Voros McCracken wrote on Baseball Prospectus over eight years ago to release one of the most controversial findings in the history of sabermetric thought: he said that pitchers only control whether a hitter strikes out, walks, hits a homerun, or hits the ball in play, but have little to no control over whether a ball in play is a hit or an out. Many have argued that the reason this is true is because walks, strikeouts, and homeruns are outcomes that have nothing to do with defense. Anything else has the potential to land on the field or be caught beforehand, and so it is the defense’s job to cover ground effectively. The pitcher had to keep the ball from being hit out of the park, or from being hit altogether.
In that article, he states that pitchers control three main things: strikeouts, walks, and homeruns, and that pitchers who were above average in recording strikeouts, avoiding walks, and avoiding homeruns one year were frequently above average in those categories the following year. However, the pitchers who were above average at preventing high Batting Average on Balls in Play (BABIP) one year were no more likely to be above average at preventing high BABIP the following year than pitchers who were below average.
BABIP is defined as follows:
McCracken argued that the hitters the pitcher faced and the defense behind him were the key determinants in a pitcher’s BABIP. Comparatively, a batter’s BABIP not only shows a much higher year to year correlation, but you can also learn a lot through breaking it down by batted ball type.
Correlation coefficients are a measure of how well two variables move together. These year-to-year correlations test whether pitchers’ strikeout rates, for example, tend to be high if they were high the previous year (a positive correlation), low if they were high the previous year (a negative correlation), or no more likely to be high or low if they were high the previous year (zero correlation). Correlation coefficients take on values between -1 and 1, where a correlation of 1 means that the two move together in perfect unison, 0 means that they are unrelated, and -1 means that the two always move in opposite directions. Look at the following table stating the year-to-year correlations for each of the following statistics using 2003-2008 data on the pitchers who threw at least 100 innings in consecutive seasons.
PITCHERS' STATISTIC CORRELATION WITH SAME STATISTIC
THE FOLLOWING YEAR
Strikeouts per Batter Faced .7686
Unintentional walks per Batter Faced .6682
Homeruns per Batter Faced .3769
BABIP .2242
Pitcher's BABIP minus overall team BABIP .1490
As you can see, pitchers do seem to have some persistence in their BABIP, but a large portion of that correlation is actually the defense behind him that keeps the whole team’s BABIP low or high (as shown by the fact that the BABIP correlation is smaller when you only consider the difference between the pitcher’s BABIP and that of all pitchers on his team). Hitters seem to control their BABIP a lot more. Check out the same table for hitters who had 300 plate appearances or more in consecutive seasons from 2003-2008.
HITTERS' STATISTIC CORRELATION WITH SAME STATISTIC
THE FOLLOWING YEAR
Strikeouts per Plate Appearance .8467
Unintentional walks per Plate Appearance .7751
Homeruns per At-Bat .7420
BABIP .3657
The following two diagrams illustrate this point pretty well. The first graph shows pitcher’s strikeout rates one year versus the next year. As you can see, the pitchers who had low strikeout rates one year had low strikeout rates the next year.
This graph shows the difference between a pitcher’s BABIP and his team’s overall BABIP allowed in one year compared to the same difference the next year. It does not seem like being vulnerable to hits on balls in play one year makes you any more or less likely to be vulnerable to the same thing the next year.
If you try to explain this concept to the average baseball fan, you’ll get some funny looks. After all, it goes against all common understanding of the game. How could a pitcher not really affect his BABIP? Pitchers who have good stuff must be able to induce weak contact, right? It goes against logic, and while we can look at the numbers and see it is true, you are likely to get a response like “oh, you can manipulate numbers to tell you anything.” When you go to a ballgame and you see the visiting team blast a line drive into the gap, you blame the pitcher, don’t you? This is where the problem lies. If he left a meatball over the plate, it seems like it is probably his fault.
McCracken himself has since softened his stance and most sabermetricians in the know generally believe that pitchers have some control over BABIP, but very little. They certainly do not seem to have any special ability to avoid line drives-the correlation in my sample between line drive rates one year and the next is 0.00! In fact, old BABIP data is nearly irrelevant in predicting their performance because it is correlated with the other factors, like strikeout and homerun rates, which matter more.
Groundball rate is somewhat correlated with BABIP (positively) only because flyballs that are not homeruns are more likely to be turned into outs than groundballs. The correlation is not all that high (.11), and is barely far enough from zero to assume there is a real correlation. There is also a slight negative correlation (-.18) between a pitcher’s strikeout rate and his BABIP the following year, too. In reality, the question is why BABIP has such a poor correlation for pitchers.
