Pitch Question

[PoopyJoe69]

I know there was probably a thread on this somewhere, but I was wondering if it was better to have two good pitches and a crappy 3rd pitch (i.e. changeup, slider) or just have the two good pitches.

[Padreman]

Well the more the better I guy who throws two good pitches and one mediocre pitch he's Will be more successful than the guy who only throws to

[OldFatGuy]

In real life or OOTP?

I guess you're asking in OOTP and probably the only ones knowing the real answer are those that programmed the game but....

I think the answer depends on the role. If you're using the guy as a reliever, it's probably better to have just the two or it makes no difference. If you're using the guy as a starter, it's probably better to have a crappy third than no third. But I have no data to back this up, so take it with some serious grains of salt.

[andyhdz]

I agree with the oldfatguy. lol love your username.

[NoOne]

if it's a poorly rated pitch, i thnk they are better off without it. assuming accurate ratings -- lower 1/3 to 2/5ths of your scale and worse... plus the higher portion of that range is really bad because unlike a "12/100," it will be thrown occasionally.

if it's ~1/2 of the scale, it's a benefit for sure, even for a reliever. it's essential for a starting pitcher to have 3 or more pitches. you can have good seasonsn with a dominant 2-pitch starter, but it will take numerous attempts before you get one lucky enough to string together consistently great years relative to their amazing ratings. it's virtually a 50/50 coin flip for them each season (a 2-pitch sp).

funny, i just posted a similar reply... i have no qualms removing a crappy pitch from one of my prospects... i consider it a demand from the organization. however, siince i have scouting inaccuracy on, i have a self-imposed rule that any decision i make to remove a pitch before i open the editor must be followed through.

[BusterKing]

Originally Posted by NoOne

i have no qualms removing a crappy pitch from one of my prospects... i consider it a demand from the organization. however, siince i have scouting inaccuracy on, i have a self-imposed rule that any decision i make to remove a pitch before i open the editor must be followed through.

see I wouldn't do that, I don't like taking advantage of the AI, the AI controlled teams don't do that to their players.

[monkeystyxx]

Disclaimer: I don't know what I'm doing and this is probably useless data.

So, I'm not sure if this is the right way to go about it, but go about it I have. I took Clayton Kershaw, and ran three separate tests (with scouting, injuries, coaching and morale all turned off to remove as many external factors as possible).
One, 'Controlshaw' perfectly normally as he is in the default data.
Two, 'Threeton Pitchshaw' with no changeup.
One, "Badshaw" with a terrible changeup (20 rated in the editor, 25 rated on his profile).

Everything about the game setups were identical except the rating of Kershaw's changeup.

I've runs 5 separate game files for each version of Kershasw, lasting 2 seasons each (because after that he tends to go to a team with vastly different characteristics which isn't great for comparisons. For instance, Controlshaw went to Houston once and had an ERA over 3.5, which was the only ERA over 2.8 he ever had, so I deleted all Season 3 data after that and didn't bother simming past S2 in future). This gives me ten overall seasons of data for each version of the player. To clarify, only one version of Kershaw existed in each game world.

A few notes before the results:
1, in both edited versions of Kershaw, his Stuff dropped from 75 to 60 (as you should be aware, Stuff is a rating which the game calculates from other ratings, including velocity and individual pitch ratings).

2, the in-editor Stuff rating dropped more when I removed the pitch entirely than when I ruined it (this is pretty logical, something adds more value than nothing at all, but obviously the game engine doesn't just use the calculated stuff rating so this isn't empirical proof that having a bad 4th pitch is better).

3, the in-editor ERA projection was slightly lower for bad pitch than for 3-pitch, but not by a lot. Both were significantly above the normal Kershaw.

The results themselves were surprising to me...



Now, I need to reiterate that this probably isn't how you're "supposed" to do it, I'm not a "maths guy", but I found them interesting nonetheless.

The first thing that immediately jumped out at me was that his control in both of the modified versions was improved (worse with a bad 4th pitch than none at all, but still better than regular Kersh). I'm assuming this has something to do with other stats I don't know enough about which tell the game engine what to expect Kershaw to do. His WHIP is extremely close in all three versions. I'd assume that because his lack of a good 4th pitch reduces his overall Ks, he has to be better with his control in order to average out as the game expects from his other ratings?

