All numbers presented herein are based off 2009 American League totals, but I do not expect them to differ much in previous years or in the National League.
Pitchers with high ground ball rates are good. Like Felix! Felix is good. You should prefer ground balls to fly balls. Fly balls are bad. The standard rationale goes like this: home runs allowed per fly ball is a stat that does not appear to be under much control by the pitcher and since ground balls never leave the yard, the more balls you keep out of the air, the fewer home runs you allow. Not allowing home runs is good.
Some people counter that argument with the fact that ground balls go for a hit more often than fly balls. While true, the difference between the two is vastly overstated. Strictly between ground balls and fly balls (excluding bunts, home runs, line drives and infield flies) ground balls go for a hit about 3% more often. That is not much.
It is even less once you factor in the salient facts that ground balls become double plays far more often, non ground-balls see a higher slugging percentage allowed on average and that excluding line drives is a misleading distinction. You can read this all and just believe it, but how about some handy numbers to give you some nice foundation support when you go to convince others?
The average ground ball (including bunts) generated 0.04 runs and 0.80 outs
The average non-ground ball (fly balls, line drives, pop ups) generated 0.23 runs and 0.62 outs
On a runs per out basis, balls in the air generate almost 7.5 times as much offense as balls on the ground do.
Non-ground balls were:
A fly ball 49.3% of the time
A line drive 31.5% of the time
A pop up 12.5% of the time
A home run 6.7% of the time
Ground balls were:
A ground ball 96.3% of the time
A bunt 3.7% of the time
Out factors were derived from average number of outs recorded per batted ball type. Run factors were derived from changes in score and the run expectancy matrix after each play per batted ball type. Errors were factored in as well.
7 recs | 124 comments
One arguement I do see sometimes
is that pitching up in the zone increases strike outs but also increases flyballs. If I am a pitcher, how much do I need to increase my strikeouts by pitching up in the zone if I decrease my GB% from 47.5% to 42.5%?
Edgar for Pres - February 6, 2010
Good question.
It’s tough because of how inter-dependent these all are, but some back of the envelope numbers would point toward needing roughly an 8% increase. If you had the SP average K rate of 16%, it would need to increase to about 17.3% to offset the loss of ground balls.
Matthew - February 6, 2010
Interesting
That is a pretty reasonable shift in K% to make up for the GB% drop. Its interesting at least since its in the realm of possibility. Its not like a drop in GB% requires a minuscule or massive change in K% to come out even.
Edgar for Pres - February 6, 2010
Somewhere, Brian Bannister is furiously scribbling down notes.
The Typical Idiot Fan - February 7, 2010
In his mom's basement.
killer_ewok18 - February 7, 2010 via mobile
So for a pitcher with "average" GB rates
Are you saying that a change in pitching style that induced more flyballs would have to be accompanied by an 8% increase in Ks to compensate for each 5% increase in FB%? That doesn’t seem right if FBs have 7.5 times the offensive contribution, so i must be misunderstanding something.
wandergeist - February 7, 2010
Strikeouts are pretty devastating from a runs scored per out basis.
Matthew - February 7, 2010
True enough
You only need three, and they don’t advance any runners. Still, 8% for every 5% is what, 1.6 times? I guess that makes sense from the run expectancy matrix (going from bases loaded with zero outs to bases loaded with 1 out drops the runs expected by 1.43; going from first-and-third with one out to first-and-third with two outs drops it by 2.3)
wandergeist - February 7, 2010
Those drops are factors
(ie division, not subtraction)
wandergeist - February 7, 2010
Until GB rates become a 'hot stat' in baseball this arguement is pointless
while we all can see how groundball pitchers will tremendously help the mariners and their defense mindset this year the truth is- baseball is a fan of the bigger stats, like strikeouts. So many times a pitcher is highlighted by just his era and K/9. Granted a groundball pitcher with an amazing defense will have a standout ERA, people still look for the strikeout as a marker. To say that any pitcher is going to consistantly pitch low in the zone to consistantly induce groundball isn’t all so feasible in the money hungry market we call professional sports. I would like to see more factors taken into consideration in to what makes a pitcher stand out, but right now the way baseball is, K’s, ERA, WHIP, and speed control are the big ones. I cant remember the last time I heard anyone on baseball tonight refer to a pitcher and his excellence based on his ability to consistantly induce groundballs.
mariseanerhawk - February 9, 2010
What people perceive to be important is more or less irrelevant
Graham MacAree - February 9, 2010
I can't remember
the last time I heard anyone on Baseball Tonight say anything relevant to serious analysis of baseball.
pdb - February 9, 2010
People talk about good ground ballers all the time.
