Stats-Based CPU Sliders [MLB 13 Version]

[nomo17k]

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

I saw the third post of this thread you compared the speeds of the real life pitches to the pitch speed slider and discovered that pitch speed 10 is the most realisitic. I was wondering has anyone done a similar study for base running speed?

In fact, the 22nd post in this thread exactly deals with this.

I've tried but its hard to find good footage of a guy running out an infield hit so you could compare that with the same player ingame. I dont have MLB tv and my internet is even too slow to really do youtube but just curious. It seems rather important to me, many other sliders depend upon that one. I see you are using 5 in your tests, but other people use 3, and bobhead is going with 3 in his sliders.

There must be one number that is more accurate to real life than others just like with pitch speed right? I want to start tuning my sliders to get the right number of singles/doubles etc but not sure where to put the base running speed

I think 4 - 6 would stay within the real-life range. But making the choice that is most realistic in terms of speed like this doesn't always lead to the most realistic outcomes in the game.... because obviously not everything is truly realistic. For example, the new baserunning engine this year tries to simulate the slow down of the runners on the basepath rounding bases. Is the implementation rock solid? I actually don't think it is yet... the concept is good, but I think runners slow down too much when they shouldn't this year.

I'm not sure of the reasoning behind lowering BR Speed slider, but even at the extreme setting you wouldn't be that far off from reality.

[BrianU]

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

In fact, the 22nd post in this thread exactly deals with this.




I think 4 - 6 would stay within the real-life range. But making the choice that is most realistic in terms of speed like this doesn't always lead to the most realistic outcomes in the game.... because obviously not everything is truly realistic. For example, the new baserunning engine this year tries to simulate the slow down of the runners on the basepath rounding bases. Is the implementation rock solid? I actually don't think it is yet... the concept is good, but I think runners slow down too much when they shouldn't this year.

I'm not sure of the reasoning behind lowering BR Speed slider, but even at the extreme setting you wouldn't be that far off from reality.

I should have read a little further! you really covered all your bases thanks for pointing that out. I agree funky baserunning especially around 3rd does negate some from balancing the speed. I am just thinking along the lines of what speed would best represent not just running to 1st base but also double play oppurtunties, which is also connected to arm strength of course.

Im following the tests your currently putting up and I know that while the runs are low now and the 2B/HR stats are a bit off its only been ~10 games and checking the pitching matchups I expect those numbers to rise with more samples. I am curious what made you bring down throwing errors to 4, the error numbers seemed to look good at fielding 6/throwing 5.

it is great to see the progress from each tab as there is more information to interpret and we can get a better understanding of how the sliders and game engine work in conjunction. Keep up the good work!

[nomo17k]

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

I should have read a little further! you really covered all your bases thanks for pointing that out. I agree funky baserunning especially around 3rd does negate some from balancing the speed. I am just thinking along the lines of what speed would best represent not just running to 1st base but also double play oppurtunties, which is also connected to arm strength of course.

In the end, unless you have a specific goal of being precise with numbers or anything of that sort, what makes you feel pleased should be setting you should use. For most, default would be fine.

Im following the tests your currently putting up and I know that while the runs are low now and the 2B/HR stats are a bit off its only been ~10 games and checking the pitching matchups I expect those numbers to rise with more samples. I am curious what made you bring down throwing errors to 4, the error numbers seemed to look good at fielding 6/throwing 5.

The change isn't really necessary... I just lowered it to see how much it goes down... I'm actually looking at (throwing) errors per balls in play, and it's been a bit inflated compared to the MLB ave... it's been so consistently, so I thought I might just try. It's not a necessary change. Per game per team number can be misleading when the overall offense number isn't thoroughly balanced... when the offense was up like it was for my last few test sets, you expect more balls put in plays. Ideally you adjust for it, and it can be done like this.

