Yuki Matsui: Start 3

Yuki Matsui made his third start of the season on Wednesday April 16th, facing 26 batters in five innings and 101 pitches, giving up three runs, while striking out four and walking six against the Softbank Hawks. Like usual, let's start with the graphs and start with his velocity chart.

Unlike the last two starts, Matsui's velocity peaked in the middle of the outing instead of around the first of the outing. His slowest pitch overall was a curveball early in the outing. His fastball velocity dropped a little toward the end until his last start. Here are his average pitch velocities from the outing:

At least compared to his first start, his fastball velocity is down over a full kilometer, and his change is down about two kilometers. His slider and curveballs were actually faster though. Whether or not that is a classification error or not is tough to say.




Here is a pie chart that shows his pitch selection:

We still see a relatively low fastball usage, and it has actually been decreasing compared to his other two starts. He really fell in love with his slider in this particular start.

Again, Yuki Matsui struggled to throw pitches in the zone, as this chart shows:

 Since he struggled again to throw pitches in the strike zone, let's see where he located his pitches:

Once again, Matsui was very heavy to the glove side and low. Armside and high was his second favorite spot. He kept the ball out of the very middle of the plate though, even though he still gave up five hits.

Here are the results of each of Matsui's pitches broken into a pie chart:

His whiff percentage was pretty impressive, though his called strike percentage wasn't. To get a better look at pitch results, let's bring the velocity chart back up but instead of labelling each pitch with pitch type, let's label them with the results.

When he was throwing his hardest in the middle of the outing, he was throwing balls, which probably means that he was overthrowing. His breaking balls at the end of the game were also almost all balls, except for the one whiff. He got a couple of called strikes with early breaking balls, just like the early fastballs.

Here are the pitch locations for the balls put in play off Matsui:

As you can tell, things were pretty balanced, but his increase in pitches thrown arm side and high seemed to hurt him, as that is the location where most of the balls were put in play.

Here are the pitch types that were put in play off Matsui:

Not surprisingly, a plurality of them were fastballs, but his slider and change had higher percentages of contact than pitches thrown, meaning they were more hittable than the curve and fastball in this start.

For comparison, here are the pitch types that Matsui got whiffs with:

The slider was a risk-reward pitch for him in the outing as he gave up more contact than he should have with it, but it got him half of his whiffs.

Masahiro Tanaka: Start Two

Masahiro Tanaka followed up his solid debut with the New York Yankees with a strong second start, throwing seven innings and striking out ten, walking just one, and giving up three runs to the Baltimore Orioles. There was a key difference in the start; he didn't get grounders.

Like last time, let's look at the home run at-bat first:

The whole graph is the strike zone, and the splitter (notice how MLBAM has fixed the tags by calling it a splitter instead of a changeup) is what was hit for a home run. The slider was actually called a ball despite being in the traditional strike zone, though clearly no one wants to throw a slider there. Now, let's look at how he pitched the same batter (Jonathan Schoop later in the game (both at-bats), with the results as the labels this time.

You can see that Tanaka was able to lower his pitches the next two times he faced him, though he worked both sides of the plate. While we are on the subject of location, here is where he located his pitches on average (using MLBAM tags).

The MLBAM tags have changed all of his 2-seamers and cutters to sinkers, which I don't think is right, though he kept it arm side like a 2-seamer. His slider control was horrible, as he couldn't even get it glove side enough to be in the strike zone on average. His curve also stayed high and arm side, though his fastball was glove side and in a decent spot. His splitter remained low, with the home run a big exception. When looking at his pitch selection via the spin and speed chart, we see that he really didn't use the slider much, perhaps because he was struggling so much to locate it:

Here is what Masahiro Tanaka's release points looked like from a relatively normal view:

Just for curiousity's sake, I wanted to look at a closer view of the release points labelled with the MLBAM tags:

Going back to location, I wanted to look at all of his pitches, labelled with the results:

For a cleaner look, here are the average locations of the results:

Tanaka also pitched extremely away from left-handers, but kept the ball away from right-handers as well, as this average location graph shows:

Lastly, let's look at how Tanaka pitched by count:

It seems that you can see a general trend of him staying extremely arm side with no strikes in the count, and going glove side with two strikes.