His home run numbers were functionally the exact same across all three tests, which again is expected because that's tied to movement, which would've been unaffected. Maybe you'd expect slightly more homers hit off his terrible changeup, but that didn't show in the results. In fact, the Kershaw with the terrible changeup gave up the FEWEST homers of all the Kershaws, but not by a margin significant enough to show in his HR/9. Of course, me being an idiot I included the raw BB and K numbers but not the raw HRs...

His WAR was interesting, being the exact same on average with a bad 4th pitch vs a normal 4th pitch, but it was worse than both when he only had 3 pitches.

The numbers seem to suggest that, while there isn't as big of a difference as I'd expect (possibly due to the above control thing? Or the admittedly mediocre sample size?), it seems that having a terrible 4th pitch is, in fact, marginally better than only having three?

I also don't know how to interpret the fact that on average, the edited Kershaws struck out slightly more pitchers in their first season than the control Kersh, but then significantly fewer in their second season.

Oh, and I'm really sorry that I missed the column headings off the averages on the left. I'm almost not-lazy enough to fix it, but not quite.

Some ideas for further tests:
1, fiddling with his other pitch values and his velocity, to get the in-game Stuff rating back near where it is by default (for instance, lowering his changeup but increasing the other pitches slightly to compensate). One obvious flaw in this data is that by lowering his changeup or removing it entirely, I'm objectively making Kershaw worse. So it's hard to draw meaningful comparisons between the control and the other two versions (though comparisons between the two edited versions should still be valid).

[NoOne]

Originally Posted by BusterKing

see I wouldn't do that, I don't like taking advantage of the AI, the AI controlled teams don't do that to their players.

nah, that's just for my teams.

but, if i happen to see a really good talent with a bum pitch (like ~10potential out of 200), i will remove it... it's not taking advantage... i am improving a competitor's pitcher, lol.

i don't often touch other team's players for this reason... but if it's a potential HoF or record-setting talent, i'll fix them for sure... that crap pitch does nothign bt hurt them (*low on scale, not a ~mediocre+ pitch).

[NoOne]

the stats may or may not be off by a statistically significant amount given that sample size... could easily accomodate for what you see..

stuff is a conglomerate of many other things, so like overall, it is a summary and different inputs can add up to the same summarized value. in this type of experiement i wouldn't even track stuff or let it influence how you look at results by accident from knowing the resulting change. the pitcher's stuff rating is never used... the individual pitch thrown is the input during an AB, along what everythign else that goes into it at that incremental step.

different inputs = different results despite the same "Stuff" ratings. pitches have their own profiles. there's just many too variations of pitch selections you'd have to look into. also different ratios of resulting individual pitch ratings, since that influences how often they throw it.. then, if you make too many edited players to try to speed up the process of learning you also will change the league too much to give good data at some point.

heh... real players are a totally different animal than fictional, too... so the info may or may not translate 30 sim-years down the road, either. now, eras... remove power, add power, blah blah... this is a hornets nest with no answer except for the very specific situation you test, which may not apply to many others, lol.

luckily....

the lower the rated the pitch, the less it's thrown... if they have a 10/100 pithc, it's frequency of use should be relatively inconsequential (unless you are reading the pbp in the WS and see it thrown, lol, you may think of it differently).

i remove it to avoid the bad-luck lottery game,but it's inconsequential to the bigger picture for sure. (99.99% players in my organization)

you could argue that any mediocre pitch (sub-1/2 scale, but not total poop), is that much worse in the playoffs against competition who's baseline average shifts up significantly from regular season talent. many more players that mash those mediocre pitches, possibly?

anyway, a bad pitch, is a bad pitch.. it can't help much unless it's the "3rd" pitch for a SP, and it should be near 1/2 scale at least (under-the-hood, not what you see).

the most dangerous "bad pitch" is the one rated high enough that it's actually used more than rarely. that's the happy zone you'll eventually find. (or, is that anti-happy zone?)