Steve McCatty On Strikeouts

[enuttage]

Not that anyone really cares about what I think. Especially since I've been gone for so long and am so far behind on all the latest statistical analysis.

But FWIW, I can see both sides of this argument.

There are emotional appeals on both sides, much like the Galileo quip earlier.

And as a brief aside, everyone should read the full Galileo story again when you have the chance, especially in the context of this conversation. Because Galileo's theory was right, but his data analysis and proof to back it up were entirely wrong. So you can come to a right conclusion with wrong data for what its worth. Galileo was also a giant dick that had enough hubris to fill the entire Mediterranean Sea, which was the real reason, along with some political background goings on, that got him his house arrest. Not his actual theory. But all that for another time. Would be a great movie.

Anyway.

I see that Curve Ball here is trying to say that its about more than the statistics, and that most everyone else is saying that they don't lie. I see that there is probably some common ground here somewhere, but I'm too lazy to try and bridge it.

My only analogy comes from the racing world, where there is a more accepted synergy between engineers and data crunchers and drivers than there (still) is in baseball, it seems.

But drivers are often hesitant to adhere to normative boundaries placed on them by engineers in test conditions that are often somewhat vacuous compared to actual race conditions. Sometimes, this proves foolish, or even deadly, to the drivers. Their crews know a great deal about the car and all of the numbers that it produces when it runs. MOST of the time, the engineers are right.

Of course, you can't quantify everything that a driver does in a race, no matter how hard you try. The reading you get from a car become less meaningful when other humans are involved. There is some predictive behavior in certain situations, but in the end, the driver sometimes has to make decisions that clearly mean a slower lap than ideal, but end up winning them the race because of their ability to predict the behavior of other drivers/conditions/etc.

And sometimes conventional wisdom based on years of data collection proves futile when conditions and situations arise that fall outside the boundaries of what is collected. Those conditions arise EVERY race. Sometimes they are critical to the outcome, and sometimes they aren't.

My point is that perhaps in an emotional moment with the press, McCatty was driving a point home and was somewhat hyperbolic in his assessment. And perhaps, there are still intangibles in baseball that are better left to the situational awareness of the player in question (e.g. TRYING to let the ball get into play). At risk of becoming a proponent of the God of the Gaps, which I am not, I'll say that some of the fun of baseball is not having all of that data at our fingertips just yet. Perhaps one day certain statistical analysis programs will be banned from sport much like performance-enhancing drugs are now.

Anyway...rant over. Point on my part is that perhaps there is some common ground here y'all can find if possible. And that everyone should go read about the Galileo Affair and how much of an a-hole he was.

The only

[Kobie]

Originally Posted by Curve Ball Dave

Screw the sum total of over a hundred years of experience and knowledge compiled by all of the players who took the pitcher's mound at the start of a game with the goal of giving their team a chance to win. Math whizzes who never put on a pair of spikes know more about pitching than we do.

It's about math, man, not baseball.

Let's replace pitchers, as their they are only as good as the number of strikeouts they can compile and hitters they can avoid walking, and replace them with unskilled pitching machines that can fling the ball to home plate and strike out 7 batters per 9 and won't walk anyone. Pitching machines won't get tired or complain when they're pulled from the game.

Because that's clearly the argument here. Bad strawman is bad.

[Curve Ball Dave]

Originally Posted by enuttage

I see that Curve Ball here is trying to say that its about more than the statistics, and that most everyone else is saying that they don't lie. I see that there is probably some common ground here somewhere, but I'm too lazy to try and bridge it.

That's exactly what I am trying to say, and contrary to what some may think, I am not absolutely rejecting stats as a tool. I pitched a long time, I also got paid once to do statistical analysis. Where you lose me is when you say that because there is no statistical metric at this point in time for something, it must therefore not exist. No trained researcher would make such a claim.

My point is that perhaps in an emotional moment with the press, McCatty was driving a point home and was somewhat hyperbolic in his assessment. And perhaps, there are still intangibles in baseball that are better left to the situational awareness of the player in question (e.g. TRYING to let the ball get into play).

