In my previous posts on the Gameday data, I first looked at the release points of the Rangers rotation, then looked at each player’s pitch types. Here in part three I will put the two together, in an effort to see if they are tipping their pitches by where they release them. As I noted the other day, it would be counter-intuitive to discover such a thing, because if a pitcher releases their different pitches from different locations, batters will quickly catch on and be able to tell what they are throwing. On to the charts:
Loe shows three pitches, a fastball/sinker, changeup and curve. I have not been able to differentiate between a regular fastball and a sinker in his data, so I am treating them the same at the moment. From his chart I see little or no differentiation between pitches, none of the colors stand out as being separate from the others. Without mathematically analyzing the three groups (something I may do later), I would say he is not showing hitters anything from where he is releasing the pitch. Interestingly, his last two starts were very good, after spending a couple of days in the minors, but I have not looked at those starts to see what they might show differently.
Cursed with only two pitches, a fastball and a slider, he’s also cursed with tipping them a little. Okay, it’s not much, but I can see that the blue sliders are higher than the red fastballs. In the small group at the bottom right, which I believe was a glitch in Gameday which caused one day’s data to measure off a little, you can clearly see the difference (in fact, although every other chart today is on the same scale, I had to increase this one vertically by a foot to show that extra data). Overall, although the horizontal release point is very similar, I would guess the vertical release point is about three inches higher for the slider. I know what you’re thinking, three inches is not that much, especially from 55 feet away (where Gameday measures release points from home plate). But remember, these guys are able to hit a ball that is 2 7/8 inches wide, travelling at 95 mph. They are able to tell what type of pitch based on what the stitches on a ball are doing as they come towards them at that speed. I think a three inch difference would help them a lot.
I didn’t color this one very well, but I was trying to diminish the effect of the fastball, because it was so dominant. I also wanted to keep it at the same scale as the others, to show how much smaller the area of McCarthy’s release point is. Click on the picture to go to my Flickr site and see it larger if you want to. What it shows is that his pitches are very similar, except for the curveball (red), which he appears to release a little further up and to the right compared to the others. Not much, but as noted they may not need much. The advantage he has is that it is still in an area which is filled with the other pitch types. If a hitter was to see the ball coming from top right, he wouldn’t necessarily be able to tell what it was, but if it came from bottom left he might be able to know that it is not a curveball.
The most troublesome pitcher to identify his pitches, and I took these ones a step further than in my previous study. I believe I found a way to differentiate between his changeup and slider, and have marked those pitches in this study. I will elaborate on that in a later analysis. In the meantime, Padilla is all over the place. His changeups are mostly in the top right, while his sliders are mostly bottom left. The other two pitches are scattered all around. I note that Padilla went on the DL today, and in part of the reasoning they said that he was pitching okay for a couple of innings but then his elbow would tighten up and not allow him to throw properly. Could his wide area of release be caused by the injury? That seems like a prime cause, if you can’t throw the same way each time you’re going to be all over the place.
Once again I save the best to last. Kevin Millwood is a veteran pitcher going through a tough year. We’re not really sure what’s wrong with him, but this is a huge clue to me. A bunch of bright orange to the top right, all the blue and green bottom left. He’s throwing the curve and slider in a similar position, but the fastball is being released
about 7 inches right and 5 inches higher. Tell me that’s not a huge difference! I believe a major league hitter would pick up on this and be able to tell fastball or not, and that could very easily be the difference in being able to hit it or not.
We can see that McCarthy, Loe and Padilla are throwing their pitches throughout their zones. Padilla probably due to his injury, and Loe due to being a little uncontrolled, but McCarthy appears to have good control (a tight release zone) and pitches spread throughout. This suggests he has been the best of the Ranger pitchers (remember my first study which showed that the tighter your release points, the lower the ERA), and in fact right now he is the only Rangers starter with an ERA below 6.00. Tejeda is already in trouble by only having two pitches, but with the possibility that he is showing them by the way he releases, that’s a double blow for him. Noting that his ERA has gone up and up as time goes on, other teams might have caught on to this. Millwood shows even more differentiation in his pitches, which could lead to him being hit more as time goes on. I can’t imagine that a veteran could have gone so many years without this being noticed before, so it is possible it is a new and correctable problem.
Now we know where they’re releasing their pitches and what they are throwing. Next up will be a look at when they are throwing it: vs left or right, what count, what score, what baserunners. This will be a more complicated analysis, and I will have to rein myself in to not do too much at once, and bury the signal within the noise. At this point I have several hundred pitches for each starter, but I will try and not chop it down so finely that the number of pitches is meaningless (the old “9th inning or later, score tied, runner on third, with the temperature below 58″ problem). The next article will hopefully only take a week or so to post.