Wang and Webb got tons of press for that very ability. It’s under appreciated but by no means ignored and frankly I am unsure what this has to do with anything.
Matthew - February 9, 2010
Look at the Rockies
clearly teams are paying a lot of attention to this.
Bearskin Rugburn - February 9, 2010
Well thank God we have a smart GM then, huh?
seattlebruin - February 9, 2010
It's almost like Jack Z thinks about these kinds of things!
seattlebruin - February 9, 2010
"To say that any pitcher is going to consistantly pitch low in the zone to consistantly induce groundball isn’t all so feasible in the money hungry market we call professional sports."
We don’t have to look very hard to find a pitcher that contradicts that notion. Joel Pineiro purposely adapted a low in the strike zone, heavy groundball approach to the detriment of his strikeouts last year. He certainly thought it was worthwhile to forgoe strikeouts if it meant he could pitch more effectively. He also received one of the best contracts given to free agent starting pitchers this offseason.
nathaniel dawson - February 9, 2010
How about fly ball from groundball pitchers?
This isn’t really related to the article, but I’ve heard to that when groundball pitchers allow flyballs they are more likely to be homeruns because flyballs from groundball pitchers are generally on mistake pitches? If it’s not too much trouble, could you look at the average HR/FB of pitchers with at least 50% groundballs and that of pitchers with less than 35% (or something like that)?
Also, I think it’s been shown that there is a definite tradeoff between groundballs and swinging strikes.
vivaelpujols - February 6, 2010
HR/BIA is as unstable as HR/FB
That should have groundballlyness built in
Graham MacAree - February 6, 2010
GB pitchers have a higher SLG against on FBs than FB pitchers do
Or so I remember hearing. I could be wrong though.
Dan Turkenkopf - February 6, 2010
Because people are, clearly, going to bring this up.
I want to strenuously point out that this article does not say that ground balls are the end-all, be-all of pitching. There are other factors involved and some of them are directly correlated to a pitcher’s proclivity for ground balls. This post does not say that you should only evaluate a pitcher based on his ground ball rate. This post is merely trying to quantify how much more beneficial ground balls are opposed to fly balls. That’s it.
Matthew - February 6, 2010
But Johan Santana is awesome! Rabble!
OlSalty - February 6, 2010
Wait...what?
You mean strikeouts and swinging strikes are useless for evaluating pitchers?
Sidi - February 6, 2010
Gah...
I meant “aren’t useless”.
Sidi - February 6, 2010
How much of the greater run effect for fly balls is line drives?
And do ground ball pitchers allow fewer line drives than fly ball pitchers?
Dan Turkenkopf - February 6, 2010
Or is that what you meant by "excluding line drives is a meaningless distinction"?
Dan Turkenkopf - February 6, 2010
A fly ball, by itself
averages 0.17 runs and 0.74 outs which is still about 4.6 times as much offense as a ground ball.
I’m looking into the second question.
Matthew - February 6, 2010
Another question
Are the flyballs allowed by ground ball pitchers worth the same as flyballs allowed by fly ball pitchers?
The assumption being that the flyballs allowed by ground ball pitchers are better hit on average than those allowed by fly ball pitchers – and of course the uncertain distinction between flyballs and line drives can affect this.
Dan Turkenkopf - February 8, 2010
That is a fairly complex question to answer
so I looked at some possibly proxies. Namely, home runs. My theory being that if the fly balls from groundballers were indeed ‘worth more’ that would be reflected in a rise in home run rate on those fly balls.
