pitch removal for middle relievers
 

the online league I am in has instituted a pitch removal rule. 2 mr's can each have 1 pitch removed. my question is- does this pitch removal help a mr improve/

e.g. fb-10 cb-9 changeup-2 does removing a 2 changeup mean that the mr will not hurt his effectiveness by throwing a weak pitch in his arsenal?

Removing a weak pitch will not help the pitcher be better at all.

The pitch ratings make up the stuff rating of the pitcher. There is no circumstance where the presence of a pitch makes the pitcher's rating worse. A "bad" pitch isn't going to help much, but it will not hurt.

the overall stuff is irrelevant.

a pitch's rating is what matters. if they throw the weaker pitch, it can be a detrimental thing... the arguement from ootp in past threads is that they throw that 1/100 pitch so seldomnly that it doesn't matter, either way... until the 1 time they throw it in the WS for no reason other than RNG and the very fact it exists, when in RL a pitcher would never throw such a weak pitch in any game, ever.

but, the naked eye won't tell the difference... maybe with a less extreme context it would be worthwhile to cut that extremely weak, but high enough to be thrown more than 1/10000 pitches etc.

whether the overall stuff goes up or down is irrelevant. the game doesn't even use that value for anything. it's correlated, but not causal in any meaningful way.

it can only hurt, but it hurts very seldomly is the actual answer, becaue it is thrown and when it is thrown it has a very high % chance of being hammered. that game would argue with you about just how important it is, lol.

i have a house rule... if i decide to cut somethign out, there's no return... i do it no matter what i see in editor. i never do this for a sub-28 aged guy or i've already made a decisions he's in the pen for other reasons etc. this isn't a stretch from reality... players are property, no matter how you look at it... they do as they are told or they can be buried and blacklisted, lol. a nail that stands out gets hammered. outside of natural talent they are born with and the efforst to develope it, they have no real choices about their career. tehy get drafted, then they are owned for a decade...

that makes sense

I created a 4-team league with:

- all players having identical ratings
- neutral ballparks.
- all pitchers in the league had Fastballs, Sliders, and Curveballs rated at 140/200 ... so pretty good.
- two teams I added a fourth pitch Changeup = 20/200 ... so pretty bad

In other words, the only difference in these teams was that two teams had every pitcher with a crappy side-car changeup. I then ran a season and compiled the results. Here's the data:

ERA: No Change = 6.26, Bad Change = 5.82
K/9: No Change = 6.97, Bad Change = 7.61
BB/9: No Change = 3.78, Bad Change = 3.48
HR/9: No Change = 1.28, Bad Change = 1.28

Bottom line, this data says that a crappy sidecar change-up makes a pitcher better. To be honest, I'd be shocked if I found a different result, but it's nice to see.

A subpar pitch that a pitcher may only use less than 1% of the time can actually be a positive due to the surprise nature of it. Even a hanging curveball can fool a hitter if the hitter isn't expecting it and has never seen one thrown by that pitcher. So I would disagree that the only way to look at it is in a vacuum, as it should be evaluated as a part of his entire repertoire. A ****ty pitch that is featured often by a pitcher would definitely be a bad thing. However, a subpar pitch used sparingly, just to show it to hitters and make them think about it, can have a positive influence on how effective that pitcher is.

NoOne, I believe that you are misinformed in how OOTP uses the Pitch/Stuff ratings. The coding behind the pitcher/batter matchup only uses the Stuff rating, not the individual pitch ratings. There is nothing in the programming where if the "bad" pitch is chosen, it has a more likely to be hammered. The pitch ratings (plus velocity) are what determine the Stuff rating. The Stuff rating will always be "better" with more pitches, no matter their quality.

is surprise modeled? in RL, i agree to a point... that euphis pitch didn't do so well, lol.... you throw a crap changeu p in the RL MLB and it's a HR often.

there's another factor at play. the efect of multiple pitches and however ootp fits that in, obviosuly.. although, if that's just one season, i doubt that's a suitable sample for what we are discussing.

1 pitch thrown every 1000 pitches is going to take a very long time to deduce what that 1 pitch's weight is truely worth... unless there's a significant 'other' factor invovled.

i think 3-pitch relievers got a boost last year? if you haven't tried one with ~max stuff and good otherwise, you really should in the pen.. they are ~1.00ERA and 130ip, lol.

it's some shortcut in the stat engine, if it is better in game to have a horrendous pitch, lol. some band-aid fix for somethign else that allows it to be true.


