Part II: Curveball Velocity, Location, or Movement: What is more important?

Stated in as simplest terms as possible, the goal of pitching is to get batters out without allowing runs to score. There are three ways any given pitch can get a batter out. A pitch can either be swung on and missed, taken for a called strike, or batted in such a way that the batted ball does not result in the runner reaching base. Batted balls involve the defence and are therefore less directly related to the pitch’s effectiveness at getting outs. That leaves us with swinging strikes and called strikes as the two best ways to measure a pitch’s effectiveness.

In Part I of my research on curveballs, I looked at what makes a curveball effective from a swinging strike perspective. I used an outcome variable that I like to call: ratio of effectiveness. Ratio of effectiveness is simply a ratio between swinging strikes and home runs hit. In Part II of my research, I will look at the effectiveness of curveballs from a called strike perspective. This work will aim to answer two basic questions: 1) are curveballs taken for strikes more often than fastballs? And 2) what are the characteristics of curveballs most often taken for strikes?

Are curveballs taken for strikes more often than fastballs?

In 2011, right-handed batters were thrown a total of 252,495 fastballs, 48,286 of which were taken for called strikes. When it comes to curveballs, right-handed batters were thrown 33,515, with 6,431 of those being taken for called strikes. Doing some very simple math, we can see that both fastballs and curveballs are taken for a strike at almost exactly the same rate (19% of the time).

Why curveballs are taken for strikes at the same rate as fastballs is not entirely clear. Maybe batters look curveball just as often as they look fastball? Maybe pitchers throw curveballs for strikes just as often as fastballs? Or maybe it is a combination of reasons (ie. batters look curveball less often, and therefore take it more, however, pitchers might also throw less curveballs in the strike zone, making the called strike rate the same).

Regardless of the reason, it is clear curveballs are taken for strikes at the same rate as fastballs. For the remainder of this article, I will not investigate the possible reasons above (maybe in a future article), but instead; I am more interested in what are the characteristics of the curveballs that are taken for strikes.

What are the characteristics of curveballs most often taken for strikes?

The characteristics I have looked at are the same as in my previous articles: velocity, location, and movement. The difference this time is, instead of looking at a ratio between swinging strikes and home runs as my outcome variable, I will look at a ratio of called strikes to the total number of curveballs thrown. Again, I have limited my data to right handed batters.

With respects to velocity (Figure 1), it is apparent that curveballs thrown in the velocity range between 70 to 80 mph all have a similar called strikes rate of about 20%. At 84 mph, it looks as though the called strike rate drops to about 10%. On the surface, this might look as though it means 84 mph curveballs are easier to hit, but if we look back to my earlier post examining swinging strike to home run rate, we know this is not the case. 84 mph curveballs are considerably more difficult to hit. The only two reasons I can think of why the called strike rate drops here are: 1) maybe faster curveballs look more inviting to swing at, or 2) maybe faster curveballs are thrown in the strike zone less often.

Figure 1 shows the percentage of total curveballs taken for called strikes, broken down by velocity.

With respects to location (Figure 2), it is difficult to draw any meaningful conclusions. In the horizontal direction, it looks like marginally less pitches are taken for called strikes down the middle. As you move away from the middle, more pitches are taken for called strikes…to a point. Eventually, the percent of pitches taken for called strikes drops off the further you move away from the middle of the plate in either direction. I don’t think the percent of called strikes drops off because of what the pitcher or batter is doing, instead I think umpires are less likely to call boarder line curveballs in the horizontal direction. In the vertical direction, location seems to be almost linearly related to called strike %. The higher the pitch, the more likely it is to be taken for a called strike. This makes very little sense, considering from my previous work we know that higher curveballs are hit for home runs at an extremely high rate.

Figure 2 shows the percentage of total curveballs taken for called strikes, broken down by both horizontal and vertical location. For the horizontal direction, zero indicates the middle of home plate and negative numbers are inside to a right-handed hitter. For the vertical direction, zero is the ground.

With respects to movement (Figure 3), it again looks like movement matters very little in terms of a curveball being effective at getting called strikes. The idea of a curveball starting off the plate in the horizontal direction, then breaking back over to be called for a strike just doesn’t seem to be shown here. I’m not saying that occasionally batters aren’t fooled into taking a curve for a called strike because the batter thought it was going to be a ball; I’m just saying that I don’t think it doesn’t happen very often. Or, maybe it is just that the umpires simply don’t call pitches that start out looking like they are off the plate for strikes, even if they do end up crossing the plate.

Figure 3 shows the percentage of total curveballs taken for called strikes, broken down by both horizontal and vertical movement. For the horizontal direction, negative numbers are inside to a right-handed hitter.

When comparing good pitches to bad pitches, looking at called strikes does not give us nearly as much useful information as swinging strikes to home runs. I think a big problem with looking at called strikes is the confounding factor of the umpires call, especially from a location perspective. Eventually, I would like to get to the point where I am looking at the entire ‘art’ of pitching in an analytical way, however right now, I have not yet developed a sufficient set of tools to look at how the pitcher sets up his pitches with previous pitches, and how he uses this to be successful.