Yuki Matsui: Start Two

Yuki Matsui made his second career official NPB start for the Rakuten Golden Eagles on Thursday, and packed 86 pitches into three and one thirds of an inning. Out of the 17 batters he faced, he struck out 7 and walked 5, with 2 runs and 3 hits. Similar to his last start, what follows will be a few charts and graphs on his start. Like last time, let's start with his full velocity chart from pitch to pitch:

While he threw less pitches, it does appear that he maintained his velocity a little better, though again, he peaked out early in the outing. When only looking at fastballs (all Matsui throws is the 4-seam fastball, or at least that is the only classification he has through two starts), I think we see this exaggerated a bit.

His fastball averaged about 87.48 MPH in the outing, between Dan Haren and Jason Vargas in the Majors in 2014. How does his pitch selection compare to his last start? Here is the pie chart from start two:

All three off-speed pitches saw increases, with him throwing less fastballs overall. The slider is what increased the most. The pitch actually lost about .2 kilometers per hour on average from his last start, slightly less than the fastball (.7 kilometers per hour on average). Here is how he located the pitch:

Low and glove side, where you want a slider to be, was the most common place, but too many of them stayed high or in the middle. Surprisingly, more than half (13 of 23) were actually thrown in the strike zone. Perhaps this is why he threw it so much, as he had problems throwing in the strike zone overall:

So the slider may have been the only pitch that Matsui could actually get over, and since it wasn't hit (his problem was walks not hits), then he had no incentive to go away from it. So where did he locate the majority of his pitches?

Because of the slider, low and glove side was the most common location, though he threw more pitches in the middle and arm side in this outing. A simpler chart:

He didn't throw hardly any pitches low and arm side or middle and arm side, which made me curious about his changeup locations, which you can see here:

Weirdly, he throws the pitch to the glove side, or at least he did in this start. Just one of the seventeen changeups turned out to be in the zone, so it is easy to see that he doesn't have much feel of the pitch.

Now, let's take a look at Matsui's results:

The higher volume of balls from last start isn't surprising, though he increased his whiff percentage and called strike percentage. When he threw strikes this time, he was more effective than his first start, as there was even a lower foul percentage.

Masahiro Tanaka: Start One

Masahiro Tanaka made his much anticipated debut for the New York Yankees on April 4th, pitching seven innings with three runs allowed, striking out eight and walking none against the Toronto Blue Jays in Toronto. Tanaka got twelve groundballs versus six flyballs, but famously gave up a home run to the first batter he faced, Melky Cabrera, a switch hitter (so in this case, batting left handed). I thought I should start the post by looking at that at-bat.

The pitches are labelled by the MLBAM tags and by the pitch number in the at-bat. Tanaka obviously attempting to keep the ball away from him at all costs, and kept the first two pitches low in the strike zone, with the slider staying far enough arm side that it was a ball. Tanaka then hung the splitter in the strike zone, and even though it was away, it turned into a home run. However, I thought it was interesting what Tanaka did to Cabrera the next time he faced him, when he struck him out.

This time, Tanaka stayed high in the zone on the first four pitches, even with a curve and a slider, which were both balls, and both probably mistakes. However, he went inside with a fastball and got a foul, and then went away with a fastball and got another foul. Tanaka finished the at-bat with his patented splitter, keeping it arm side (away from Cabrera) and low in the zone. 
'
Before taking a closer look at his stuff, let's take a look at his release points:

So his release point seems to be a little lower and more out than I had thought in my original scouting report. The original scouting report gives us an idea of his stuff, but here is his speed and spin chart from his first Pitch F/X start:

 We see five real clusters here, suggesting about five pitches. The curve is the pitches on the far left and bottom, the slider in the middle and low, the splitter around the 88 MPH mark and high, with two clusters of fastballs, suggesting a four seamer and a cutter or moving fastball (the MLBAM tags have a little of both, and he threw both in Japan). I don't want to focus too much on velocity in these posts, because that is something someone can easily find elsewhere, but I thought it would be interesting to compare his velocity in Japan according to the data in my original scouting report and his velocity according to the general MLBAM release point (for these posts, I will be using the modified Brooks Baseball data, but this part will be the general pitch velocities you can get on FanGraphs and other places). So here is the difference, with the plus or minus compared to what his pitch velocity numbers were in Japan.

FF: + .8
SL: + 2.1
FT: + 2.1
Splitter: + 1.7
CU: 3.54

So as I predicted, thanks to better pitch tags, and a cold gun in Japan (and perhaps some adrenaline since it was his first start), he is throwing harder in the Majors.