There’s some interesting stuff here, counter to what I expected. All data from 2007-2009 MLB with pitchers who gave up at least 100 batted balls in a season.
1. Looking only at batted balls classified as fly balls (no LD or IF), as a pitcher’s ground ball rate increases does a pitcher’s home runs allowed via fly ball per fly ball increase?
No. In fact, it remains perfectly the same. A best fit equation for home run rate on fly balls alone is -.1 * GB rate + 11%. I expected this.
2. Looking only at batted balls classified as line drives, as a pitcher’s ground ball rate increases does a pitcher’s home runs allowed via line drive per line drive increase?
Also no. I was surprised by this and it might then run contrary to my supposition below that ground ball pitchers will see a higher SLG allowed on non-GB than fly ball pitchers simply because their line drive rate is expected to be higher. It turns out that while GB-pitchers might be expected to allow more line drives than FB-pitchers, the line drives themselves might be less harmful. My guess is that this would be due to some fuzziness in the distinction between hard hit ground balls and line drives and that GB pitchers see a greater share of line drives credited against them that have no chance of clearing the fence.
3. Combining the previous two questions, as a pitcher’s ground ball rate increases does a pitcher’s home runs allowed per non-ground ball increase?
No. Less surprising now given the answer to the two previous looks, but worth mentioning for completeness.
Also worth noting is that on all of these the R^2 values are incredibly low, on the nature of 0 to a max of .03 and the slopes, even the negative ones, were not dramatic. The best rule of thumb I can state from this look is that a pitcher’s ground ball rate has no impact on his various rates of yielding home runs.
Matthew - February 8, 2010
Thanks Matthew, this is some great work
I’m not sure whether HR is the best proxy for what I was asking though. It might be, or we might want to look at SLGIP, especially since we’re not seeing a difference in homers/BIA.
It seems contradictory but I seem to remember reading research somewhere that suggested no difference in HR performance, but a difference in SLGIP.
Dan Turkenkopf - February 8, 2010
If slugging is up on BIA but home runs aren't any more frequent...
That implies that the ball isn’t being hit any harder, but it’s going for more bases anyway. My guess is that the outfielders play in a little bit on extreme groundball pitchers in order to get to get to ground ball hits more quickly.
Of course, we could check if the ball is being hit harder without using HR as a proxy with hit f/x data, if we can get to it.
Graham MacAree - February 8, 2010
Alternative hypothesis: FB pitchers get more popups
Seems more reasonable
Graham MacAree - February 8, 2010
Yeah, I'm working on a way to get that info
but it’s not simple.
Matthew - February 8, 2010
I have count adjusted run values, which should actually be better than slugging for these purposes I think
Doing a straight correlation by pitcher for pitchers with at least 500 BIP from 07-09, I get this:
The equation is rv = .26gb – .05, with an R-Squared of .12.
So for every 10% increase in ground balls, you’d expect an extra .026 runs per pitch, at least if I’m interpreting that right. That’s 2.6 runs per 100 pitches, and with a not terrible R-Squared (and it’s significant at .0003).
That seems like a huge effect to me, and I wonder why it’s so different from what Matthew got.
vivaelpujols - February 8, 2010
It's different because you measured something different.
Your interpretation does not jive with your vertical axis. You are plotting against fly balls, not per pitch, so that would be 2.6 runs per 100 fly balls allowed, not per pitch.
Matthew - February 8, 2010
Aah, you are correct
So that’s about .1 extra runs on fly balls per every 10% increase in ground balls. And if you consider that the more ground balls you allow, the fewer fly balls you allow (duh), it’s not a linear increase in run value per pitch.
So it appears the effect is significant at least, but just not very big.
vivaelpujols - February 8, 2010
That should be .1 extra runs per 100 pitches
As about 4% of all pitches result in fly balls.
vivaelpujols - February 8, 2010
My goodness. This is fantastic work. You should add it to the post as an update so it appears on the front page.
Decatur - February 9, 2010
Yes, they allow fewer line drives, but I don't think that's precisely what you are interested in.
I suspect the precise question that is more interesting is: “Does a ground ball pitcher allow a higher rate of line drives per balls in the air?”