To project future hitter performance, not only is it useful to know his homerun, walk, and strikeout rates, but it is important to know his BABIP. However, to project future pitcher performance, it is still useful to look at his homerun, walk, and strikeout rates, but knowing his BABIP will not help the projection at all. When projection systems like PECOTA make ERA projections that turn out to be accurate, even though they do not seem in line with the previous year’s ERA, they are using a trick-they are mostly using strikeouts, walks, and homeruns to predict ERA, and are not using last year’s ERA much at all! Of course, the pitcher does seem to have a little tendency to control BABIP, but that effect is captured by looking at his strikeouts and groundball rate. If hitters had a high BABIP against him the year before, that might be an accident. It’s best to look at how many people he struck out and that will tell you whether he can induce weak contact better than how much weak contact he caused last year.
There is clearly a dilemma here. On one hand, we see a meatball lined into the gap, and we know the pitcher probably messed up. On the other hand, we keep looking at all these numbers that tell us pitchers who give up a lot of line drives are no more likely than other pitchers to give up a lot of line drives the next year. What gives?
The concept is no more complicated than rock/paper/scissors. If you have ever played rock/paper/scissors before, you know that you can’t keep doing the same thing over and over, and you can’t be any more likely to do rock, paper, or scissors at any given time. You need to randomize. In game theory, we call this approach a mixed strategy.
When it comes to pitching, we know that if a pitcher threw a fastball on the outside corner every single pitch all year long, the hitters would eventually discover this and start hitting the ball the other way. The result would be a lot of opposite field hits and a high BABIP. Pitchers who are predictable get told by their coaches to “keep the hitter honest.” What they are saying in game theory terms is “play a mixed strategy.” Even if you are a little predictable one day, you can switch your approach. The result is that sometimes the pitcher throws an 0-2 fastball right down the middle and the hitter is caught looking. Pitchers are taught to randomize their pitches and locations, and as a result, a line drive often comes from a hitter guessing right. Chipper Jones recently said, “For me, plate discipline is being able to know what pitch you want to put in play before you step in the box and not swinging at anything else but that.” That is certainly one way to guess right, and that’s why Chipper Jones’ BABIP has been .343, .352, and .388 compared to league averages of around .300 in the last three years. For pitchers, the best way to have a low BABIP is apparently just to face Chipper Jones less. It’s not that pitchers don’t control BABIP-it’s that pitchers barely differ in their abilities to control it, because the only control they have is to try and stay unpredictable.
Thank you for reading
This is a free article. If you enjoyed it, consider subscribing to Baseball Prospectus. Subscriptions support ongoing public baseball research and analysis in an increasingly proprietary environment.
Subscribe now
The middle just felt weird to me. My understanding is that while BABIP is not useful for predicting a pitcher's future success per se, it is useful to show how much of a pitcher's performance was affected by luck. So thus, looking at BABIP is useful to see if a pitcher's performance is repeatable.
I also thought the paragraph starting with "If you try to explain this concept to the average baseball fan..." was basically the same thought rerepeated 10 times.
But Richard is correct that removing BABIP luck is certainly a good way to improve ERA projection.
That being said, I thought thing picked up considerably after that, and agree with others that the graphs were particularly enlightening, especially to the "novice" reader.
Overall, I think Matt does a fine job of explaining a rather difficult concept for the average fan.
The author says, "It's not that pitchers don't control BABIP—it's that pitchers barely differ in their abilities to control it..." This is completely wrong. Simply because there's a weak correlation between two numbers does not mean that pitchers have little control. The measure in question could just have a lot of noise so the correlations naturally weaken. There are tons of possible reasons for this noise and a throwing in a game theory explanation isn't really necessary or helpful.
Anyway, I sincerely liked the first part of this article. Stopping before the last paragraph, I would consider it the top entry.
Thus, for most pitchers, it becomes a rock-paper-scissors game of avoiding pitching patterns while also avoiding the chance that the hitter makes a lucky guess on what the pitcher will throw. The better pitchers are one who can vary their pattern the most, or in the case of those like knuckleballers, don't know what their pitch will end up doing in the first place.
As far as my mixed strategy theory, I don't really find it so far fetched:
(1) Line drive BABIP is about .730, flyball BABIP is about .140, and groundball BABIP is about .240, and line drives makes up 20% of balls in play-- meaning that line drives are where a large portion of hits on balls in play come from.