Thank you. You're the first poster to realize that the game is played by humans, not statistical generating robots. It is that situational awareness that allows smart pitchers to allow hits when they know they don't have to bear down, walk hitters without making it obvious because the guy they really want to face is up next, or not care if they give up a solo home run because their team is still up 8 runs. All of these things reflect negatively in his empirical statistics in the spreadsheet, but none had any bearing on the outcome of the game. It is also that awareness that tells an experienced pitcher that the hitter is expecting a certain pitch, just by looking at how he's holding the bat or how his feet are set. You cannot quantify this, it's skill and it exists.

I can't state how many times in my career when I had a meeting with my catcher and said, "I don't care if this guy gets on". I knew the situation and my ability. Giving up a hit to that one batter because I was conserving my energy for the long haul reflected negatively in my stats. It was also irrelevant to the outcome of the game. Hits per 9 tells you some things, but not everything...to give just one baseball example.

As I write this I am watching an MLB game. I'm watching pitchers fool some hitters and induce easily fielded balls, and not fool others and give up solid hit balls. Yet some in this thread are arguing that what I'm watching with my own two eyes is not happening because there is no stat for how many hitters a pitcher fooled in a game. The data may show the pitcher had just about no bearing on the outcome of the at bat that resulted in a ball in play, but geez just watch a hitter bailing on a breaking ball and hitting a weak ground ball to third base and argue the pitcher's skill had nothing to do with that outcome because of a study someone had done.

[Curve Ball Dave]

Originally Posted by Kobie

Because that's clearly the argument here. Bad strawman is bad.

Then what is the argument???

I am stating that a pitcher, by varying the speed and location of his pitches in such a pattern that will maximize the probability of a hitter mistiming and incorrectly targeting his swing, has control over how well a ball will be put in play. The less squarely the ball is hit, the greater the probability of an out. The ability of a pitcher to do this is called skill and some have more skill than others. Pitchers have been taught for 140 years to vary the speed and location of their pitches to induce weakly hit balls.

The counter as far as I can tell is that someone who to the best of anyone's knowledge never played the game did a statistical study that "proved" that there is no such thing as what I just wrote because the data just doesn't bear it out. What I've stated cannot be measured, therefore it does not exist. The pitcher's ability, if we can call it that, has no bearing on the outcome of a hit ball because that's what the numbers say. Period, there is no other argument.

Correct me if I have the counter argument incorrect.

[rudel.dietrich]

I have never had cancer, but I believe in cancer research.

[Curve Ball Dave]

Originally Posted by rudel.dietrich

I have never had cancer, but I believe in cancer research.

I believe in research too. I also accept that there limits on what can be properly quantified. Simply because something cannot be properly quantified for analysis does not mean it does not exist.

I witnessed, in person, a hitter face Chris Sale of the White Sox. Sale threw a change up. The hitter, who was gearing his timing for a fastball, began his swing far too early and barely tapped the ball. The result was 1-3 put out.

Those who point to the studies and data want to argue that Sale's ability to throw off the hitter's timing on that pitch had no bearing on the outcome of the that at bat because the ball was put in play and the pitcher has no control over the ball once that happens.

Unless I have the argument wrong. I've yet to see it in baseball terms. I'm just told to do the math or read the study.

[Curve Ball Dave]

For those who want to site McCracken, consider what he wrote here

The critical thing to understand is that major-league pitchers don't appear to have the ability to prevent hits on balls in play. There are many possible reasons why this is the case, and I don't really have a concrete idea as to why it is.

emphasis mine

Because Voros you are looking strictly at numbers, not the actual playing of the game of baseball. You are fixated on the statistical data, not the actual confrontation between the pitcher and batter. You can't qualify things that cannot be quantified. And simply because you don't see it numbers, it does not mean it is not there. They may not appear to have that ability because the numbers don't show it, but it's there. Pitch a few innings Voros, and you'll see.

