About a week ago I looked at the strike zone from the pitcher’s point of view, and showed how it did not appear that the result of the pitch was in any way influenced by the location within the strike zone. I broke all the pitches for an individual pitcher into types, and then into zones, and could not show the expected pattern, which would be something along the lines of if you throw a pitch down the middle of the plate, hitters will feast on it. The results, using several different pitchers, suggested that where the pitcher threw the ball (once you took out balls and called strikes and leaving just pitches the batter swung at) did not matter, that the rate at which various events occurred was similar anywhere you looked in the strike zone. The one place I found was that down and away to a right hand batter, they were much more likely to swing and miss than anywhere else, but otherwise results were pretty much the same across the strike zone.
Once I found that, I wondered if the same would be seen from the batter’s point of view. We’ve all seen those hot and cold zone charts, where they break out a particular batter, and tell you how they are doing in that zone. For example, batter X will hit balls middle inside at a .350 clip, but middle outside he’ll only hit .250, and so on. If that’s the case, we should be able to see at least something from some different batters that would interest us. My suggestion last week was that perhaps pitchers are so homogeneous because they are an agglomeration of multiple hitters, and if the hitter is really the one who influences the location it might be hidden in the pitcher data because of the large numbers of different hitters they face, all contributing to the pitcher’s chart.
So this week I performed the same test as last week. I took the top six hitters from my Gameday database, based on number of pitches tracked in the database, and ran the same set of charts on each of them, then looked at the data. The hitters in question were Ichiro, Adrian Gonzalez, Mike Cameron, Richie Sexson, Rafael Furcal, and, in case you were wondering why I chose six instead of five like the pitchers, Michael Young. I pulled their data, dropped their charts into my templates, analyzed the numbers, and guess what? Well, let’s look at some charts to see. Since this is primarily a Rangers blog, I’m going to show you Michael Young’s charts:
First, the overall picture. If you look closely you might detect some patterns, but I’ll make it easier on you and break it out by result type.
Here are just the foul balls. As always, this is from the catcher’s perspective. You might already tell a slight pattern, which is that there are more pitches to the right, and you would be right. In fact, of all the pitches Michael Young swung at, about 55% were to the right hand side of the strike zone (away, of course, since he is a right hand batter). As a matter of fact, once we take into account the number of pitches into each area, we find that Michael Young fouls off about 40% of all pitches he swings at, with the exception of down and in, where he only fouls off about 25%. More on that in a moment.
Here are the in play, no out results, which is basically his hits. These appear to skew slightly to the inside, but that is an optical illusion caused by one thing: there is not a single IPNO result on the right hand side, below 2 feet. And on the left side, below 2 feet, there aren’t very many either, but enough to make you think it skews left. Is this surprising, that he doesn’t get many hits when the ball is down, and doesn’t get any when it is down and away?
Here are his outs. In a repetition of the results found for pitchers, he shows a much higher rate of making outs when the ball is inside, especially down and in where the rate is about 50% higher than the other zones. In the pitcher article, I attributed that to the ball coming inside and batters hitting grounders to third or short. This seems to back up that theory. This in fact is the partner of the foul ball: there were many less foul balls inside, and down and in especially, because there were many more outs inside, and down and in. In fact, when adding the two types together, their rates stayed remarkably stable, suggesting that fouls and outs are closely tied together. I’m not sure why this would be, except to say perhaps that, as above, instead of hitting the ball foul down the left field line he is hitting it fair on the left side.
If you think about that, you realize that a right hand hitter will foul most of his pitches to the right side, and not many to the left. Imagine, if you will, a baseball diamond from above, with the 90 degree angle of fair territory. Imagine a hitter hits to, say, 120 degrees of angle. If the angle was even with fair territory, he would hit fouls to 15 degrees on the left and 15 degrees on the right. However, since he is a right hand batter, he tends to foul to the right, so place the left baseline of the 120 degree angle down the third base line, and the hitter will foul pitches off an extra 30 degrees to the right. Vice versa for a left hand batter. The closer to center, the more likely a hit, so in the case of these pitches inside, if a batter hit evenly across fair territory they would merely be fouls down the left field line, but since he tends to hit more to the right, they stay fair and end up with the third baseman or shortstop throwing him out.
Okay, that was not very well explained. I’ll have to make a picture of some kind, and come back to that later. Let’s move on.
Here are his play results where he put it in play and runs scored as a result, not too many so not necessarily a large enough sample to deduce anything, and in fact there were really no areas where the rate was higher or lower than any other.
And, once again, our final picture is the most interesting one. If you’ve watched Michael Young bat as many times as I have, you know he is very susceptible to swinging and missing at the pitch down and away. Here is that pitch, shown quite clearly in this chart. Four times as many swinging strikes on the outside as on the inside. Most of those strikes are in the middle or lower. All those previous charts where he wasn’t hitting the ball much down and away, they all end up here, as he swings and misses those pitches. In fact, he is over two and half times more likely to swing and miss down and away, as he is to swing and miss anywhere else. The middle away zone does show a lot of dots, but you also have to remember there were a lot of other pitches there too. There are 22 swinging strikes out of 59 pitches down and away, and 24 swinging strikes out of 124 pitches middle away.
Now, back to something I said at the top. I said I’d show you Michael Young’s chart, because this is a Rangers blog. Well, that is true, but there’s another reason. All five of the other batters showed very similar patterns. Oh, there were minor differences, like Ichiro’s left-handedness causing his swing and misses to be down and to the left, and their areas did show different shapes, Sexson’s being a little taller presumably due to his height, but really the overall sense was the same as Michael Young’s: the outcome when they swung did not matter on where the ball was, except they would ground to third (first, if they were a lefty) more on inside pitches and swing and miss more on outside down pitches.
So again I ask the question: What is the use of location within the strike zone? If you’re just throwing it randomly in there, you’re going to get a fairly random result. In fact, if you want, just throw it down and away all the time, hitters won’t have much chance of doing anything with it. That is crazy, of course, because hitters will adapt. If you decide to just throw to one spot, pretty soon the hitters will be waiting for a pitch there, and they will tee off on you. And I’m not advocating complete lack of command, either, because that would be fatal in other ways. If that were the case, you’d throw too many balls, and the whole point of this study was to ignore what the pitcher was throwing outside the strike zone, but rather to look at what the batter felt (for whatever reason) was worthy of swinging at. No, the point is to induce the batter to swing at the pitch, and it’s only then that this study appears to show that the result when they swing is random.
So now that I’ve gotten this far, what’s next? If you rule out location as a factor in a result, what are you left with? First, speed, followed by the various breaks (vertical, horizontal, amount of break on the pitch), angles, types of pitches. Does it seem like one (or more) of them should stand out? It does to me, because if it doesn’t, then what is the skill of batting, other than just randomly swinging at what the pitcher throws? I’ll start working on that, and let you know what I find.
I would also like to find somewhere that actually shows me the heat zones I referred to earlier (or even scouting reports that say something like “don’t throw inside to this guy”), and see if I can replicate them with the Gameday data I have. I question whether I could, based on this admittedly small sample of data. But I don’t know, maybe people have imagined things, or simply not measured them accurately, and written down their expectations? It wouldn’t be the first time that “experts” were shown to be wrong when someone actually bothered to measure something. It also wouldn’t be the first time that I’ve looked at some data and come to a conclusion that’s not necessarily supported by deeper analysis. More study is necessary, I think.