Starters v Relievers

[theobscure]

Originally Posted by mrbill

If only that could be quantified by some real stats that boiled down to a general rating across all pitchers, or to just one rating for each pitcher (hidden or not), I'd be pretty excited.

some kind of "pitching IQ" rating, you mean? I guess I.Q. isn't necessarily the right "word". It would cover a pitcher's ability to make adjustments, but then there are other factors like pitch assortment, ability to change speeds, and how willing they are to do their homework on opposing hitters.

Not to mention the impact that catchers have. The more you think about it the more you realize that no matter how many stats they come up with, there are so many things in baseball that go beyond their capability.

[koohead]

Originally Posted by theobscure

The more you think about it the more you realize that no matter how many stats they come up with, there are so many things in baseball that go beyond their capability.

And there my friends is the beauty of baseball and why we play the game

[cmdroverbite]

Originally Posted by RonCo

I have read many studies that attempt to prove clutch hitting, and never seen one that has bene successful. Do you have a link? I would love to be thrilled.

Have you seen this? http://www.dolphinsim.com/ratings/notes/clutch.html. I think that most modern-day statheads have looked at it and bought into the validity of its conclusions:

1. Clutch hitting is an important skill in baseball.
2. The difference between a good and a bad clutch performer is about 28% the difference between a good and a bad hitter, a much larger effect than had previously been thought from sabermetric work. So it is unlikely that any 0.250 hitters turn into 0.400 hitters in clutch situations, but there are 0.285 hitters who turn into 0.300 hitters.
3. Because of random effects, it is extremely difficult, if not impossible, to peg a specific player as a clutch performer or choker with a high degree of certainty. (For that matter, it is extremely difficult to ascertain much of anything about a player's batting skills to an accuracy better than 20 points of OBP based on one season's stats.)
4. That said, power hitters that perform better in the clutch are fairly rare, as are singles hitters that perform worse in the clutch. This can be used to make an educated guess of a player's clutch tendencies.

[Ktulu]

Doesn't #3 invalidate #1?

[bababui]

Originally Posted by Ktulu

Doesn't #3 invalidate #1?

No

3 just says its hard to predict who will be a 1...not that 1's dont exist

Edited to add: 1 should be a baseball postulate

[Ktulu]

If you can't say who is a clutch performer how can you say that clutch hitting is a skill? Clutch hitting exists, but that doesn't make it a repeatable skill.

[bababui]

Originally Posted by Ktulu

If you can't say who is a clutch performer how can you say that clutch hitting is a skill? Clutch hitting exists, but that doesn't make it a repeatable skill.

This is like that problem in physics where you can either know a photons location or its vector but never both at the same time. We know there is clutch hitting and we know that players do it but we can never say when it will happen.

I hope this clears things up.

[Ktulu]

Still, knowing there is clutch hitting obviously implies that players do it, but I don't see how you can automatically jump to the conclusion that clutch hitting is a repeatable skill.

[bababui]

Perhaps part of the problem is in the definition of clutch hitting. I was thinking of avg with runners in scoring position but the problem with this is in a blow out does driving in runs really constitute clutch hitting?

With precise definitions might it not be possible to make this determination?

[Ktulu]

It might... I just didn't see anything in that study that supported "Clutch hitting is an important skill in baseball." That is all.

He sorta just threw that conclusion out there without really justifying.

He goes from "Regardless of how one chooses to define "clutch" situations, it is clear that there is indeed a statistically significant difference between how players perform in clutch and non-clutch situations." to "The fact that clutch hitting is indeed a real skill in baseball does not necessarily mean it is an important one."

[Spielman]

Originally Posted by cmdroverbite

Have you seen this? http://www.dolphinsim.com/ratings/notes/clutch.html. I think that most modern-day statheads have looked at it and bought into the validity of its conclusions:
[/LIST]

Not from my experience, they haven't.

Another take on it here:
http://sports.espn.go.com/espn/page2...s/ortiz/060405

Interesting to me that while the list of players from the Prospectus guys includes a couple of the same guys mentioned by Dolphin, (Henderson and Gwynn) it also lauds players like Kent Hrbek, Jeromy Burnitz, and Dante Bichette, who apparently don't fit the Dolphin conclusion about power hitters.

[RonCo]

Originally Posted by cmdroverbite

Have you seen this? http://www.dolphinsim.com/ratings/notes/clutch.html. I think that most modern-day statheads have looked at it and bought into the validity of its conclusions:
[/LIST]

It's an interesting study. I had seen it before, yes, but had forgotten it. I don't believe many statheads have bought into the approach, honestly...but maybe I'm just not reading the right statheads. Fundamentally, the study uses a talent model to attempt to duplicate data populations--one model with a clutch element added, and another without. It's a valid analytical tool, so the study does carry interest.

It basically says that the best that can be said about clutch hitting is that its effect is so tiny that you can't see it over a season's span. I do believe any prudent stathead will agree that the studies that have been done do not rule out clutch hitting as a skill, but that they have said that _IF_ clutch hitting exists as a skill, it's too small to detect.

