Changing the Way We Look at Offense, Part Three: Pitching

To understand everything about offense, you have to know at least a little bit about pitching. After all, part of building a hitter's offensive game is understanding what pitchers will do to pitch to him successfully.

Before I can talk much about pitching, I want to talk about why I am only including starters. Simply put, they provide more data (I only had two years to pull from). They pitch in larger samples, which is what I really needed to do analysis with the time frame I had available to me. After all, Inside Edge hasn't collected WHB for a long time, and this is partly because the company simply isn't that old (relative to the age of the game of baseball), so there isn't a ton of data available. This made me set relief pitchers aside and focus on starters.

Now to the theory: just as hitters have three goals at the plate, pitchers have three goals. However, they are not the same three outcomes that come from DIPS theory. For pitchers, they want to avoid hard contact (not just homers!), avoid walks, and rack up strikeout totals. When these things are done, runs are prevented and positive outcomes are had.

Before I go into each of the three key outcomes, I want to talk about the relative importance of each aspect. When I ran a regression, the relationship between the frequency of WHB (WH%) and WHOBP produced an R value of .82, which suggests that the largest and most important part of offensive production (as far as this metric is concerned) is WHB.

This doesn't come as much of a surprise, either. As was noted in my second installment, power is a huge result of producing WHB. This was shown again when I tested the relationship between WHOBP allowed by pitchers and SLG allowed by pitchers (SLGA), which produced an R value of .51.

The next important thing to test is the importance of avoiding walks. For this portion, I will be using walk percentage a lot. The first thing I did was test BB% against ERA+. When I ran this, the relationship between the two produced an R value of .32. This suggested that a higher BB% often leads to a lower ERA+, which really isn't a surprise. However, time to get back on topic. How does BB% relate to WHOBP? Well, I'll tell you: the relationship between BB% and WHOBP produced an R value of .52. Again, this suggests that walks are a big part of the offensive game.

Now, before I get to strikeouts (definitely my favorite part of this installment), it's time to look at some other basic events. Here is a table with some relationships that I will explain below:

Let's start with the first one. PPPA stands for "Pitches Per Plate Appearance." I have heard often through media outlets (especially color guys on television broadcasts) that pitchers throwing more pitches are being inefficient and that they are wasting energy. I wanted to test this against WHOBP to see if there were any noticeable consequences.

The resulting R value was .18, which really suggests no true negative relationship between throwing many pitches and having success (after all, it's entirely possible for a hitter to get a WHB on the first pitch of a plate appearance and strikeouts require at least three pitches).

The next group involves "Plate Discipline" metrics that are available on Fangraphs (albeit under different names). The first is In-Zone% (IZ%), which is just the frequency by which a pitcher is throwing pitches inside the strike zone. When related to WHOBP, no positive relationship was found, which suggests that it's not necessarily vital for a pitcher to throw a ton of pitches in the strike zone for him to be successful.

The next was Miss%, which was explained in my second installment. This relationship produced an R value of .40, which begins to suggest that missing bats is good (not that that is a surprise).

The relationship between K% and WHOBP produced an R value of .59, which suggests that racking up the strikeouts is a big part of preventing offensive success. This supports the DIPS theory and my offensive theory pretty strongly. But again, I'm more interested in the process, so what is being done to record strikeouts, and what is being done to most prevent offensive success?

The answer is Chase%, which is referred to on Fangraphs as O-Swing%. This is the percentage of the time that a pitcher gets a batter to swing at pitches that are outside of the zone. It is a reflection of deception, command, and movement. The relationship between Chase% and WHOBP produced an R value of .64, which suggests that getting hitters to chase pitches out of the zone is a huge part of preventing offensive success.

It also does a lot to show why scouts are so big on a pitcher being able to command his pitches. Being able to trick batters, deceive them, and to throw what you want when you want is what leads to outs ("the best pitch is a well-located fastball" and such).

This revelation was really the key in deciding I was ditching DIPS for WHB theory. I believe strongly that pitchers heavily influence whether balls fall for hits or not. With Chase% relating to success, I have to believe that pitchers can use deception and movement to limit the quality of contact, especially by getting hitters to swing at pitches out of the strike zone. Remember, a WHB has to be hit on the barrel. If the ball isn't hit on the barrel, then the quality of contact is weakened and it's easier for fielders to record outs. It's a sign that a hitter isn't doing something right, which is a sign that the pitcher has been just effective enough for it to matter. Obviously there's still a lot of luck involved, but I'll get to that in later posts.

The most important relationship tested here, however, was the relationship between WHOBP and ERA+. If WHOBP is going to have any relevance to pitching, it needs to be able to relate to ERA+. The resulting .65 R value shows that there is indeed a relationship between the two metrics, which suggests that keeping the opposing offense's WHOBP down directly influences the ERA+ of a pitcher in a positive manner.

Now to get to those strikeouts. Here is a table showing some relationships:

This was pretty surprising. No significant relationship appears to exist between striking out more batters and preventing WHB or BB. However, K% is a key in preventing SLGA and ERA+ with respective R values of .50 and .47 being observed. This suggests that K% is key in preventing offensive success, but it doesn't necessarily mean a pitcher is going to walk more batters or prevent balls from being hit harder.

Conclusion

These are the basics of how WHB relate to pitching. The DIPS theory is lightly supported by my WHB theory. The essential difference between the two is that instead of sticking to the "True Outcome" of home runs, I am looking at the "True Outcome" of Well-hit balls.

By doing this, I am able to account for the amount of quality contact that a pitcher is giving up, because a pitcher absolutely has the ability to influence how many WHB he is giving up. After all, Chase% relates very well with success, which suggests that a pitcher has the ability to limit strong contact (thus theoretically reducing BABIP and HR/FB).

Looking at pitchers provides an avenue for better understanding WHB. After all, different events impact hitters and pitchers differently. In the next installment, I will join offense and pitching together to look at the importance of Ball-In-Play data (BIP). It will look into the importance of fly balls, ground balls, and line drives as they relate to creating and preventing offensive success. As a sneak peak into installments beyond the BIP installment, I will say that the expansion of WHOBP is just on the horizon. Soon we'll be able to get into the raw guts of what WHB means for baseball.

---------------

Credit to Inside Edge and Baseball Reference for the data.