(2) Line drive rate has a correlation of actually 0.00 among my sample of 494 pitchers who have thrown 100 innings or more in consecutive seasons from 2003-2008 (retrosheet changed its definition of line drives starting in 2003), meaning that in practice, pitchers really don't seem to be able to affect line drive rate by even the slightest amount at all.
(3) You need to objectively ask yourself what the line drive rate of a pitcher with average skills would be if threw the same pitch in the same location every time. Imagine a fastball on the outside corner. If every hitter knew that was coming, would they still only hit line drives 20% of the time? I can't imagine that anybody believes this. If not, then given (1) and (2), it does seem that preventing line drives is done by playing a mixed strategy so the hitter can't predict what's coming. Hence, the small correlation in BABIP is due to similar abilities in preventing those .730 BABIP line drives.
I'm sure that mixed strategies are important, but your conclusion doesn't really make sense. If a pitcher threw the same pitch in the same location, the correlation for strikeouts and homeruns would also be pretty high. And they are. The correlation for BABIP would probably be smaller because there would be lots of noise. This is exactly what we see.
So, the evidence doesn't really make your point. I do assume that all pitchers use a mixed strategy (though an article about this using pitch type/location data would be really interesting). But your argument seems to be: pitchers want to prevent line drives, there's no correlation in line drive rates for consecutive seasons -> pitchers used a mixed strategy? Not sure how we reach that conclusion. Are you saying all pitchers used a mixed strategy and mixed strategies lead to randomness? If so, why the correlation in homerun rates (which pitchers are also definitely trying to prevent)?
I'm just missing the link. Are all pitchers using a mixed strategy? Equally well? And then mixed strategies lead to randomness? Or they just lead to better outcomes in general?
Could you explain what you mean about the denominator being smaller for BABIP? It seems like I'm using the same number of pitchers and there are actually more balls in play than anything else. Do you mean that K/PA has a bigger denominator than Hits/(Balls in Play)? I think that will have some effect but not drastically, but if you mean something else, please let me know.
Of course doing a strategy of throwing the same pitch in the same place repeatedly would also make a pitcher to strike out fewer and give up more homeruns. But when you see a hitter swing and miss at a 100mph fastball, that's pretty clearly something that only certain pitchers can do. When you see a hitter line a single the opposite way on an outside fastball, it's usually about the hitter guessing right or the pitcher making a bad/predictable decision that he learns from and adjusts next time around.
I guess I would say that mixed strategies clearly improve performance, and since everyone does them, the true difference in pitchers seem to be an ability to avoid contact, an ability to find the strike zone, and an ability to cause pitches to arrive at different trajectories and therefore be hit on the ground or in the air.
So by varying location and pitches, that in effect applies game theory to tilt the odds more in the pitcher's favor because a batter has to react not only to the pitcher's stuff, but the batter is also uncertain of what kind of pitch will be received and what the location of that pitch would be.
Not sure if you're a poker whiz or not, but I keep thinking of Sklansky's Theory of Poker when considering this concept. The idea is that if the batter knew what pitch was coming and where it was thrown, he would perform optimally by picking the correct swing timing, etc. The pitcher's job, then, is to remove the batter's certainty through varying pitch type and location so that the batter handles each pitch in a suboptimal fashion.
Denominator: No, you interpreted correctly (PA vs. BIP). Not the main influence, but it probably has some effect.
Mixed strategies: I just don't see how mixed strategies explain the low correlation in BABIP. A few posts below you talk about what would happen if Eck had always thrown the same pitch on 3-2 counts. Yeah, he'd probably be a bad pitcher. But what does this have to do with BABIP?
Are you trying to explain the low line drive rate overall or the low correlation in BABIP? If all pitchers are using mixed strategies equally well, then it's probably not the reason for the low correlation.
Here's an analogy: What would happen if Eck could only throw 55 MPH? His line drive rate would probably skyrocket. Since his line drive rate was low, I can assume he threw more than 55 MPH. I guess I can assume every pitcher throws more than 55 MPH because of the low line drive rate in the majors. But that doesn't mean I can say anything about it explaining BABIP correlations. I get that mixed strategies are good for pitchers' performance stats. They have to be. But why does this affect BABIP correlation? There's a major link missing here in your argument.