[enuttage]

Originally Posted by Curve Ball Dave

For those who want to site McCracken, consider what he wrote here

[/I]emphasis mine

Because Voros you are looking strictly at numbers, not the actual playing of the game of baseball. You are fixated on the statistical data, not the actual confrontation between the pitcher and batter. You can't qualify things that cannot be quantified. And simply because you don't see it numbers, it does not mean it is not there. They may not appear to have that ability because the numbers don't show it, but it's there. Pitch a few innings Voros, and you'll see.

See now here's where you start to lose me. McCracken humbly admits that he doesn't have all the answers, which is very cool of him. And right NOW, that can't be qualified. But its dangerous logic to assume that one day it won't be as we move forward. Thus the God of the Gaps mention earlier.

Also, its not good logic to say that McCracken needs to have personally pitched in order to understand what makes a good pitcher, or even pitching mechanics or dynamics. I can only go back to my racing example for that knowledge. You can drive a racecar...REALLY well, and intuitively understand what you need to do. You need to be completely fearless, lightening quick, and understand what the car is doing so you can anticipate what to do next. But just because you've never raced a car, doesn't mean you can't understand the dynamics or statistical output of the car. In fact, sometimes you understand it better than the driver. Sometimes not. This is why as a driver, I would go to Kelly Moss or other major tuner shops in order get my car set up. I may know how to drive it, but I don't understand fully the things I need to do in order to be able to really have that extra inch on the competition. And most of those guys have never raced. But if they tell you how to setup your suspension, you'd damn well better listen.

I feel the same way about McCracken, et al.

Per my example earlier, there is a meeting in the middle of knowledge and experience. Baseball pitchers (excepting perhaps a very rare few over the years - lookin' at Mike Marshall) don't have the capacity to be able to concentrate on both. So pitchers would do well to listen to McCracken and learn whatever they can, knowing, as he has admitted, that he certainly doesn't know everything.

[Curve Ball Dave]

Originally Posted by enuttage

See now here's where you start to lose me. McCracken humbly admits that he doesn't have all the answers, which is very cool of him. And right NOW, that can't be qualified. But its dangerous logic to assume that one day it won't be as we move forward. Thus the God of the Gaps mention earlier.

I didn't say that one day someone might be able to come up with a valid metric that explains this. I hope someone does. The quote is aimed more towards those who take McCrackens original conclusion ex cathedra when the researcher himself admits that he really doesn't know how to explain his findings.

Also, its not good logic to say that McCracken needs to have personally pitched in order to understand what makes a good pitcher, or even pitching mechanics or dynamics.

He doesn't have to have actually pitched, but if nothing else he needs to step back from just the pure numbers and start trying to get into the mind of an experienced pitcher. The answers are not always in the data. The answers are in the game.

Per my example earlier, there is a meeting in the middle of knowledge and experience. Baseball pitchers (excepting perhaps a very rare few over the years - lookin' at Mike Marshall) don't have the capacity to be able to concentrate on both. So pitchers would do well to listen to McCracken and learn whatever they can, knowing, as he has admitted, that he certainly doesn't know everything.

I don't really know what a pitcher is supposed to learn from McCracken. They've known for over a century to try to not to walk batters and situational strike outs are nothing new either. There is nothing in the data, as far as I can see, that would have made me or anyone else a better pitcher. As I have stated in my examples, pitchers do all kinds of things that may seem wrong to a statistician but make perfect baseball sense in the context of the situation of a particular game or even a season.

[enuttage]

Originally Posted by Curve Ball Dave

I don't really know what a pitcher is supposed to learn from McCracken. They've known for over a century to try to not to walk batters and situational strike outs are nothing new either. There is nothing in the data, as far as I can see, that would have made me or anyone else a better pitcher. As I have stated in my examples, pitchers do all kinds of things that may seem wrong to a statistician but make perfect baseball sense in the context of the situation of a particular game or even a season.