That answer is also yes. Based on 2007-9 MLB data, as a pitcher’s ground ball rate increases, the rate of line drives allowed, as a percentage of all non-ground balls, goes up. However, it goes up slowly, 0.36 points per every 1 extra point in ground ball rate. Given that the response variable is LD/BIA and not LD per Balls in Play, you’re looking at a pretty small difference.
I will not look into it specifically at this time, but this would also suggest that yes, ground ball pitchers will see a higher SLG allowed on non-GB than fly ball pitchers by dint of the slightly increased line drive rates.
Matthew - February 6, 2010
I love it when Matthew facts me.
Janic - February 6, 2010
I love that it only took him 20 minutes to do so.
Talk about technique.
Ormson - February 6, 2010
Very interesting stuff
Thanks
Dan Turkenkopf - February 7, 2010
That would make sense, when you think about it
A groundball pitcher gets hitters to hit the ball on a lower trajectory. Some popups become flies, some flies become line drives, some line drives become groundballs.
nathaniel dawson - February 7, 2010
I'm dreading the day when you guys run out of things to say.
These things are just so interesting to read.
Fear - February 6, 2010
That day will come
When the teams just determine the season by running the equivalent of DiamondMind (and generating photorealistic animations to entertain the folks watching on TV). Humans are just too quirky and inventive, and understanding the gross phenomena just leads to discovery of more subtle factors.
Also, from listening to the ever-lengthening podcasts, I think it’s clear these guys will never run out of things to say.
wandergeist - February 7, 2010
That day will come when the anti-numbers guys win out in the end ...
… and all numbers are dropped, including batting average, errors, hits, and runs scored.
Teams then are ranked by having them play in front of professional scouts, who have devoted their lives to understanding the game. Those scouts will then decide which team won the game by applying the keen observational skills.
Steve Nelson - February 7, 2010
What, no robots playing the game instead of humans?
nathaniel dawson - February 7, 2010
No, because you can't make a robot as small as Eckstein
So we have to go with VR simulations instead, Like Avatar, but with baseball bats.
wandergeist - February 7, 2010
I'm waiting for the day that baseball has been so completely described through mathematical analysis
…that the sabremetric crowd and their following that has descended upon the sport like a plague of locusts with their so called expertise and technologically assisted analysis have nothing left to wrangle over but the dry bones. Then they’ll move on to the next great thing, and leave baseball to true fans and real baseball men.
Kermit. - February 7, 2010
I'm not sure which is the better parody
Steve or Kermit
wandergeist - February 7, 2010
This was a pretty entertaining read for some reason.
Pretty interesting topic too. Answered a question I have wanted to answer for a while.
Kirk - February 6, 2010
Should take out Bunts and Line Drives to make stat more useful
Line drives screw with these stats. What happens if you take them out of the fly ball category?
In some ways, a line drive is very nearly a ground ball.
Calling them “in the air” is a little misleading as a lot of them aren’t even above eye level or land just out of the infield. I’d say line drives and ground balls also come from similar swings. (more so than flies and line drives)
Does this logic hold up if you put Line Drives with ground balls? If you take LD out completely?
AaronSawyer1 - February 6, 2010
Line Drives frequently go for HRs, grounders never do
That’s the distinction. “Line drive” is a fairly subjective determination, but they’re never misclassified as grounders, and grounders are never misclassified as line drives. The subjectivity all comes down to whether a fly ball is actually a line drive, and vice versa.
And we’ve already established that the groundball rates are a property of the pitcher, not of the hitter, so the type of swing shouldn’t matter, should it?
wandergeist - February 7, 2010
I can't imagine "never" misclassified as grounders or LD's is correct
What about balls that are struck in a straight fashion (i.e., no a pop up) that travel the first 80 feet in the air before skipping through the infield (or into the fielder’s glove)? 90 feet? 100 feet?
There has to be a point a which a batted ball could be classified as either a groundball or a line drive.
dnc - February 7, 2010
If it touches the ground in the infield it's a GB
If it doesn’t, then it’s a FB or LD, as decided by the person doing the classification. At least that’s my understanding. It either touches (or would’ve touched but for hitting a glove or player) the ground within the basepaths or it doesn’t.
wandergeist - February 7, 2010
But why group LD with FB?
