CPU plate discipline logic has been re-written

[Armor and Sword]

Re: CPU plate discipline logic has been re-written

I agree with Brian wholeheartedly about the experience is different for every player.

On 13 I manage to get some great chase and looking strikeouts far more than ever before.

So I can only imagine the improvement for 14!!! Because on 13 it was night and day from 11 and 12. Clearly better on 13 for me.

Everything for me is how you set up batters….understand their hot and cold zones and use that information to your advantage. Also the hitters plate vision plays a huge role from my experience.

April 1st can't get here soon enough. This is one of the enhancements I can't wait to test drive.

[Qb]

Re: CPU plate discipline logic has been re-written

Looking forward to the potential improvements. I found more success striking batters out with fastballs in the zone than offspeed stuff moving out of the zone. As a result, I got pretty good at setting hitters up for a called third strike fastball at the knees away, but rarely seemed to be able to get them to chase the breaking ball out there.

[Maverick32093209]

Re: CPU plate discipline logic has been re-written

For me it really depended on the pitch and the set up. I could get guys to chase pitches out of the zone if I set them up correctly. a slider away out of the zone for a swinging strike after hit the outer edge with a comeback 2 seamer.

some pitches didn't make sense...i could only get swing and misses on splitters in the zone...no matter how I set it up. But I also feels splitters don't drop enough in the game.

curveball was a mixed bag...i could go out of the zone...but just out to the zone...no swing and misses on pitches in the dirt.

also...as a Yankee fan using Mariano Rivera...i could get swing and misses on cutters in on lefties and away from rightes quite regularly

[MrOldboy]

Re: CPU plate discipline logic has been re-written

For me it really depended on the pitch and the set up. I could get guys to chase pitches out of the zone if I set them up correctly. a slider away out of the zone for a swinging strike after hit the outer edge with a comeback 2 seamer.

some pitches didn't make sense...i could only get swing and misses on splitters in the zone...no matter how I set it up. But I also feels splitters don't drop enough in the game.

curveball was a mixed bag...i could go out of the zone...but just out to the zone...no swing and misses on pitches in the dirt.

also...as a Yankee fan using Mariano Rivera...i could get swing and misses on cutters in on lefties and away from rightes quite regularly

Certain pitches in The Show have worked to get players to chase. High fastballs work. Sliders/cutters inside to LH and away from RH work. Changeups in the zone or close too it can get swings and misses.

But the big hole in the AI for me has always been the low pitches that are designed to induce a chase. Splitters in the dirt, sliders in the dirt, curveballs in the dirt. These just do not result in nearly enough instances where the AI will swing at them. There has been long discussions on this in the past including data collection on the frequency of the AI batting results. I'm too lazy to do much data collection, but I beleive Nomo17k and others note a lot of stats for games they runs, including CPU vs CPU.

But from a purely observational perspective to me the AI is too easy to exploit on some pitches (high fastballs being one of them) and not nearly exploitable enough (balls in the dirt) to represent a realistic representation of MLB at bats.

[Armor and Sword]

Re: CPU plate discipline logic has been re-written

Yes….agree on the splitters and the 12-6 curves. I can not get a chase on those from the CPU.

Typically if I have a pitcher on the mound with a very good curve…..I throw that as my first pitch a lot…..and if he is going good I can get ahead 0-1 as the CPU tends to look at first pitch curveball strikes.

Other than that it is really tough to get them to chase out of the zone breaking balls and splitters.

SO hopefully that is one of the real improvement areas. But fastballs, cutters, sliders….I can get them looking or chasing.

[Hockeynut99]

Re: CPU plate discipline logic has been re-written

I would think that would be like writing a thesis for Brian. To try to explain it. Lol

[nomo17k]

Re: CPU plate discipline logic has been re-written

Certain pitches in The Show have worked to get players to chase. High fastballs work. Sliders/cutters inside to LH and away from RH work. Changeups in the zone or close too it can get swings and misses.

But the big hole in the AI for me has always been the low pitches that are designed to induce a chase. Splitters in the dirt, sliders in the dirt, curveballs in the dirt. These just do not result in nearly enough instances where the AI will swing at them. There has been long discussions on this in the past including data collection on the frequency of the AI batting results. I'm too lazy to do much data collection, but I beleive Nomo17k and others note a lot of stats for games they runs, including CPU vs CPU.

But from a purely observational perspective to me the AI is too easy to exploit on some pitches (high fastballs being one of them) and not nearly exploitable enough (balls in the dirt) to represent a realistic representation of MLB at bats.

I haven't really gone as far as tracking pitch-by-pitch result by count (only did so for strike %), so I cannot really have hard evidence but my own play experience and looking at CPU vs. CPU games do agree with what many guys are saying in this thread.

I do give enormous credits to Brian for improving this aspect of the gameplay though. I think CPU discipline has been steadily improving (since MLB 10 when I started playing). And one of the major pain point (I think) of the issue, chasing pitches out of the zone, has started happening more in MLB 13 compared to earlier iterations of the game. I can have CPU chase breaking pitches more now than I used to. I always test the game using the same pitchers, and the difference has been noticeable without tracking numbers.

However, it still doesn't happen as often compared to real-life numbers, meaning, I haven't been able to bring the chase % up to MLB numbers by tweaking sliders for example.

I would agree that fastball remains the most effective strikeout pitch for both in the strike zone and outside the strike zone. I would also agree that CPU probably takes the third strike more often than it probably should. Overall, pitching against CPU still revolves around how to effectively use fastball in sequence, since most breaking/off-speed pitches aren't as effective as they probably should be.