Alright. I'll join the others in bowing out now. If pitchers have learned nothing new in 100+ years and can learn nothing new from new data application, then you're right. There is nothing to gain from new knowledge, or its application. Carry on. Keep doing the same thing. Don't use all the tools given to try and be better.

This wall now has the blood of enough heads on it. Good luck, take care, and I hope you too are able to step back far enough to try and see a bigger picture that is more inclusive and more progressive.

[Curve Ball Dave]

Originally Posted by enuttage

Alright. I'll join the others in bowing out now. If pitchers have learned nothing new in 100+ years and can learn nothing new from new data application, then you're right. There is nothing to gain from new knowledge, or its application. Carry on. Keep doing the same thing. Don't use all the tools given to try and be better.

(sigh)

I am not saying that statistical analysis does not have benefits and I'm not saying that advanced stats don't tell us things we previously did not know. I'm saying the answers to strange relationships in the data, like McCracken's statement that he has no idea why his numbers are what they are not necessarily in the data. Statistics have their limits. A statistical researcher reaches a point where numbers cannot explain everything and it's time to start looking past numbers and delving into what the subject of your study actually do when they practice their trade. If number crunches only produces more dead ends, doing more number crunching isn't going to solve your dilemma.

Baseball players play the game on the field. The statistics that are produced are the bye product of what they do on the field. The production of stats is not the end goal of a player. The player is trying to help his team win either a single game or a series of games. Stats happen. If you're focused on winning you don't pay attention to your stats because you know the only stat that matters is the team's won/loss record. So in the process of playing the game to win the player does all kinds of things that negatively impact his stats, but they benefit his team. The stats tell us what happened. They don't necessarily tell us why.

As for learning new things from the stats, what lessons are there for pitchers that would have made them better at what they do? The stats say the best thing a pitcher can do is strike out every batter he faces and therefore exert as much energy as possible to achieve this. That's great stats logic but horrible baseball logic. What is good for a statistician is not necessarily good for a player.

[Skipaway]

Originally Posted by Curve Ball Dave

A statistical researcher reaches a point where numbers cannot explain everything and it's time to start looking past numbers and delving into what the subject of your study actually do when they practice their trade. If number crunches only produces more dead ends, doing more number crunching isn't going to solve your dilemma.

Science doesn't work this way. If an animal behavior scientists use statistical method to study how bees scout, and reached a limit on how much he can explain, asking bees aren't really a better idea. Bees actually don't know why they are doing what they are doing.

Those who practice their trade don't necessarily know more simply because they are doing it. Quite often it's the opposite: the fact they are doing it makes them biased and too confident even when their knowledge is false. A lot of technicians do things in certain way because they believe that's how things should work without knowing if there is any good reasons. Ball players are like technicians: can do their jobs perfectly even without any real understanding of the science behind things. Same as drivers are fine knowing nothing about car mechanics.

Dead ends are resolved by finding more info, typically through observing factors that are not previously observed. And you design experiments to prove the theory.

[Curve Ball Dave]

Originally Posted by Skipaway

Science doesn't work this way. If an animal behavior scientists use statistical method to study how bees scout, and reached a limit on how much he can explain, asking bees aren't really a better idea. Bees actually don't know why they are doing what they are doing.

If your data was contrary to your observations, both of which are methods of gathering information per the scientific method, you can either try to gather better data or examine your observation techniques. Regardless, doing more of the same thing that got you to the dead end won't get you out.

Those who practice their trade don't necessarily know more simply because they are doing it. Quite often it's the opposite: the fact they are doing it makes them biased and too confident even when their knowledge is false. A lot of technicians do things in certain way because they believe that's how things should work without knowing if there is any good reasons. Ball players are like technicians: can do their jobs perfectly even without any real understanding of the science behind things. Same as drivers are fine knowing nothing about car mechanics.