It screws with the way you interpret the stats because LD’s are so weighted. If you group LD’s with GB (and a lot of batters hit GB with LD swings, while HR’s come from FB swings) it makes a lot of sense from a batting perspective instead of an trajectory perspective. I think LD should be taken out completely from the analysis.
AaronSawyer1 - February 8, 2010
Because line drives are fly balls
The only sensible division we can make of batted ball types is the one between grounders and non-grounders.
Graham MacAree - February 8, 2010
Can you please explain Chien-Ming Wang's 2009 season?
What has he done differently in 2009 compared to his previous seasons in turns of GB to flyball ratio?
brian_sun - February 7, 2010
He got hurt
seattlebruin - February 7, 2010
Have a look
At this from early in the season. Wang was releasing the ball higher than when he was having success, yet he thought the problem was just the opposite. Meanwhile, his fastball had lost velocity and horizontal movement, and his slider didn’t have as much snap either. That was the result of any one or a combination of: — injury, bad mechanics leading to injury, bad mechanics resulting from or compensating for or trying to avoid injury, and/or simply not being able to find his way back to his old form after injury. Assuming he’s healthy, a good pitching coach looking at the video and Pitch F/X data from back when he was good should be able to “fix” him, but who knows how long that will take.
wandergeist - February 7, 2010
...
http://www.beyondtheboxscore.com/2009/6/12/906896/rob-neyer-duped-by-lazy-pitchf-x
vivaelpujols - February 7, 2010
Interesting
wandergeist - February 7, 2010
It will be interesting to see how this all changes when Hit F/X is fully operational
Particularly with classifying line drives. I’m assuming there will be some sort of equation that has run value on one side and angle and speed on the other?
Fett42 - February 7, 2010
There already has been!
http://www.hardballtimes.com/main/article/using-hitf-x-to-measure-skill/
vivaelpujols - February 7, 2010
Well shit
Fett42 - February 7, 2010
Interesting stuff
OlSalty - February 7, 2010
The "Line Drive" bucket
Is subjectively determined, correct — since we’re not yet using Hit FX trajectory data, there’s a human doing the classification. And I seem to remember reading that certain parks produce more LD than others, which suggested not a park effect but an observer effect?
wandergeist - February 7, 2010
There's probably both effects at work
Matthew - February 7, 2010
Yeah pretty much
Breaking hits up into three types kinda sucks I think but its better than nothing.
Is Hit f/x actually going to track the full ball path or is it going to just measure “off the bat” values?
Edgar for Pres - February 7, 2010
As I understand it
What they have as “Hit F/X” right now (though the data is not yet generally published) is the speed and angle “off the bat” values (they can get this from the equipment they’re using for Pitch F/X). But they don’t have ball spin off the bat (see this conversation at The Book blog) , which obviously affects the ultimate trajectory significantly. Some future upgrade to “real” Hit F/X will produce the spin values, which should give them everything they need to calculate where the batted ball would go in a still atmosphere, which is arguably the thing you want to know the most when looking at hitter abilities / characteristics, independent of climate and park effects.
Obviously tracking the balls all the way would tell you more about stadium effects which is interesting too, though it is (mostly) separate from the pitchers and hitters. Of course we know some hitters try to jack balls in certain directions to take advantage of the wind, and there may be pitchers who actually consciously attempt to induce pop-ups into the prevailing wind on certain days in certain stadiums too, and there may be a skill there to capture and measure (and for pitching coaches to develop, and maybe eventually for stadium architects to consider).
I actually think the “not quite enough” warning track fly-outs are the most interesting batted balls to track, because there’s such instability in the outcomes: we know about the “just enough” HRs at HitTracker, but we don’t know about all the “almost weres” — and if just handful of one become another it changes the entire season (and perception) for a batter or pitcher, so the factors that affect those are disproportionately interesting.
wandergeist - February 7, 2010
Good comment
Yes, I think the biggest problem with Hit f/x is going to be spin – which can have a huge effect. Hit f/x is going to be better than subjective batted ball classifications, but it’s not going to be the mesiah like a lot of people think.
vivaelpujols - February 7, 2010
You can account for a little bit of spin
Incorporate hit f/x with gameday’s angle data. Not great, but better than neglecting it.