When it comes to pitch movement, it is hard to see clear classifications, so speed and spin is probably better to classify pitches for him, or at least easier:

 But what about location? We saw that it was important when it came to the at-bats against Cabrera, so how did he generally locate his pitches? This first graph is all of his pitches, labelled by whether they were strikes (s), balls (b), or put in play (x).

 The biggest cluster seems to be mid strike zone in height and on the arm side, with a couple of other major clusters, such as low in the zone and arm side, and mid zone and glove side.

Breaking down the average locations by pitch type (with the average release point for reference) looks like this (using MLBAM tags, I combined the non four seam fastballs into one pitch):

I think his slider location left something to be desired, as it stayed up more than his other pitches and didn't get glove side. His splitter stayed low, as did his moving fastballs, and his curve was a glove side curve (though it stayed high as well). When you look at his average locations by pitch result, it looks something like that:

Strangely, his foul balls were the highest pitches on average, but not surprisingly, his whiffs were the lowest. Notice how everything is arm side. There also isn't much difference between his called strikes and his pitches that resulted in contact. Here are all 19 pitches that were put in play by Blue Jays hitters:

Most are really arm side, and the home run was the furthest up, but he got a lot of grounders on pitches low and arm side (though he got two groundouts just slightly lower than the home run).

I plan to have a post like this after each of Tanaka's starts, and if you have any ideas at what I should look at each start, then let me know in the comments or on twitter.

Yuki Matsui: Start One

Yuki Matsui, the first pick in the 2013 NPB draft, made his NPB debut for the Rakuten Eagles on April 2nd against the Orix Buffaloes, pitching six innings, giving up three runs, striking out six and walking five over 112 pitches. Below is a short summation of his start using graphs and charts.

First, let's take a look at the velocity of his pitches by his pitch count (note, there a few blanks because the velocity of the pitch wasn't posted, as all data was copied from Yahoo, though all the charts are mine).

It seems that Matsui's velocity peaked early, and never returned to that peak, though he still threw a few pitches over 140 kilometers at the end of his outing. It also appears he used more slower breaking balls late. The next graph shows how his fastball velocity fared throughout the outing:

 Again, we see that his velocity was the best after his first few fastballs. However, there wasn't a real sharp decline, and stayed steady from the middle of the game to the late part of the game. Here are the average velocities of the four different pitch types he used:

As expected, Matsui isn't a hard thrower, though that level of velocity from the left-hand side in the NPB should play pretty well. His slider was faster than his changeup on average, and his curve and slider velocity difference is pretty normal. The following is his usage breakdown, showing how much he used each pitch.

His fastball usage is about normal (at least by American standards), and the change more than the slider usage could be explained by the right-handed heavy lineup of the Buffaloes. Here is how his pitch results broke down:

A less than 60% strike rate is not something to be desired, and his whiff and called strike percentages were a tad low, and the percentage of pitches he threw in the strike zone was a little low as well:

So we should dig a little more into location, and the following pie chart shows where each pitch was generally located.

As you can see, he was very heavily glove side, working across his body the majority of the time. As the next chart shows, this hurt him when it came to hits.

He was actually more likely to get whiffs when he stayed arm side.

One reason that he may have been working so glove side is the way he releases the ball:

As you can see, he is leaning quite a bit to the right when he lets go of the ball. This may cause him to crossfire, and throw in to right-handed hitters, something you usually don't want to do when you are not throwing especially hard.

How Much Does Rakuten Believe Kazuya Fujita's Defense is Worth?

The Rakuten Golden Eagles have re-signed (he did not exercise his domestic free agent option, so instead of just agreeing to a contract for the year, they extended him) starting 2nd baseman Kazuya Fujita to a 3 year deal worth 240 million yen, or about 80 million yen a year. Fujita is mainly known for his defense, and in my opinion, is the best defensive 2nd basemen in Japan, and possibly the best defender in the NPB as a whole, regardless of position. However, the NPB doesn't have the advanced defensive metrics (that are flawed as well) we easily access when looking at MLB players. The only thing close is range factor, which is crude, but does like Fujita a lot. In this post, I wanted to take a stab at estimating how valuable, in terms of money, Rakuten thinks Fujita's defense is.