The reason that mixed strategies have to do with BABIP is that the typical counterargument to the statement that pitchers don't really control BABIP is "if that were true than there is nothing a pitcher could do to have a high BABIP, but we know that's not true." I am explaining away that hole in the argument so that it's clearer to people. It's true that if Eck threw 55 MPH or the average high school pitcher went out there and did the same, BABIPs would be high, but among major league pitchers, there aren't pitchers like that. I've read that minor league pitchers with very high BABIPs are likely to fail, since they are too far from major league level but the skill distribution at the top in terms of BABIP is very small. Not playing mixed strategies would make them terrible, but when they are playing smart mixed strategies, the only real difference is missing bats (leading to Ks), missing the strike zone (leading to BBs), or throwing pitches that have a trajectory such that they will be hit in the air more (leading to flyballs, and hence, HRs). Line drives really do seem to be a product of good guesses and quick hands of the hitter.
I'm trying to explain the low line drive rate actually. The reason I do that is because we know that pitchers differ in their groundball/flyball ratio and that any team is bound to have a higher BABIP on groundballs than flyballs. Picking through the data, I found that the reason that the correlation was so small anyway was the inability of pitchers to prevent line drives was the same across players, and that line drives make up a very large share of hits on balls in play anyway. That's why the correlation is low-- the key component has no year to year correlation.
1) I don't think it's good practice to use the predicted BABIP to get a standard deviation measure of true ability. For any regression, you have explain and unexplained variance. You've only used the explained variance but since you have a poor (noisy) measure of true ability, this doesn't really give you the right number (eg. regress wages on IQ and get predicted wages -> you wouldn't claim the std dev for the predicted wages is the correct std dev for wages since there's so much unexplained ability). Anyway, this actually isn't that big of a deal to me.
2) At the risk of coming off snarky, your theory is completely wrong (or at least illogical). You're just saying mixed strategies lead to better performance. So does throwing harder than 55MPH. Your point seems to be - all pitchers use mixed strategies (equally well?) and mixed strategies lead to randomness. By your logic, we should see no correlation in HR rates as well. It would be incredibly interesting (though, I'm guessing, difficult) if you did a study on pitch types/location to study patterns, looking at whether predictability affects BABIP. But you just can't prove it with the stats you've presented.
The explanation for this theory really has to center around why HR rates are highly-correlated across years, but BABIP is not. The general answer is noise. Take a HR ball and move it over 1 foot - probably still a HR. Now take a grounder and move it over one foot - the probability that the outcome changes (out->hit) is higher. Or don't move it over at all and the probability of an outcome change is still non-zero (fielder is a step slow in one scenario). This "extra randomness" just means more noise. Noisy measures are less-correlated.
Anyway, I really do appreciate the responses and I did vote for you (one of 2).
1) The point about using a regression to find SD is a fair criticism. I guess that only works if you assume I have found the true function using the regression, which only is a good argument if you agreed with me in the first place. Tell me if this one helps (I have a feeling you will have no problem following my math but tell me if I'm leaving something out, and I apologize to any other reader who I'm skipping steps with)...
a) netbabip has a standard deviation of about .0202.
b) the average pitcher in my sample has about 560 balls in play and a .297 BABIP meaning that if there were no skill in BABIP at all, you would get a standard deviation of around sqrt(.297*.703/560)=.0193 (using the binomial distribution).
c) The variance in ability + the variance due to the binomial distribution should add up to the total variance in the sample. meaning that the true standard deviation, SD, should be such that SD^2 + .0193^2 = .0202^2. That gets me SD=.006 which is close to the .005 I got using the regression.
...Does that make sense? I recognize your disagreement on the log wage vs IQ regression example, so I thought this might be another way.
2) I still think my theory is right, but what you say is also true about the one foot difference in HR vs GB. There are a few things that are true...one is that mixed strategies will improve pitching performance, and that everyone using mixed strategies will lead to a smaller distribution in performance than some people mixing and some being predictable. I think we already agreed on that point. The other is that the more variance in the skill level involved in a certain statistic among MLB pitchers, the more variance in performance.
The variance in HR/9 comes primarily from variance in flyball rate. There is not much variation to HR/Flyball. So the variance in angles that a pitcher's pitches reach the batter is high enough that flyball rate is highly varied even though all pitchers are mixing strategies (though not as much as if pitchers additionally varied between smart mixed strategies and dumb pure strategies).
The variance among major league pitchers in throwing pitches that the batter can square up and hit a line drive against when they guess right (and the variance among major league pitchers in throwing pitches that the batter can square up and hit a line drive against when they guess wrong) are both not so highly varied so the primary variance in line drive rate comes from randomness in how often hitters guess right.