Graham MacAree - February 7, 2010
This might be what you're thinking of:
Here
nathaniel dawson - February 7, 2010
I'd just like to point out that I care more about this post, than I care about the superbowl.
Baseball is awesome
ARock - February 7, 2010
Damning with faint praise
Praise nonetheless
wandergeist - February 7, 2010
I cleaned my garage while listening to the Superbowl on the car radio ...
… except for the fourth quarter. Unlike many Superbowls, this one actually had an interesting fourth quarter.
On the whole, though, pitchers and catchers report is one hell of a lot more intriguing to me than is “Call the coin toss while it is in the air.”
Steve Nelson - February 7, 2010
Do fly ball pitchers generate more pop ups then ground ball pitchers?
Michkin - February 8, 2010
Yes
morrow - February 8, 2010
If my memory does not fail me
pop up rates are generally not stable and not considered pitcher controlled. There was a big argument about this centered around Barry Zito in his FA year because much of his value as a pitcher came from a very high pop up rate.
Actually, looking at Zito’s splits page on Fangraphs (Splits! on FanGraphs!) I can see that his career pop up rate has an enormous platoon split – 7% v L, 14% v R. This suggests some pitcher skill/control.
Bearskin Rugburn - February 8, 2010
Pitchers have some control
Graham MacAree - February 8, 2010
I was wondering at what point does the higher pop up rate for flyball pitchers offset their higher HR rate.
If that question makes sense or if that is possible to calculate.
Michkin - February 8, 2010
Looking at the tRA formula, that is.
Michkin - February 8, 2010
Also
scanning the IFFB leaderboard for last year I see a pretty good mix of pitcher types. I doubt there’s going to be any real correlation once you adjust it to pop up per ball in air.
Bearskin Rugburn - February 8, 2010
Yeah but if you don't adjust it
then FB pitchers probably have more popups.
Edgar for Pres - February 8, 2010
Additionally, from Dave's new Fangraphs post
Batted Ball League Averages 2002-2009 by wOBA
OlSalty - February 8, 2010
Ok, who got the bunt double?
Eyebrows - February 8, 2010 via mobile
That's what I want to know
OlSalty - February 8, 2010
Ok I should just read the article
http://www.baseball-reference.com/blog/archives/2099
Eyebrows - February 8, 2010 via mobile
Wow, I can't believe Furcal has 3 of them
OlSalty - February 8, 2010
That's what I said
Fett42 - February 8, 2010
Also, I would think that the OBP for grounders would be higher than the average due to errors
seattlebruin - February 8, 2010
Errors don't count for OBP.
Eyebrows - February 8, 2010 via mobile
Really?
I always thought RBOE counted against average and for OBP
seattlebruin - February 9, 2010
The flies line looks like Greg Halman's future
Fett42 - February 8, 2010
Poor Greg Halman. 814 OPS though!
Kirk - February 8, 2010
When Ichiro does the drag bunt he usually does it to get on base. What is his percentage.
He must do it a few times a year to get the run in from third base. How successful is he?
wolfmanshowlforever - February 8, 2010
His career bunt success rate is pretty ridiculous
.641/641/641 1.282 OPS .572 wOBA 276 wRC+
OlSalty - February 8, 2010
Thought this might be germane to the discussion
Tango says if you remove homers, FB = GB.
Of course GB pitchers give up fewer homers.
Dan Turkenkopf - February 8, 2010
But you can't remove homers so what's the point in doing so?
OlSalty - February 8, 2010
I need to know more about exactly what Tango is asserting to form a worthwhile argument,
but I will say that based on my looks at changes in run expectancy and score, the average ground ball does not = the average fly ball with home runs removed.
If by fly ball, Tango meant FB + LD + IF, then I might agree. If it’s straight GB vs FB, then I do not, barring further information.