Rainking:

i don't mean the raw stuff rating from editor.. i mean the profile values sans inaccuracy, because it includes gb% and velo etc.. whatever adds up to an individual pitch's stuff is on the profile. (and futhermore vsL or vsR is used, of course)

well, from how they explained it in a previous, that pitch is thrown and ti's not important because of how often it is so it won't affect gameplay.. unless they were just glossing over while trying to explain in less direct way why it doesn't matter. if the pitch is thrown, it's not using overall stuff, it's using the stuff for that pitch (not raw stuff).

if it doesn't match up pitch thrown, that would be horrendous, lol. that along with a single seed rolled for outcome per AB (at least for hit/out and type) would just be a big joke, lol.

the more i think about it... it does makes sense relative to how to abuse the game... i'd love to test that out... get a pitcher with a really weak pitch and reload when i have a consistent out 'Seed' and see if it's different than in past experiences. it wouldn't be a doomed out, if they eventualyl throw that weakt-tea pitch. a little hexeditor could make this easy to figure out while playing. from the rampant cheating i heard about earlier in year (still?) it sounds like someone knows the memory address for the seed generated.

this is truly unfortunate, lol. really glad i don't play games out, if this is true. you might as well jsut spam "1" and get it done quicker, lol.

I know you're not a big fan of the value of understanding the Log5 equations, but if you do, you'll understand why/how it uses overall stuff rather than individual pitches. Of course, maybe I'm wrong. I'll leave it to anyone else to make their own calls.

ron, that isn't even relevant to my issue with it. i don't have a problem with the plug-n-play equation they are using to determine outcome, unelss it's per PA and it is all summarized... yes, similar overall outcomes guarnateed, but not the same as far as getting there.

mostly stream of thought below, but reasoning as to why, too.

it's about resolution of what is being used.. it's cutting a corner early instead of fleshing it out further.. can the math say the same average at the end while doing less, yes. does the game do it inother areas, yes. is it acceptable in many areas, yes... just depends on how far down the rabit hole you go with every relevant factor they chose to code. relative to this concpet, to flesh it all out you'd have to have understanding down to the individual atoms invovled... it can get that deep or even deeper, depending on how accurate our udnerstanding of the phyiscal world truly is. a corner is inevitably cut for a video game and due to lack of knoweldge, but not so early as per PA.

the equation doesn't care of you use summarized or individual... no extra work, not any harder to code etc etc... only reason it wouldn't 'fit' is if the incremental step is as large as 1PA. it all has to be @ the same rate, each incremental step or lap through the logical loop that is 1 PA or 1 pitch thrown etc.

if it's only resolved down to "all pitches summed up" and used, that's the issue i take. it's the resolution of a factor (pitch thrown), not the math. the math stays the same, regardless of resolution. hence, no more difficult to plug them in individualyl as they are used unless it does jive with some other factor that isn't resloved beyond per pitch thrown reso, whic is 100% avoidable and planned out from beginning, not an accident at that point, of course.

also, if the game only resolves down to PA, it makes many decisions you make irrelevant. (this would make sense if it does use a summarized stuff, and lines up with seed being generated once) my rational brain tells me it's summarized as i type this paragraph.. lol.. poop on that. feelings are going to make me irrational about it, lol. this game likely chugs along per pa... anythign beyond that is beyond it's resolution and just asthetic in nature... ugh, wish i hadn't thought about it, now.

however, there's some hope for me:

making a KB drastically different would be more difficult on a 1-200 scale as opposed to the resolving power of each pitch considred per throw (=200^x, conversely). instead of playing it out individually, it's averaged (weighted) to come up with a single 1-200 score that will represent every single pitch combination, gb%, velo possible. so anythign that equates to ~90stuff is the same (ignoring handedness for this, same logic applied, same math idea as i mention above, just simpler, less resolved). if we make 3 pitchers with slightly different repertoires and all 90stuff they should have the same results -- same babip ag, same k/9, same bb/9 etc etc... everything possible would be the ~same, with expected volatility.

i'm betting the guy with the KB doesn't have simialr results in all respects. if it adds up to 90, given all other factors, and 90 is used, then it should have the same results, all other things remaining the same. i'm betting the babip is significantly lower with a KB involved, at minimum, and you cannot find any other combination of 3 pitches that has the same results at same stuff rating as one repertoire that includes the KB. (could still make that happen with 1-200, you'd have a section dedicated to just the KB -- if that's true i'd get physically sick to my stomach, lol...the work required to do that is > than just resolving it per pitch, lol.)

also, it'd be more difficult for the game to resolve 3 pitches that equate to the same stuff but are in different proportions. 2 excelent and 1 ~1/2 scale... or 3 above average that equate to same ~X stuff? there'd be no way to differentiate between the two outcomes, because they are the same # used in the math to begin with. instead of easy math and threshold, yo'd have to have sections dedicated to that combination... it'd just be a mess of nonsense trying to cram all that into 1-200 and expect enough variation in results that gives an hint that each pitch type matters.