In general I'm saying that mixed strategies help with GB/FB ratio, K rate, BB rate AND LD rate, BUT there is still a large variance in GB/FB, K, and BB skills among major league pitchers, and LD rate does not have much variance in skill level among major league pitchers-- as long as they play a mixed strategy...which they do. The key is that LD rate would have persistence if some pitchers mixed strategies and others didn't.
First, that on a graph (similar to the BABIP graph) would have added.
Second, the rock-paper-scissors part might have been more compelling if it started with the premise that the goal is to prevent line drives.
Finally, further analysis could look at whether some group of pitchers *do* seem to have a year to year positive correlation in line drive rate to see what we could learn (Johan Santana or Roy Halladay, perhaps?).
You make a good point about linking the rock/paper/scissors to the goal of line drive prevention earlier.
It's tough to know about individual pitchers because even if you looked at a statistic that is totally random, there will be consistent outliers who are very unlucky a few times in a row or very lucky a few times in a row. It's easier to look at groups of pitchers. Knuckleballers have a lower BABIP than other pitchers, according to what I've read (though I admit I have not looked into this myself). High strikeout guys tend to also-- so that would explain some of Johan Santana. But that effect is small. The BABIP effect of strikeout rate might explain only a few points of BABIP.
Thanks for the comments.
The analysis was very strong and did a good job of letting the reader know why the stat is important. Gets my vote.
First, I think he unnecessarily assumed an excluded middle in his opening paragraph, which is a shame, because it was really just flavour-text.
But more than that, he titled the article very badly. This article does an excellent job of showing that pitchers don't control BABIP, and does so by wonderfully illustrating (with great graphs) that year-to-year BABIP is effectively random.
But that's not what the article said it was going to do. The title reads "Why Pitchers Don't Control Batting Average on Balls in Play", and that question not only wasn't answered, but it wasn't even approached.
Demonstrating that something is true (which this article does very well) does not constitute an explanation as to why it is true.
This is borderline for me.
I voted for 4 of the articles (to give you an idea of how permissive I was).
I felt that I explained the reason why pitchers have little to no control over BABIP had to do with playing a mixed strategy. Earlier in the article, I isolated that the true non-correlating variable (in fact, 0.00 correlation) was the line drive rate. I explained how mixed strategies are a common choice among pitchers, and why BABIP would vary wildly between pitchers if pitchers made the same predictable pitches. I read on Wikipedia (http://en.wikipedia.org/wiki/1988_World_Series) that Kirk Gibson hit his homerun off Dennis Eckersley in Game 1 of the 1988 World Series precisely because Dodgers' scout Mel Didier had said that Eckersley always threw a backdoor slider to lefthanded power hitters on 3-2 counts. That is not a mixed strategy. What happened? He hit it out of the park! Now, that's not a line drive, but if Eckersley didn't start mixing his strategy in 1989, don't you think that lefthanded power hitters would have started hitting a lot of doubles in the gap on 3-2 backdoor sliders? Sure, but pitchers adjust, and I'm sure Eckersley learned his lesson. The result? No correlation in BABIP from one year to the next.
Not to mention every player in MLB saw what pitch Gibson hit out of the park. It would be like a mini training video on how to hit Eck.
I would have liked to see more about line drive/flyball/groundball percentages, and how they impact BABIP. Although, possibly that would have led away from the "Basics" format.
I thumb this with no hesitation.
Heck, run them with OBP as well.. maybe the more patient hitters are more likely to pick good pitches to swing at, and thus, get more line drives.
If players have a higher OBP because pitchers are afraid of them, how does that explain the Alfonso Sorianos of the world? Is it that he has power, but pitchers know how to pitch to him?
There is still a lot of variance in patience among power hitters. Alfonso Soriano swings at 35-40% of pitches out of the strike zone versus a league average of under 25%. He only walks at all because pitchers throw him enough pitches far enough out of the strike zone that he lays off them now and then.
Regarding Soriano, are you then saying that his power comes not so much from being patient, but by capitalizing on additional oppotrunities outside the strike zone?
I don't think power comes from patience. I think that you need to have some mixture of power and patience to make the majors, and the correlation in walk rate and homerun rate is due to pitchers throwing fewer strikes. There is no correlation between power and laying off bad pitches, but there is a correlation between power and swinging at good pitches. Soriano is powerful in general and may be more powerful if he had a better eye. It's tough to separate having a good eye from being patient. I'm pretty sure Soriano doesn't really have either quality (eye/patience) but has enough power to keep hitting homeruns anyway. A hitter can make the majors with a poor eye only if he has enough power or contact skill to make up for it. Juan Pierre with Soriano's eye and patience would never have made the majors with Juan Pierre's power deficiencies and Soriano's eye.