The first step in doing this is to estimate how valuable the other parts of his game are, which should be easier. Defense is the toughest to quantify statistically, so we should look at how valuable his offense is, and assume that Rakuten, using simple metrics, value his offense about the same. In 2013, Fujita was worth -9.84 runs offensively against average but 11.21 runs against a fictional replacement level of .32 winning percentage. For the rest of this post, I'll be using average as the baseline instead of replacement, because it is easier, and a little less confusing when it comes to position players, as all three aspects of the game will be weighed against (Pacific) League average. Fujita's offense has been declining, as these are his adjusted OPS's over the last three seasons against league average:

112 OPS +, 100 OPS +, 91 OPS +
 
I think it is safe to say that Fujita, at age 31, will probably keep declining, or at least hold serve. A safe estimate for the three year contract would be Fujita being 10-15% worse than league average when it comes to batting, probably around 10-12 runs below average over full seasons. With positional adjustment, we can convert that to 8-10 runs a year.

Fujita is also worth nothing when it comes to the basepaths, with a grand total of 22 steals in his entire career. Out of the 68 qualified batters in the Pacific League in 2013, his 3 steals were tied for 33rd best in the league. He's not quick, but he isn't especially a baseclogger either, so we will conservatively estimate that he is worth 1-2 runs less than average per year.

So if Fujita was an average defender, we would estimate that he would be 9 to 12 runs worse than league average over a full season. A MLB comp might be someone like Jose Altuve, someone who plays 2nd with average defense (according to DRS), with below average offensive value and not much baserunning value. However, no one thinks Fujita is an average defender, and that is the purpose of this post, as we are attempting to see how much Rakuten thinks the defense is worth. However, there is another step that we haven't gotten to yet, and it may be the toughest. While we have an estimate of how much Fujita is worth, without defense, when it comes to wins and losses, we haven't converted that into money yet. In the MLB, a win is generally considered to be worth about 5 million dollars, or perhaps 6-7, depending on who you ask. However, we need to convert this into NPB salaries (and adjust for currency as well). So we need to look at how much NPB teams are paying for players, and more specifically, runs and wins.

According to the salary data collected by Yakyu Baka, the average NPB player made about 36.55 million yen, or about 365,000 American dollars (for the purpose of this post, we are assuming that 100 yen is 1 dollar, even though it is a little more than that. However, I'll keep most of the discussion to yen). So Fujita will make over double what the average player will make, but it seems far too simplistic, especially not taking in account things like service time and leverage to say that Rakuten values him twice as much as they would a league average player.

First, I wanted to look at the domestic free agents signed in 2013, and see if their 2013 WAAs had any predictive value when it came to what they will be paid in 2014. I just used a simple wins above average using runs created offensively, and just ERA WAA (split up between leagues) for pitchers.

The first problem we encounter is that there are only 8, which is really too small of a sample size to make any conclusions. Secondly, the worst player according to the simple WAA is the most expensive player, paid mainly for his past success. For these reasons, it seems tough to use those free agent contracts to project value.

So it may be more helpful to look at how Rakuten values wins and dollars, or at least how much they paid in 2013 for players versus the output they got. I limited the look at 7 individual players, the 5 foreign players and the two players on the team that were former MLB players. These players would most simulate "open market signings", which we are assuming Fujita is.

So Rakuten paid 622 million yen for the 7 players for 4.9 Wins Above Average or roughly 126 million yen per win over average. It should be noted that Rakuten has offered Masahiro Tanaka a 800 million yen a year contract to stay with the team after a 5.4 WAR season. This is a little over the rate of the other players, but at 148 million yen per win, isn't too far off (and it is safe to say that Tanaka has more leverage than Fujita had and Tanaka holds more sentimental value to Rakuten).

So with the 8 free agent signings in the NPB this year having a wins above average of just .51 total, and about 1068 million yen dolled out for the 8 players, Rakuten was much more efficient with their "foreign players" than the teams that dipped in the domestic free agent pool in 2013 (assuming these players don't perform much better in 2014). This should be kept in mind when discussing Fujita's contract, as it seems like NPB free agents were well overpaid.

If we keep the guesstimates about Fujita's non defensive value that we made above, and we assume that Rakuten is willing to match their 126 million yen per win above average, then Rakuten values Fujita's defense as worth anywhere from 15-18 runs per season, or about 189 million yen to 227 million yen per year (without subtracting what an average player would get anyway).