i think we all agree that 2 better pitches and one average pitch almost certainly has better outcomes than 3 ~equal pitches that add up to the same stuff. could be wrong on that, but that's my experience too.

if you have experience with a KB pitcher, you know they are a different animal even if they have the same stuff as another pitcher.

still on the fence.. it certainly isn't 100% clear either way just looking at it without the proper tools and approach to flesh it out.

i think whatever bandaid they have for # of pitches affecting outcomes is the real factor here. 3-pitch rp got a 'boost' several patches ago ('18? '17?). they are way too good as RP now, imo. i've used a number of 3-ptich RP that blow 2-pitgh guys away with lower stuff, lower movement and lower control, lol. accross the board lower rated vsL and vsR, yet they are consistently RP of the year type seasons. i can get 3-4 war out of 100-130ip with them. very few 2-pitch guys are capable of that in the pen, yet i can find multiple 3-pitch guys that can do it every single year. as i play thi sgame more, the more i realize how many "bandaids" exist... really glad i don't play out games. can't really tell from a season-wide view.

what i see with kb'er gives me hope, and i don't play games out. so, the drawbacks to such a limited resoultion wouldn't affect what i see... but, on principle of a model, that's ridonkulous if it is that way due to how easy it would be to gear it per pitch thrown and how much more accurate to life that would be by doing it that way. but, it makes no sense that a random seed is only generated 1 per PA, either... it's gotta be a bandaid to cover up some flaw, or i just can't see why it wouldn't go the obvios step further for a modicum of effort..


side note --- Quotes from bill james, and english major:

"My math skills are limited and my data-processing skills are essentially nonexistent. The younger guys are way, way beyond me in those areas. I’m fine with that, and I don’t struggle against it, and I hope that I don’t deny them credit for what they can do that I can’t."

"I hope I’m summing this up effectively — is that WAR does not connect directly to wins. "

the most hilarious thing about 'moneyball' is watching people that don't really get statistical analysis try to use it like a hammer. it's nice to read that bill james recongnizes the fundamental problems with his metrics. i have a totally differnet perspective of him, now. it's plebs blindly using his work without the fundamental and necessary knowledge that is the problem.

I believe you are either thinking too hard, or not hard enough. :)

have you even taken calculus? graduate degree? unless you are a math major or better, my grasp of these concepts very unlikely worse than yours. i aced my calculus and statistics classes, albeit ~20 years ago. it's actually very easy stuff to learn. i didn't even have to study. just went to class and took exams -- that easy.

resolution, bud... if the model resolves down to 1 PA, it can't see beyond that. most everything beyond that is just asthetics for your eye and meaningless otherwise, if true. i wouldn't doubt more than a few things you do in a game are irrelevant, or just to provide immersion but no real function as far as affecting outcome.

it's almost certainly simpler than what the facade shows... more rounded off. less resolution than alluded to.

Well, if we're comparing resumes I think I'm okay. Engineer by degree, and a not inconsiderable focus on the mathematics of statistics over many years of process design and management. I also admit I am aware of some of the details of OOTP's defensive schemes--though I would be reluctant to say I really know anything for certain because I don't actually code the game, and I don't think I'm allowed to say much more anyway. I have, however, been beta testing for, I don't know, 15 years on and off...and I did what must be the most comprehensive defensive study in the history of the game (back some years ago--though I think the online league I did it in is now defunct so all the links that used to exist around here are now dead).

That said, here's a post I recently made on my online league's forum that compiles defensive performance based on a script I run against game logs. The concept of this script is a little watered down from the full-on study I did back then, but it's in the same direction. The old approach could trace things to individual players. This only goes to the team level. That said, it's pretty quantitative when it comes to defense and it's values. Perhaps I could do better...but...shrug.

http://montybrewster.net/forums/view...p?f=23&t=26826

There are others on the site that look at specifics if you dig around a little.

And finally, yes, I've been reading Bill James since he was publishing his old Baseball Abstracts. Still have them on my bookshelf and occasionally read only articles just because he's a helluva writer and it's fun to hear him put words together.

The magic question is "How does the engine resolve a plate appearance?"

The answer seems to me to be both obvious and highly complex, and, in general, leads to results that are--in most cases--as deeply nuanced as baseball itself.

so how'd you misread what i said? lol.

it's programming, it's planned out. it's not a maybe. either it steps per pitch or it steps per PA, and it was designed that way ahead of time.

(mba-finance, computer science undergrad with business minor, elective short or so of chem/bio minors, but that was just for fun while there)

are you guys serious

I think this can easily be resolved by revealing the unrefutable truth:
OOTP is magic!