But in general, the game seems to produce more throwing errors than fielding errors. It was like that last year as well.

it is great to see the progress from each tab as there is more information to interpret and we can get a better understanding of how the sliders and game engine work in conjunction. Keep up the good work!

This game is such a good baseball simulation that I actually like seeing baseball the sport through the eyes of someone who makes the game, "how do you make things realistic?" Makes you think about what baseball is about... that's part of the fun.

[OliDegu2008]

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

Hey Nomo, thanks for the hard work on these. I am new to the CPU vs CPU aspect but I was wondering, do you foresee any major changes to the set out now. I'd like to start a chise with em, but will wait if you think there will be any big changes in the foreseeable future.

Thanks and keep up the awesome work.

[nomo17k]

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

Hey Nomo, thanks for the hard work on these. I am new to the CPU vs CPU aspect but I was wondering, do you foresee any major changes to the set out now. I'd like to start a chise with em, but will wait if you think there will be any big changes in the foreseeable future.

Thanks and keep up the awesome work.

Just updated the second post in the thread for my "suggested" slider set. It's still not final but I don't foresee any major changes (i.e., more than one click) now.

Might tweak error and BR/speed related sliders but I think those will be within the range where the gameplay won't be affected that much.

I think the biggest unknown still is where Pitcher Control will be... but then I think it will be either 2 (current) or 3, depending on how stats stabilize after quite a few more games. But so far I'm liking what I'm seeing... walks are generated but the game overall plays very tight. Hopefully things work out along this line.

[ralphieboy11]

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

Looking good, Nomo. My sliders are going to be similar so it's nice to have your stats to fall back on to strengthen my belief in what I'm seeing.

For what it's worth...

In my recent tests, I played around with fielder reaction. I was looking for a way to drop batting average without affecting too much else. Through almost 50 games, by moving it from default to 7, I've seen batting average drop by around 20 points. Singles really seem to be the only area affected. So it's the quicker reactions of the infielders that appear to be taking away a couple hits a game. I guess this could also be outfielders catching shallow fly balls a bit better as well.

I haven't watched a lot of these games but from what I have seen this adjustment does not make fielders look superhuman. In fact, on default this year I felt like the infield reactions seemed a tad slow.

So although this was a small sample size I do feel comfortable in saying that a slight increase in fielder reaction should drop batting averages a few points if anyone is looking to do that. I will probably leave it at default or 6 to start one of my seasons and then I'll know that I can use it to adjust averages slightly if I need to.

[BrianU]

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

I like the idea of raising fielder reaction a notch or two just from an eye test point of view. You can't test it as with base running speed as easily, its more of checking a replay and asking yourself if that situation made sense and if the fielder would reach the ball IRL. It seems others have felt the need to raise it too. Of course you don't wanna do that at the sake of throwing singles out of balance for right now im leaving it at 5 as I start playing but would like to possibly raise it

How do you feel about relief pitcher stamina nomo? I am curious about that one. From the number of pitchers data the CPU it looks like they are using a bit too many pitchers throughout the game. Starter IP's don't seem out of whack but the relievers could use a look at. I'm not as alert on what changes sliders make precisely but in my head it would seem if you raised relief pitcher stamina then that would decrease the amount of different RPs used? Looking at the sliders most people go with for USER vs CPU the tendency is strongly to lower this slider. What is your take?

[nomo17k]

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

Looking good, Nomo. My sliders are going to be similar so it's nice to have your stats to fall back on to strengthen my belief in what I'm seeing.

Are you still on Strike Frequency 3? Wonder if the offense stays at the right level. I don't know what generated the offense inflation in my tests haha...

I think the biggest difference between yours and mine would be strike %s in various counts. Strike Frequency (supposedly) only changes strike % in early counts... I've actually been seeing the biggest difference in full-count strike % by varying Pitcher Control slider. Also, lowering Pitcher Control seems to lead to more meat balls. Offense has been actually down so far with my current set, but XBHs are generated, which weren't the case previously.

