I mentioned Johnny Cueto the other day, but I wanted to take an even deeper look today. I hope you’ve been reading some of the excellent analysis done with PITCHf/x already. If not, you’re in for a treat, as I think it can lift analysis to new and unprecedented levels. This will mark my first real foray into using these date, and before I start, I’d like to throw out a few thank yous to the people who have made this possible.
First, a big thank you to MLBAM and Sportvision for having the vision and capability to install the PITCHf/x system in major league ballparks as well as their continued efforts to improve the system.
Second, a hearty thank you to fellow THT writer Josh Kalk. In all of my PITCHf/x analysis, I will be using the data he has collected, classified and corrected. This sounds like a daunting task to say the least, and he has not only done a fantastic job with it, but has been kind enough to share it with us. Bravo, Josh. In addition, he’s helped me with a ton of questions I had about the system and uses for the data.
I’d like to also thank fellow THT writers Mike Fast and John Walsh for their assistance with additional questions I have come across as I begin to use these data and for their (Josh included) excellent, field-leading research that has paved the way for people like me. I don’t want to leave anyone out, so to everyone else who has contributed to our understanding of PITCHf/x and have come up with new uses for the data, thank you as well.
Now, let’s get into the analysis!
Note: All numbers in this article exclude Cueto’s start from May 21. I don’t currently have the PITCHf/x data for that start, so for the purpose of continuity, all stats exclude this start. Not a whole lot really changes, although his strikeout rate would drop a little.
In this and in future analyses, we’ll start with the usual numbers to see what a player has been doing, then moving on to the PITCHf/x stuff to see how he is doing it.
YEAR AGE GS IP ERA WHIP W SV 2008 21 9 51.7 5.75 1.20 2 0
YEAR AGE GS IP LIPS DIPS WHIP K/9 BB/9 xGB PERC 2008 21 9 51.7 3.43 1.18 9.23 2.26 35
YEAR AGE GS IP LOB% BABIP HR/FB LD% RS* TEAM R/G* 2008 21 9 51.7 61 0.278 18 17 3.34 4.27
*Note: I’ve added Runs Support (RS) and Team Runs per Game (TEAM R/G) to the stat line. A pitcher who isn’t receiving his fair share of a team’s offensive production is likely to have fewer wins than he deserves.
Cueto clearly has talent. His LIPS ERA is a very healthy 3.43, and he’s striking out more than a batter per inning to go along with above-average control. The only thing to worry about, as I noted at the beginning of the season, is his low ground ball rate and Great American Ballpark’s tendency to inflate home runs by 28 percent.
He will likely post a HR/FB above the league norm of 11 percent the rest of the way, but 17 percent is far too high. This is partially the cause of his large ERA to LIPS ERA. The other culprit is his far too low left on base percentage, which we discussed the other day. His skills look fine when pitching from the stretch, and I would expect a LOB rate much closer to league average going forward.
When these two stats regress, Cueto’s ERA should plummet. This information alone would make Cueto an excellent buy low candidate. Let’s go deeper, though, to see how Cueto is doing on a start-by-start basis.
True Quality Starts
If you’re new around here, you can read up on True Quality Starts here. If you’re not looking to read a long, detailed explanation, True Quality Starts basically uses linear weights on a pitcher’s skills (strikeouts, walks, batted ball breakdown) to calculate a “TQS Score” and takes a standard deviation approach to classify every start a pitcher makes into one of six categories: Great, Good, Above Average, Below Average, Bad and Awful.
Here are Cueto’s TQS numbers so far this year. The first row gives the raw number of starts in each category since we’re still early in the season, and the second row gives the percentages.
True Quality Starts
Great Good AbAv BlAv Bad Awful TQS* GG* BA* GG/BA* 1 2 3 1 2 0 6 3 2 --- 11% 22% 33% 11% 22% 0% 67% 33% 22% 150%
Note 1: Cueto’s starts were plugged into the 2007 run environment because it is simpler this way and deals with a larger sample size. It is also probably more reflective of what the final 2008 run environment will look like than the current 2008 run environment would be.
*Note 2: TQS is the number of Above Average or better starts. GG is the number of Good plus Great starts. BA is the number of Bad plus Awful starts. GG/BA is simply a ratio of the two.
While these numbers look pretty good, they aren’t amazing. They aren’t quite as impressive as his raw peripherals are.
Here’s why. Cueto hasn’t developed the skill (or maybe the confidence from his manager given his 5.75 ERA) to go deeper than six or seven innings into the game. While he has posted some very good strikeout and walk rates during a lot of these starts, you really have to be stellar in order to get a Good or Great rating if you only go six innings.
Cueto also doesn’t induce a lot of ground balls, so in games in which he induces less than his average of 35 percent, this can also affect the numbers. That isn’t an excuse, but I wanted to mention it. The innings thing, I think, shows that while his starts aren’t technically “Great” yet, the potential is there should he gain more stamina.
