this a little on MLB 12, and over a longer run it really doesn't affect the game results in any discernible way (to me, unfortunately... I want to see a bigger effect).

And you don't seem to realize that you cannot actually compute O-Swing % from the stats you said that you will be collecting right now. Please go ahead with your own experiment and draw any conclusions you like. It will not affect my opinion (nor that of the devs I can say) at all.

I'm sorry but I can only go SMH for how some people here in this forum can be so arrogant while being totally ignorant of what they are doing.

Please just do not waste my time by mentioning me. Good luck.

Physician heal thyself :-). When I started programming we were using keypunch machines. I remember when the IBM 129 came out in 1971 which allowed the operator to both punch and verify on the same machine - state of the art technology- and even then there was a saying "garbage in, garbage out".

This is the biggest change in gameplay in years and it was done using data gathered for 0-swing% using a completely random strike zone. Not only is it not statistically accurate but it doesn't pass the eye test. I just played a game Cubs vs Pirates in which 25 outs were made on chased pitches. McCutchen, reigning NL MVP struck out on consecutive at bats when he took two called strikes and then swung at a pitch which was so far inside it nearly hit him - on consecutive at bats. Reigning NL MVP has no idea whether a pitch is a strike or a ball. If he randomly swings at a pitch which is randomly a strike he has a better than average chance of getting a good result.

It appears to me the determination of whether or not a batter swings at a pitch is random and has nothing to do with pitch location. Pitchers with low bb/9 ratings instead of giving up a lot of walks are getting a lot of outs on chased pitches.

I don't just let the games run and then accumulate data when the game is over, I watch the games. mlbts14 doesn't pass the eye test.

[nomo17k]

Re: CPU plate discipline logic has been re-written

On that we agree.

I wish you agree with other statements with more substance (to save the time of both yours and mine).

The bottom line is, your accusation is very misguided for many reasons:

(1) You are (unintentionally) undermining the ability of the devs to make their own decisions and fine-tune the game by implying that they succumb to baseless feedback by dumb and stupid random OS forum members like myself. Do you honestly believe that an extremely intelligent developer like Brian@SCEA changes up things just to please dumb and stupid random OS forum members like NOMO17K (and others who do not agree with *your* opinions)? It is not going to happen. The (offline) game has been remarkably stable and remain great while going through steady, incremental improvements because the developers filter in only good feedback but filter out bad ones. And you are saying that a long-standing complaint about eagle-eyed CPU, which had been shared by many who actually play the game and solidly supported by data a couple people had collected, basically destroyed the game balance of your MOM. First of all, that particular feedback didn't break the game. Second, your feedback is not the only feedback that counts. Finally, it is the developers who choose which feedback is of more value. If my feedback (and that of others) is indeed bad and the devs succumb to it, then why don't you blame the devs instead?

(2) You have already been banned once for being annoyingly persistent, ineffectively redundant, and overall nonconstructive about bringing forward your single and only wish about the game (that MOM be fixed/improved). Then Brian came along and fixed most of shortcomings (out of blue), and now the only thing that you ever value in this forum is what he writes.... not knowing that it is actually Brian himself who already knew that the CPU discipline in the game needed to be improved. He did, in fact, successfully improve the aspect the last couple years, but this year's game shipped with a part of game (BABIP) relatively unbalanced for some reason. If you know anything about BABIP (much easier to collect data on than plate discipline), it's pretty clear THAT is the main issue, not plate discipline. A classic case of confusing correlation as causation. Just because of your own bias against improved plate discipline. I mean, do you have any other meaningful points to make than just singling out CPU plate discipline? As you say, one change definitely has ripple effects on others, but why don't you even think of a possibility that something else can be the cause of low offense production in this year's game at default setting? Because you have a narrow-minded agenda, and that is why.

(3) If you really want to think the variable umpires causes noticeable differences in plate discipline or any other aspects of the game to the extent that it invalidates all my opinions, there is a very efficient thing you can do to test your hypothesis by using the data that I have already accumulated. Just aggregate and group the data by the umpires, test for statistically significant differences among them. Given the data is there for use already, that should take only less than 0.5 hour if you already know basic stats (if you don't, then it will take longer but then how should anyone trust your analysis when you don't know how to analyze?). That will save you from additionally running 100 MOM games on your own, about 60 - 70 hours worth of your life and more at least, when you can just start playing using what others have already learned as the baseline (on which you can make your own adjustments).


My reason for making the data that I collect public is exactly for people to do this kind of secondary analyses themselves using the data if they wish, without them spending as many meaningless hours as I have running all these stupidly boring CPU vs. CPU games themselves, taking numbers out and recording them onto at-times stupidly unresponsive Google Spreadsheet (not to mention fatter electricity bills from running PS3 and PS4 concurrently all the time).

Guess what, nobody does. Only thing that people like you ever do is to make baseless accusations. It is exactly this kind of meaningless accusation that makes me no longer want to make the data public in future, let alone discuss them. It's meaningless to do so in this forum because of a dude like you. :P

I know I should just not take you seriously, but I'm just being very honest of my displeasure on unreasonable accusations from people who are indeed very ignorant.

[nomo17k]

Re: CPU plate discipline logic has been re-written

Physician heal thyself :-). When I started programming we were using keypunch machines. I remember when the IBM 129 came out in 1971 which allowed the operator to both punch and verify on the same machine - state of the art technology- and even then there was a saying "garbage in, garbage out".

...

Mentioning things like this (combined with your overall behavior in the forum) only says that you are a very old person who have failed to update his knowledge. Punch card was well before my time, but I have coded using machine language so if you think that you having used punch card impresses me, it unfortunately does not. :P

Just because you have programmed using a punch card doesn't automatically make you more knowledgeable than late comers. It only says that you are old and have very rigid thinking, and nothing about your knowledge or ability.

Eye test, yes, of course. I watch plenty of games myself. Certain aspects still don't pass my own eye tests. I feel CPU guesses and takes pitches too often when it should protects. I have plenty of feedback like that. I just do that on top of accumulating data in hope of adding feedback that are actually substantiated.

[Bobhead]

Re: CPU plate discipline logic has been re-written

I would certainly like to see evidence of a game with 25 chased-pitch outs.

[Bobhead]

Re: CPU plate discipline logic has been re-written

... I just played a game Cubs vs Pirates in which 25 outs were made on chased pitches.

Wait it just occurred to me that you might be referring to 25 chased pitch outs total amongst both teams in the game? In that case, that is not even out of the ordinary. The average is about 26 misses a game (including misses on strikes), and 42 chases a game (including chases where contact is made).

So 25 chased pitch outs is completely within the scope of an average MLB game.

[dfhsjhu]

Re: CPU plate discipline logic has been re-written

I wonder exactly how that impacts the game and in what way it was changed.