When you mention a mixture of power and patience, I'm almost wondering if it's a supply/demand problem... the supply being the kinds of pitches a pitcher throws versus the demand of the kinds of pitches a hitter will swing at. Probably outside the scope of the discussion of this article, but might be worth some additional thought...
Now if only I knew an economist... :)
since you've made your case in exchange for a vote to other readers in these comments, i'd love to hear some thoughts or analysis on this in turn for my vote..
At the same time, it's better to include them because they are a skill. Hitters especially have a lot of control over their infield fly rate, and infield fly rate is naturally correlated with foul fly rate. So it makes sense to include avoiding or recording foul outs as a skill, albeit while understanding that park effects play a large role.
Good information as well. I was hoping that there would be a big string at the end to draw everything together. If a pitcher has to try to be more random, are some pitchers better at it? Why some years are they more random than others? I think the key component to that is the catcher. I think adding the catcher in as a variable is almost as important as HRs, BBs or Ks.
This is easily one of the top 3 articles posted in this round.
...and yet you didn't really address quality of opposition faced, or how much it might affect year-to-year correlation of BABIP.
Pretty good overall. I'll agree with the people who have called the first sentence "a disaster", and also those who found it disingenuous to present your discussion of randomized strategies as an *explanation* of the low year-to-year correlation in BABIP. Hint: why doesn't the same explanation lead to low HR/9 correlation?
I discussed the HR/9 thing a little in other comments, but the real reason is fascinating. The reason that pitchers have higher correlation with respect to HR/9 is that those who give up a lot of homeruns are flyball pitchers. The rate of homeruns/flyball has very little correlation year to year. There certainly is a high correlation in flyball rate, since pitchers throw different pitches that come in at different angles. That's where most of the variance in HR/9 comes in. I always wondered that about BABIP too, and the HR/Flyball issue resolved it for me a good deal.
1. The other pitchers on a team do NOT face the same hitters, especially under the anti-balanced schedule with interleague play. The differences are significant, and BP reports them in a canned report on the Statistics page. This deserved at least an aside.
2. The problem is not whether it is true that HR/9 is pitcher independent and BABIP is not. The problem is that the argument/explanation you gave for BABIP would apply equally to HR/9, and so can't be right -- or at least not sufficient. That's a flaw of logic and exposition, not of fact.
2) My goal is to explain the lack of correlation in line drives. Flyball rate clearly has a higher correlation, and this is why HR/9 has a higher correlation. HR/Flyball (adjusting for park) does not have a high correlation at all. The theory I'm using works fine if you assume that homeruns are not "correct guesses" much more than other flyballs are. Line drives tend to be a product of good guesses. Homeruns are flyballs hit by more powerful hitters.
For all of you statisticians out there, there is a whole wide world of pitching formulae, ratios and analytical tools for you to develop.
Bogus
I understand that the conventional wisdom says that pitchers can make hitters pop it up or avoid hitting line drives but the fact is that were true, the same pitchers who succeeded in one respect would be more likely to the following year. With line drive rate, it's simply not true. Groundball pitchers actually have higher BABIPs since flyballs have lower BABIP than groundballs.
Btw, the statement "Groundball pitchers actually have higher BABIPs since flyballs have lower BABIP than groundballs." is worded awkwardly at best and is iffy kind of logic at worst... you are basically submitting a proof by surmising its opposite is true but not providing proof of that opposite. I realize that there are stats that flyball pitchers get more popouts, so have a lower BABIP, while groundball pitchers are more likely to give up hits because balls can sneak through fielders, but the statement as you word it does not say that..
http://www.baseball-reference.com/leagues/split.cgi?t=b&year=2008&lg=MLB
GB-BABIP: .236
FB-BABIP: .142
LD-BABIP: .718
Then they separate by groundball pitchers, flyball pitchers, and neutral pitchers
Groundballers' BABIP: .305
Flyballers' BABIP: .290
Neutral guys' BABIP: .302
I would also point out that if you're really talking about "missing bats" you should exclude called third strikes. After all, many good strikeout hitters get their strikeouts by hitting corners. And if the hitter reaches for these balls, instead of taking it, he's likely to hit it weakly; purpose achieved either way.
That said, over two rounds of submission Matt has proven himself one of the best at structuring his articles, builing a sound organization around his thesis, executing with solid prose, and connecting the numbers to real-world examples. I especially found his original piece creative and engaging, and I hope future rounds give him greater opportunity to display that potential.