In one “Above Average” start, he posted a 9.0 K/9, 1.5 BB/9, and a 33 percent ground ball rate but went only six innings. In another, he posted a 10.5 K/9, 3.0 BB/9, and 46 percent ground ball rate but again went only six. The skill is there, he just isn’t putting up the innings yet.
Now let’s look at what Cueto is doing with his pitches to achieve this early success. First, his movement chart, which is one of the most interesting I’ve seen since the inception of PITCHf/x.
Before we comment on it, let’s look at one more.
Notice Cueto’s sliders and change-ups. Not only do they have similar vertical movement and blended, seamless horizontal movement, but they are also thrown within the exact same range of speeds.
Because of this, they were actually all classified as sliders to begin with. You’ll noticed that I artificially separated them into two groups—sliders and change-ups—at the 0.00 horizontal movement mark. The sliders probably extend a little bit into negative territory, but does it really matter what we call them? Their characteristics are so similar and they blend together so easily that I don’t think it matters much in the grand scheme of things which is technically a “slider” and which is a “change-up.”
This certainly is strange to see, but is it a good thing or a bad thing? Let’s look at a few more graphs before we answer.
I first saw this graph (and the next one) in Josh’s curveball and slider articles that have run over the past couple of weeks. A big thanks to Josh for the patience to help me learn to put them together myself.
As we would expect given the nearly identical vertical movement on the two pitches, Cueto’s slider and change-up take almost the exact same path when we look at it from a side angle. Let’s see how they compare when we take an overhead view.
Here, we see some divergence, which occurs pretty close to the plate. We see that the two pitches come in on very similar paths, but then they dart in opposite directions 20 feet or so from the plate. This has to be difficult for batters, seeing the pitches coming in with the same vertical movement, at the same speed, and from the same release point (something we’ll examine a bit further in just a second), but not knowing which way it will dart until the ball is almost at the plate. I think this kind of deception is giving Cueto a big advantage.
Also, remember that we are looking at the “average” slider and “average” change-up for Cueto. We know from our movement chart that both Cueto’s slider and change-up have a great deal of variability in their horizontal movement. The two have a range of horizontal movements from -7.68 to 5.97, a difference of 13.65! That is absurd. So while the gap doesn’t appear huge in the top view graph, there is actually an extremely wide array of possible movements the pitch can take, and the batter has very limited time to react.
Essentially, the batter has no idea if the pitch coming in is a slider or change-up (assuming they’ve first decided that it’s slow enough not to be a fastball). Then, once they see it moving in one direction or the other, they need to decide how far it’s actually going to move.
On this topic, let’s look at the late break Cueto is getting on his pitches. Excellent idea on the part of Mike, and a big thanks to him for teaching me how to put it together. Here’s Mike’s description of late break from his PITCHf/x primer.
The goal is to show something close to what the hitter perceives as the break or movement of the pitch. I calculate the deflection of the pitch due to two forces, spin and gravity, in the last 0.25 seconds of its trajectory before it crosses the plate, an idea I got from Tom Tango. I chose a quarter second because that’s roughly the reaction time of a batter executing a swing. I chose to include the effect of gravity because I believe that more accurately reflects what hitters see. Hitters don’t attempt to hit a gravity-less pitch; they attempt to hit a pitch that’s being affected by gravity and being deflected by spin.
Again, we see a good deal of variability in the movement of Cueto’s pitches (especially on the change-up side), this time in the last quarter of a second before it crosses the plate.
Let’s now take one more look at Cueto’s deception.
I’ve never seen these charts used before, but I think they can add a lot to our discussion. We see that Cueto is being very consistent in hiding his pitches as they come out of his hand. The change-ups and sliders are on top of each other, and they essentially mirror the fastballs.
The kinks are still being worked out of the PITCHf/x system, and I believe Great American was one of the parks to experience trouble early this year, so there is a clump in each graph that doesn’t really fit (from -3 to -2.5 in the horizontal graph and right above 5 in the vertical graph). They probably should be shifted over, but even if they aren’t, all three pitches still mirror each other in both graphs. If we ignore these or pretend that they are shifted, we see that Cueto would have a pretty compact, repeatable release point. Always good to see.
All in all, I think Cueto looks like a good play. He is getting unlucky so far in terms of his ERA, but his peripherals look very good. He’s already turned in one TQS “Great” start and two “Good” starts, and if he had a little more stamina he would have at least a couple more.
As far as his actual pitch data go, there is a lot to like about Cueto. He can bring the heat with his fastball, but he also has a slider and change-up capable of very good movement. I would have to conclude that the wide variability in the movement of Cueto’s slider and change-up do indeed benefit him. Combining this with nearly identical release points, very similar paths to the plate, good late break, and the deception of all of this combined, I think Cueto has found a very effective way of using these pitches.
As I mentioned a few days ago, I’ve already traded for Cueto in one league and I drafted him in another. I’ll be looking to buy in the rest of my leagues, and I would advise you to do the same. Cueto looks like one of the best bargains in fantasy baseball.