For what it's worth...

In my recent tests, I played around with fielder reaction. I was looking for a way to drop batting average without affecting too much else. Through almost 50 games, by moving it from default to 7, I've seen batting average drop by around 20 points. Singles really seem to be the only area affected. So it's the quicker reactions of the infielders that appear to be taking away a couple hits a game. I guess this could also be outfielders catching shallow fly balls a bit better as well.

I haven't watched a lot of these games but from what I have seen this adjustment does not make fielders look superhuman. In fact, on default this year I felt like the infield reactions seemed a tad slow.

So although this was a small sample size I do feel comfortable in saying that a slight increase in fielder reaction should drop batting averages a few points if anyone is looking to do that. I will probably leave it at default or 6 to start one of my seasons and then I'll know that I can use it to adjust averages slightly if I need to.

I've been thinking of measuring reaction time with videos... have not done yet since I don't know if that would be useful.

I think it's actually possible to adjust Fielder Reaction slider using something like range factor stats for infielders in the game. I might revisit this since the stats has been kept in my season file...

[nomo17k]

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

...

How do you feel about relief pitcher stamina nomo? I am curious about that one. From the number of pitchers data the CPU it looks like they are using a bit too many pitchers throughout the game. Starter IP's don't seem out of whack but the relievers could use a look at. I'm not as alert on what changes sliders make precisely but in my head it would seem if you raised relief pitcher stamina then that would decrease the amount of different RPs used? Looking at the sliders most people go with for USER vs CPU the tendency is strongly to lower this slider. What is your take?

I haven't touched relief pitcher stamina only because I don't know what the best stats to use to adjust them. If I were to adjust this, I might do this via looking at IPs or games pitches in simmed season...

I actually never understood why people like to lower RP stamina. My fear is that by making them less effective quickly (I think RP become significantly less effective around 50% energy mark... don't remember exactly but I thought that's when the ticker alerts the user for RP being tired?), we will see a lot more low-confidence meltdown (some call this CPU comeback). Relief pitchers are quite ineffective already in the game, so I don't really like to make them even less effective.

[ralphieboy11]

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

Are you still on Strike Frequency 3? Wonder if the offense stays at the right level. I don't know what generated the offense inflation in my tests haha...

I think the biggest difference between yours and mine would be strike %s in various counts. Strike Frequency (supposedly) only changes strike % in early counts... I've actually been seeing the biggest difference in full-count strike % by varying Pitcher Control slider. Also, lowering Pitcher Control seems to lead to more meat balls. Offense has been actually down so far with my current set, but XBHs are generated, which weren't the case previously.

I am still at Strike Frequency 3 but I'm certainly not married to it. I've basically been keeping it there to maintain some consistencies in my results. I think you are right about the strike %s in different counts. I have not been keeping as detailed of stats as you, but I have been keeping overall strike % and it is perfectly matching real life at 3. I think we can deduce from your findings that I'm probably seeing a few more balls in early counts, and a few more strikes in later counts. In the end it is evening out to come to the perfect percentage overall.

I would not mind a few more meat balls myself. I still had pitcher control at default, and as I was testing fielder reaction I noticed my home run numbers starting dipping slightly. It's not unacceptable now, but in order to get it back up to major league average the pitcher control slider tweak might help.

[BrianU]

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

I haven't touched relief pitcher stamina only because I don't know what the best stats to use to adjust them. If I were to adjust this, I might do this via looking at IPs or games pitches in simmed season...

I actually never understood why people like to lower RP stamina. My fear is that by making them less effective quickly (I think RP become significantly less effective around 50% energy mark... don't remember exactly but I thought that's when the ticker alerts the user for RP being tired?), we will see a lot more low-confidence meltdown (some call this CPU comeback). Relief pitchers are quite ineffective already in the game, so I don't really like to make them even less effective.