So, this study may play in harmony with that aspect.

On the other hand, I'm not steeped enough in the method he's using to give it a real critique. I do worry a little when he expains away a delta batting average of .009 points as not significant, but a delta OBP of 7 points as being iimportant.

In addition, if the impact is as large as this report suggests -- a quarter the size of the span between good hitter and bad hitter--then other methods employed by other Sabermetricians should be able to find it. The method linked in the ESPN article above suggests that it has found clutchness, but seems to kill itself in its own pages.

The other thing that bothered me about his method was that he didn't see a difference in the result as he varied the definition of clutch. Maybe these comments are related (significance of the gap between clutch haves an cclutch have nots, the proposed size of the impact of clutch, and inability to adequately expalin away the difference in average). I don't know. But ultimately, I guess what I'm saying is that I'm not smart enough to be convinced that this paper does what it says it's done. Just as in any other field, I need something to corraborate it.

It's an interesting start, though.

I'll have to think about this more.

However, I'll stand by the statement that most stat-heads have accepted clutch hitting as a skill is not correct. I read a bunch of them all the time, and "clutch is a skill" is not something that I've heard coming from their mouths to date.

[RonCo]

Originally Posted by Gorilla Shakespeare

That'll work too, but you'd want it for both starters and relievers in that case.

Like it or not, in real life relievers wind up with lower ERAs and the like than starters. This was the case even before LaRussa ball. In fact, it's a time-honored piece of Conventional Baseball Wisdom that some pitchers who are otherwise decent simply cannot hack it as starters. Next year, I'm hoping for improvements (one interesting wrinkle in particular is how Baseball Mogul handles pitch modelling) on this. For now, we need a way, any way, for starters to have a shot at going better than 17-6 in a season.

I've been looking around for some real research on this so we could turn "conventional wisdom" into something real. This study isn't absolutely perfect to fit the discussion, but it's close:

http://www.retrosheet.org/Research/SmithD/batlearn.pdf

It's a retrosheet study of 20 years of play-by-play data that shows batters improve their performance with every at bat they have against the same pitcher in the course of a single game. Logically, then, relief pitchers in real life would have better ERAs than starters merely due to the fact that they never (or very rarely) have to face a hitter a second time. The difference between a guy who sees only one pass through the order and one who hangs around for three passes is on the order of 5-6%. On an ERA of 4.00, that's between .2 and .24 of ERA degradation in starters that is explainable merely by the fact that they pitch 6-7 innings rather than 1-2.

[Dreslough]

Originally Posted by RonCo

Just against adding a batter-by-batter/pitch-by-pitch performance modifier that has never been proven to exist...

"The Book" by TangoTiger etc. shows strong statistical significance that pitchers become less effective with each batter:

http://www.amazon.com/gp/product/B00...lance&n=283155

And they find no evidence that its best to "jump on a pitcher early".

I've also seen studies that starters improve all their stats not just ERA when moved to the pen.

I think I've seen the first effect in Mogul but my sample size isn't that big yet. The second effect is definitely part of the game when you move a player to the pen or vice versa.

[bababui]

Nice to see another Mogul reference. Officially there are now more BM2007 references on these boards than on theirs.

[mrbill]

I wouldn't complain, exactly.

You've got a guy who is developing a baseball sim game contributing to the discussion of how to improve a competing game. He brings knowledge and experience (and links to works that inspire his model), and we'd all be wise to take what we can from him and get it applied to OOTP.

Not exactly a n00b coming in and asking for an MLB roster set :cool

[bababui]

Originally Posted by mrbill

I wouldn't complain, exactly.

You've got a guy who is developing a baseball sim game contributing to the discussion of how to improve a competing game. He brings knowledge and experience (and links to works that inspire his model), and we'd all be wise to take what we can from him and get it applied to OOTP.

Not exactly a n00b coming in and asking for an MLB roster set :cool

mr bill mr bill mr bill

'Dreslough' didnt know the city the real C Dreslough lived in until another poster filled him in. He is also getting some free advertising too for the game. Not a bad gimmick.

[Malleus Dei]

Originally Posted by RonCo

http://www.retrosheet.org/Research/SmithD/batlearn.pdf

It's a retrosheet study of 20 years of play-by-play data that shows batters improve their performance with every at bat they have against the same pitcher in the course of a single game.

Hitting is timing. The more of a pitcher someone sees, the better grasp they get of his timing. That's both obvious and logical.

[Malleus Dei]

Originally Posted by mrbill

You've got a guy who is developing a baseball sim game contributing to the discussion of how to improve a competing game.

Mr. Bill, how would you like to buy some swampland and a bridge?

[Malleus Dei]

Originally Posted by cmdroverbite

Have you seen this? http://www.dolphinsim.com/ratings/notes/clutch.html. I think that most modern-day statheads have looked at it and bought into the validity of its conclusions

To the best of my knowledge that statement is completely wrong.