All of this is true. And there had been many times in my career where a good use of statistics helped me as a pitcher. Particularly, we charted pitches so we knew how often we threw what. I got into a stretch where I was getting hit kinda hard and didn't know why. I was throwing the ball just as well and there was no evidence that I was tipping off my pitches. We checked my charts. I had become as predictable with my patterns as the sun coming up. This was only discovered by examining the data. So, I changed up my patterns and viola! I regained my effectiveness.

To give just one example. In other cases, by examining the stats we learned what match ups were most favorable.

It's not that statistical data is not useful, it's what data you look at and how. Statistical analysis is a tool in your chest, not the only tool. And just like a mechanic uses the right tool to fix the right problem, so too does a coach or pitcher when trying to fix something. The issue I have with McCracken or anyone else married to his data, is that they are not using all of their tools and in some cases not the right one. As I noted above, observation is a method of gathering information. That's what I mean when I say, "Watch a game". Observation of the behavior helps a researcher understand why some data may seem contradictory or illogical. And as I have stated, baseball players at times do things that seem illogical to a statistician but make perfect sense in the context of helping the team win. A starter who burns up all of his energy by the fourth inning trying to strike out every batter may have been doing the right thing by the statistician, but that does not mean he did the right thing by his team in their collective effort to win the game.

Dead ends are resolved by finding more info, typically through observing factors that are not previously observed. And you design experiments to prove the theory.

Which is exactly what I am saying, only in another way.

[Skipaway]

Originally Posted by Curve Ball Dave

And as I have stated, baseball players at times do things that seem illogical to a statistician but make perfect sense in the context of helping the team win.

I don't think that situation exists. I don't even know what's "illogical to a statistician" mean.

Maybe you are talking about "certain people who quotes certain stats" instead of "statistician".

Typical baseball stats are records of observations. Trajectory data of the ball are records of observations. Data IS the output of observartions, so there is no such a thing of "data being contrary to your observation", because data and observation are linked. I think you got a lot of terms mixed up and confuses.

Statistics is just a tool to help combine vast amount of info and find patterns in it.

[Curve Ball Dave]

Originally Posted by Skipaway

I don't think that situation exists. I don't even know what's "illogical to a statistician" mean.

Maybe you are talking about "certain people who quotes certain stats" instead of "statistician".

I will concede that this is better phraseology than what I used.

Typical baseball stats are records of observations. Trajectory data of the ball are records of observations. Data IS the output of observartions, so there is no such a thing of "data being contrary to your observation", because data and observation are linked. I think you got a lot of terms mixed up and confuses.

I probably did. BTW, data on trajectories is very useful statistical information. It will tell a pitcher or coach where balls are being hit and the data can show changes in trends. Those changes may be good, but they may also be bad and if it is the latter then you know there's a real problem to fix.

Statistics is just a tool to help combine vast amount of info and find patterns in it.

Exactly.

[jbergey22]

I think the talent levels at the MLB level from both the hitters and pitchers kind of make this work out.

I can certainly see CB Dave's point in that at lower levels the pitcher can certainly cause the hitter a lot of frustration and put the ball in play weakly cause less hits on balls in play for the hitter.

Maddux certainly did appear to have an influence on how many weakly hit grounders were hit on him however despite that his BABIP was higher than league average a few years. Perhaps many of these weakly hit grounders end up being infield singles?

The numbers just work out and plenty of data backs it up. I think a case could be made that the ball comes off the bat differently against certain pitchers however the results are still very similar at the MLB level.

I didnt buy this at one time either but just too much data out there right now is convincing.

Maddux may have given up weaker singles than a Livan Hernadez type but it still appeared they were working within the same range as far as BABIP goes. These were very extreme examples because Maddux was usually on the lower end of BABIP while Livan was usually on the upper end of BABIP.

Oddly enough its the fly ball pitchers that generally generate a little lower BABIP than the groundball pitchers. Hence the "limited" control addition to the original findings.

[Skipaway]

Originally Posted by Curve Ball Dave

I will concede that this is better phraseology than what I used.