Matthew - February 8, 2010
According to my run values
Which are calculated using the methodology in this article:
http://www.hardballtimes.com/main/article/searching-for-the-games-best-pitch/
I get -.07 for ground balls, -.1 for non-home run fly balls and . That’s using GameDay 2007-2009 as the source. So fly balls are better according to my numbers.
However, I see on the About tRA page, the run value of a ground ball and a fly ball are basically the same. Those values are also significantly higher than mine for all batted ball types. So I’m just confused, I could be doing something wrong.
vivaelpujols - February 8, 2010
This is a stupid question I probably already know the answer to
But does tRA assume constant run values for each outcome?
Edgar for Pres - February 9, 2010
I ask this because it seems that a "good" pitcher should have lower run values for BIP
Since its more likely that there aren’t any baserunners on. Does this make sense?
Edgar for Pres - February 9, 2010
Yes, constant.
In a perfect model, it would probably all be non-linear, but keep in mind that fewer base runners on, while meaning that the run value of, say, a ground ball would be lower, it also means the out value of a ground ball would be lower since there would be a lesser chance of a double play.
Matthew - February 9, 2010
Cool. Yeah its probably a pretty small effect
I could see it leading to undervaluing great pitchers and overvaluing horrible pitchers. Its probably small though since all pitcher really fall within a fairly small window of performance.
Groundballs might be complicated but flyballs and line drives only get less valuable with decreasing number of baserunners.
Edgar for Pres - February 9, 2010
Not that simple.
Remember to consider that as the number of base runners increases, so too does the cost of an out.
Matthew - February 9, 2010
Which means strikeouts are worth more when you have less of them, among other things.
Graham MacAree - February 9, 2010
Anyway, I think the non-linearity of it all would be pretty cool to flesh out
It would be pretty hard though and I’m not sure what the best way to go about it would be. Fitting the data to come up with a non-linear expression would be a little less gratifying than coming up with something that is based on first principles that you can easily relate back to reality would be awesome.
Edgar for Pres - February 9, 2010
Maybe off the top of my head?
Baseruns but use batted ball stats as inputs to come up with the probability of hitting 1B, 2B, etc. Does that make any sense?
Edgar for Pres - February 9, 2010
Yes. I've actually been looking this for months.
Graham MacAree - February 9, 2010
Is that an intelligent way to go about building a stat?
Are there issues with it that I’m not seeing?
Edgar for Pres - February 9, 2010
It will work.
Graham MacAree - February 9, 2010
Is there a reason why somebody hasn't done it before?
Just alot of work?
Edgar for Pres - February 9, 2010
Well the main reason I haven't done it yet is because I want to incorporate it with something else first
And also I don’t want to do it for free.
Graham MacAree - February 9, 2010
Leverage Index?
OlSalty - February 9, 2010
Good on you
goodness knows you guys should be better compensated for how good you are at this
seattlebruin - February 10, 2010
Yeah true
I agree it gets super complicated
Edgar for Pres - February 9, 2010
Yeah there is some decreasing value associated with strikeouts which I think is kind of cool
Edgar for Pres - February 9, 2010
This conversation is germane to this sub-topic:
From the Book Blog (start here and scroll down).
Incidentally, Graham/Matthew: Any thoughts on ‘SIERA,’ the new BP pitching metric?
marc w - February 9, 2010
My thought:
They’re doing it the wrong way. Starting from a nonlinear regression makes way less sense than starting from what actually happens in a baseball game.
Graham MacAree - February 9, 2010
Yeah
You might end up at the same result but you won’t understand how you got there or what is going on.
Edgar for Pres - February 9, 2010
Additionally:
Fuck you for that little insinuation, BP.
Graham MacAree - February 9, 2010
What little insinuation?
Forgive me, for my statheadedness isn’t as well-developed as it should be. I took a couple years off baseball when the M’s really sucked. (In my defense, I was in high school and I was a highly impressionable person.
thehemogoblin - February 9, 2010
So they did better than what, xFIP?
jolly good for them.
Bearskin Rugburn - February 9, 2010
You must Login with your SB Nation account and be a member of Lookout Landing to post a comment.