Would the # of pitchers used per team which you are tracking be useful? You would have to be mindful of the SP ip if you adjust anything but you could always compensate with SP stamina.

Right now it appears the CPU is using an extra pitcher on average every game. Would raising RP stamina have the effect of them using less pitchers overall? I agree I am hesitant in lowering the RP Stamina. Other people have it dropped severely into the 1 and 2 mark. If you were to adjust manager hook also in conjunction perhaps you could avoid the pitcher staying in the game when he is out of energy, but get pulled at the proper time to get the numbers to match up.

The problem I forsee is from what I read of Manager hook it causes the manager to adjust pulling you not just based on the pitchers energy, but also based on how many runs they are giving up. So that could cause unwanted effects. Could raising RP stamina slightly and adjusting manager hook slightly do the trick? ill let it go for now just throwing some ideas out there

[ralphieboy11]

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

I assume those that were lowering reliever stamina were doing so to prevent themselves from using their best relievers game after game. I can understand that logic but in regards to how the cpu handles the bullpen, I think you will see some strange things if you do that to cpu reliever stamina over the course of a season.

At default 5, less than two weeks into the season, I've already seen the Pirates bullpen become extremely tired during a stretch. During one extra inning game they had used all of their relievers, so they turned to the next day's starter at the end of the game. We saw this same pattern in last year's game.

When everyone in the bullpen is tired, the cpu will sometimes bring a guy in to pitch to only one batter. It's almost as if the logic has a little bit of a breakdown. The manager AI brings the new pitcher in, but after the at bat says "This guy is too tired. Let's bring in someone else." Eventually all reliever options are used, and the only option left is a starter.

I think these situations will be rare, but to avoid them my tendency is to bump cpu reliever stamina up.

[BrianU]

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

I assume those that were lowering reliever stamina were doing so to prevent themselves from using their best relievers game after game. I can understand that logic but in regards to how the cpu handles the bullpen, I think you will see some strange things if you do that to cpu reliever stamina over the course of a season.

At default 5, less than two weeks into the season, I've already seen the Pirates bullpen become extremely tired during a stretch. During one extra inning game they had used all of their relievers, so they turned to the next day's starter at the end of the game. We saw this same pattern in last year's game.

When everyone in the bullpen is tired, the cpu will sometimes bring a guy in to pitch to only one batter. It's almost as if the logic has a little bit of a breakdown. The manager AI brings the new pitcher in, but after the at bat says "This guy is too tired. Let's bring in someone else." Eventually all reliever options are used, and the only option left is a starter.

I think these situations will be rare, but to avoid them my tendency is to bump cpu reliever stamina up.

Is that situation with the pirates you describe during a simmed series or a series which you played all the games against them? I wanna run some tests on simmed games because I think CPU Manager Hook and the pitcher stamina sliders might not to have different values depending on if you sim a game or play a game. Even if you play all 162 games in your own franchise the games where the CPU vs CPU sims are happening may need different sliders.

I have two saved sliders one I load for when I sim any game, and one I load for when I play game. Right now the difference is injury frequency 8 for played and 10 for simmed and everything else to 5 but as I said I want to check out the other sliders too.

[ralphieboy11]

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

The situation I described is where I'm playing every team's games CPU vs CPU. No simming. So for most people that do sim games I doubt it would ever show up.

In years past I try to play out as many games as possible. When I do sim games I put all sliders back to default.

[BrianU]

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

The situation I described is where I'm playing every team's games CPU vs CPU. No simming. So for most people that do sim games I doubt it would ever show up.

In years past I try to play out as many games as possible. When I do sim games I put all sliders back to default.

I got you, yeah I made another slider set all default except injury frequency 10 for simming. I'm going to keep my eye on reliever usage by the CPU it is a tough thing to nail without playing the games. I had an idea of possibly just switching my controller over to the CPU team, making a pitching change when appropriate for them but i'd like to see if I can get them to act more realistically with sliders first