I probably did. BTW, data on trajectories is very useful statistical information. It will tell a pitcher or coach where balls are being hit and the data can show changes in trends. Those changes may be good, but they may also be bad and if it is the latter then you know there's a real problem to fix.

Exactly.

There are a lot of great developments in baseball these days. Technology advancement helps us to record ball movement accurately. Pitching motions can also be analyzed in details with high speed motion captures. Baseball people are also much more receptive to science approaches, while good money in MLB has attracted a lot of people with good statistical/science skills.

I think right now the area we need to study better is what makes arms tired or hurt. If we have that info better, we'd find a better balance on how hard a pitcher should try when.

[Curve Ball Dave]

Originally Posted by jbergey22

I think the talent levels at the MLB level from both the hitters and pitchers kind of make this work out.

I can certainly see CB Dave's point in that at lower levels the pitcher can certainly cause the hitter a lot of frustration and put the ball in play weakly cause less hits on balls in play for the hitter.

It happens in Majors too. Good pitchers keep hitters off balance so they don't make solid contact. The less solid contact hitters make, the fewer hits they are likely to get. However, there is no metric for "solid contact". BABIP does not differentiate betweeen hard hit balls and tappers. For that matter, it won't tell you if the hits were scattered over six innings or all in a row. In many ways it's a meaningless stat.

Maddux certainly did appear to have an influence on how many weakly hit grounders were hit on him however despite that his BABIP was higher than league average a few years. Perhaps many of these weakly hit grounders end up being infield singles?

Perhaps. Or, Maddux was in situations where he didn't have to knuckle down and go after hitters, so if he gave up a few more hits it didn't matter to him as far as the outcome of the game was concerned.

The numbers just work out and plenty of data backs it up. I think a case could be made that the ball comes off the bat differently against certain pitchers however the results are still very similar at the MLB level.

I didnt buy this at one time either but just too much data out there right now is convincing.

Data may or may not show trends or correllations, but that does not mean the data for BABIP may be very meaningful in evaluating how good a pitcher is. As I said earlier in the thread, the only "repeatable skill" a pitcher needs to have is the ability to give his team a chance to win. What he does to give his team that chance (as he cannot control how many run his team will score for him) is made up of many variables and not all of these can be quantified or show logical relationships.

There are stats that are useful. But stats are just tools in the box, and not every tool is applicable for the job at hand. You must be very careful in using your tools correctly. You use a hammer if everything looks like a nail to you, but baseball is not just nails. Advanced stats have their place as long a they are used correctly, not as a replacement for other tools that are also very useful, or if their outputs are considered the final say in a matter.

[Curve Ball Dave]

Originally Posted by Skipaway

I think right now the area we need to study better is what makes arms tired or hurt. If we have that info better, we'd find a better balance on how hard a pitcher should try when.

Right now no one knows why one guy breaks down and another can pitch forever. Researchers have been studying it for decades. Micky Lolich, who never would have appeared on the cover of a fitness magazine, pitched 300 innings a year on a regular basis while guys who would be magazine covers are lucky to throw 180.

[Goody]

Since I'm a Mariners fan and watch most of their games, I'll give my limited take on it.

Jason Vargas is a prototypical "contact" pitcher. He's good, and can go deep in games, but he doesn't get the same dominant stats as Felix Hernandez. He gives up a lot of home runs because...well...if you're not striking people out or trying to strike them out there's always the chance of them getting ahold of one of your "safety/economy" pitches.

Hernandez is an elite pitcher. He racks up the strike outs but sometimes he doesn't. If he is not "feeling it" he knows to switch modes. But when he is feeling it...the strike out is how he racks up his 3 or 4 dominant starts in a row 90% of the time.

It basically all works out to the style of the pitcher and his talent. Keep the ball where the guy can't hit it well. If you have to talent to do that...it'll result in more strike outs naturally. If you don't have the talent to do that...use Steve McCatty's philosophy and hope for the best.