An experiment in visualizing pitch sequencing

Share

Author’s note: None of this would have been possible without the tremendous work done at Brooks Baseball and their Pitch F/X data. I love them dearly for the hard work they do.

Foreward

I’m fascinated with pitch sequencing. The most alluring thing about the craft of pitching is how it melds art and science together in a sometimes-neat, sometimes-not-so-neat package. There’s a lot of subjective and objective observations to be made when you concentrate on the pitcher. They make the game go, so it’s a thing of supreme beauty to see a pitcher at his absolute best.

Pitching takes time to learn. One of the most basic tools a pitcher needs to learn is pitch sequencing. You hear about this a lot. Announcers love to say “He needs to establish the fastball early,” or “He needs to do a better job of mixing in his (x pitch).”

I think that a lot of us think about pitching in a linear fashion:

linearsequencing

But I don’t think we should. Pitching isn’t binary, and it isn’t linear, not in the way I think about it. Hitters have to constantly keep certain sequences in mind when they face a pitcher. It isn’t as simple as thinking fastball-curveball-slider-fastball.

What if we started to think about pitch sequencing like this:

nonlinearsequencing

 

Which is to say, what if we started to think about pitch sequencing as a series of potential outcomes? To me pitching has always been an organic exercise, so why don’t we think of it organically rather than linearly?

So, I decided to chart two Yu Darvish starts (5/16 vs. DET and 5/21 vs. OAK) and attempt to visualize how he chained pitches together in those two starts.

The chart explained

Before we get into the chart let me explain why I designed it to look the way it does.

It is color coded and the colors represent different pitches. There’s a legend in the upper left corner.

I separated and parsed the data down to sets. Each set is based on what pitch type was thrown first. There’s a chart for sequences based off first pitch fastballs, a different one for first pitch sliders and so forth. Establishing the fastball early is important (as you’ll see) but pitchers never go first pitch fastball 100% of the time.

As for the chart itself, essentially the chart follows this formula:

example1 is the first pitch thrown in a sequence. It’s the biggest object and represents 100%. Any one pitch is open to differing paths. 2 represents the second pitch, and it’s size is proportionate to how often a pitcher (Yu in this case) will go to that second pitch. 3 Represents the third pitch and that size is based off 2. 4 is the fourth pitch and that size is based off 3. I kept it to four pitch sequences because it gets a little hinky after 4.

So…without further adieu…

The Chart

yu copy

OK. That’s a lot of data. So let’s parse it:

First pitch fastball

fastballs copy

Yu went first pitch fastball 46% against 56 batters.

 

As you can see Yu Darvish relies on the fastball-fastball combo a lot. Interesting to note that he doesn’t go Fastball-offspeed that much and tends to pair the hard stuff with the hard stuff.

First pitch sliders

sliders copy

Also worth noting that Darvish likes to pair the soft stuff with the soft stuff.

First pitch cutters and splitters

cutters copy

Yu used the cutter 16% of the time against 56 batters

splitters copy

…and he used the splitter 9% of the time.

These are both the prettiest graphs IMO. Darvish’s cutter looks like an AB ender, where as the splitter works best when paired with a slider.

First pitch curveball

curveballs copy

Yu Darvish does not start hitters off with the curve often. He only used it 6% of the time over 2 starts.

This was a limited sample. Darvish did not throw Uncle Charlie too often over the two starts. When he did he usually followed it up with a fastball.

So, what’s that all mean?

Well, on it’s own it’s simply a visual representation of what Yu Darvish goes to in generalized situations.

However, I’m a visual learner. I understand concepts and numbers a helluva lot better when it’s shown to me rather than explained to me. That’s just how I absorb information, and if you happen to learn the same way I hope this is very helpful in understanding how Yu Darvish pitches.

If you aren’t a visual learner then I think this can simply be a different way of thinking about pitch sequencing. I happen to think it isn’t linear, but maybe you do. Hopefully these graphics will help you think about it in a different manner.

Really this is just the start for me. I’m doing 18 relief innings by Carlos Marmol next to contrast how simplified his approach is compared to Yu Darvish.

In the end, I hope you walk away with this with a better understanding of the information presented here.

Leave a Reply

Your email address will not be published. Required fields are marked *

*


*

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>