But we should keep in mind that it is also likely that *we* do not pitch like most MLB pitchers do, which influences how CPU reacts to our pitching strategy in the game. I play against human players a lot, and the effectiveness of pitching, which in the game is mostly simply pressing/moving buttons/sticks at the right timing, wildly varies among gamers. Most gamers throw too many strikes compared to real-life pitchers.

Balancing that human factor with maintaining realistic gameplay result/experience must be tough... when we expect experiences close to real-life, we also need to play our games realistically. But do we actually? Not very easy to see.

[cardinalbird5]

Re: CPU plate discipline logic has been re-written

If 2 quality humans play one another with some discipline, the pitch counts are pretty realistic. Once again, this is rarely the case online as most guys get predictable with their pitch sequence and swing at everything when hitting.

[ryan36]

Re: CPU plate discipline logic has been re-written

I will echo that in general more swings and misses are still needed in my opinion. And the only K pitches are a fastball, slider, and a good change. Throw anything else and not so many missed bats. Plus there are a lot of flyballs on sinkers ...stuff like that

Sent from my SGH-M919 using Tapatalk

[Armor and Sword]

Re: CPU plate discipline logic has been re-written

So last night I was playing game one of the ALCS between the 1999 Yankees vs the 99 Astros on The Show 13.

Clemens on the hill.


I was dominant for 7 innings and K'd 8 batters. 6 of those batters I was able to induce swinging strike outs with Clemens split finger fastball.


It was beautiful.


Classic Pitching.


Of course....I ended up losing game one 5-3 as we self destructed in the top of the 8th as Jeff Bagwell made a deposit with a two run shot to take the lead and Wagner came in and fanned the side.


Sigh.

[tessl]

Re: CPU plate discipline logic has been re-written

I will echo that in general more swings and misses are still needed in my opinion. And the only K pitches are a fastball, slider, and a good change. Throw anything else and not so many missed bats. Plus there are a lot of flyballs on sinkers ...stuff like that

Sent from my SGH-M919 using Tapatalk

I play manage mode and strikeout totals were very realistic in mlbts 13. If people who "joystick" the game were getting different results I suspect user input was a major factor. I wouldn't want to see manage mode turn into a strikeout fest just to appease joystickers who can't get the cpu to chase more pitches but Brian is as good as it gets in getting this game to play realistically so I believe any changes will be minor.

[nomo17k]

Re: CPU plate discipline logic has been re-written

I play manage mode and strikeout totals were very realistic in mlbts 13. If people who "joystick" the game were getting different results I suspect user input was a major factor. I wouldn't want to see manage mode turn into a strikeout fest just to appease joystickers who can't get the cpu to chase more pitches but Brian is as good as it gets in getting this game to play realistically so I believe any changes will be minor.

I do agree that the occurrence of strikeout overall is realistic, but I think what comes out of people's overall impression in a thread like this in the past is that strikeouts in the game happen more within the strike zone than outside than you would expect in real life. CPU often takes or just swings through a fastball for the third strike. I agree with that impression as well.

I think what should happen is that CPU should chase more when it's taken out the zone, so that he does not have to go down on pitches within the strike zone to keep the overall strikeout rate balanced.

[tessl]

Re: CPU plate discipline logic has been re-written

I do agree that the occurrence of strikeout overall is realistic, but I think what comes out of people's overall impression in a thread like this in the past is that strikeouts in the game happen more within the strike zone than outside than you would expect in real life. CPU often takes or just swings through a fastball for the third strike. I agree with that impression as well.

I think what should happen is that CPU should chase more when it's taken out the zone, so that he does not have to go down on pitches within the strike zone to keep the overall strikeout rate balanced.

The dilemma for the developers is to make one group - joystickers - happy without messing up the game for another group - manage mode users. One frustrating thing in manage mode was hitters taking called third strikes with risp and 2 out. While it was frustrating I can't honestly say it wasn't realistic. I won't know until May when I put a disc into my ps4 whether that has been made better, worse or no change.

And, as I said, total strikeouts for manage mode which is the best way to test the game because it removes human variables, were realistic - the most realistic in the history of the franchise. My concern is in order to appease users joysticking the game who want to strike out 15 cpu batters per game the game it could cause manage mode to become strikeout fest. I adjusted zero sliders for mlbts 13, that's how realistic the game was in mange mode.

[nomo17k]

Re: CPU plate discipline logic has been re-written

The dilemma for the developers is to make one group - joystickers - happy without messing up the game for another group - manage mode users. One frustrating thing in manage mode was hitters taking called third strikes with risp and 2 out. While it was frustrating I can't honestly say it wasn't realistic. I won't know until May when I put a disc into my ps4 whether that has been made better, worse or no change.

And, as I said, total strikeouts for manage mode which is the best way to test the game because it removes human variables, were realistic - the most realistic in the history of the franchise. My concern is in order to appease users joysticking the game who want to strike out 15 cpu batters per game the game it could cause manage mode to become strikeout fest. I adjusted zero sliders for mlbts 13, that's how realistic the game was in mange mode.

I'm talking from my experience running CPU vs. CPU games, which is essentially the same as games in MoM, so no user input there. I'm only agreeing with the impressions from people using their own stick skills, because I'm seeing the same thing in both.

[Armor and Sword]

Re: CPU plate discipline logic has been re-written

Out of the zone swing and misses would be more welcomed. But I am able to get borderline pitches....mainly lower in the zone, low and away and low and inside as swinging and looking strikes outs vs the CPu on classic pitching.

On 13 the improvement was super obvious from years past.


So 14 will only be better.