A first look at the 2008 Pitch f/x databy Josh Kalk
April 01, 2008
With the first baseball game played on U.S. soil in 2008 came the first regular season 2008 Pitch f/x data from Sportvision. This article will examine Tim Hudson's performance Sunday night.
Sportvision's plan is to have every pitch thrown this year tracked by Pitch f/x (except the two games played in Japan). At the end of last year, only two parks were without Pitch f/x cameras; Camden Yards and RFK. The completion of Nationals Park included two new Pitch f/x cameras and they recorded their first regular season data Sunday night.
Along with the tracking every pitch, the hope is that the Pitch f/x system will be better calibrated and there won't be a need to do nasty corrections to the data this year. Time will tell—it will take at least a month's worth of data before we will have a chance to tell if it's improved.
Lastly, Sportvision has added a few new variables, most importantly the type of pitch that was thrown. This pitch identification appears to classify all pitches into fastballs, sinkers, curveballs, sliders, changeups, split-fingered fastballs, cut fastballs and knuckleballs. Fortunately for me, this is exactly the list my classification algorithm used to classify pitches last year. Along with the pitch type, a confidence level is reported, which seems to go from zero being least confident to one being most confident. More data will be needed to determine how we should use this confidence level, so for now I am just going to list the pitch type that Pitch f/x has given us.
On to Hudson's Opening Day. He started a bit shakily, giving up two runs in the first inning to the Nationals, then settled down and didn't allow a base runner for his last six innings. He ended with a no-decision and the Braves lost a heartbreaker in the bottom of the ninth.
Here is a look at his 78 pitches by the movement of the pitch compared to a pitch thrown without spin. Note that negative x means the ball was moving toward a right-handed batter.
Fastball or sinker?
The first thing to notice on this plot is that Pitch f/x has identified all of Hudson's fastballs as true fastballs and not sinkers. Hudson actually throws both a four-seamer (fastball) and two-seamer (sinker) as John Beamer broke down last year. As John points out, the difference between these two pitches is very small and very hard to detect. My previous classification algorithm called all of these pitches sinkers because even the four-seamers produced a lot of "sink." Here, Pitch f/x is labeling all these pitches as regular fastballs and I think that is justified.
On average, these 56 fastballs were thrown at 92 mph with -8.07 inches of x movement and 8.28 inches of z, compared to a ball thrown without spin. Those 8.28 inches of "rise" are hardly what you would expect from a sinker. An average sinker has just under five inches of "rise" and Hudson checked in at 4.5 inches of rise last year. So either Hudson didn't have much sink Sunday or he was purposely throwing four-seamers. In either case, it doesn't look like his velocity suffered at all: His fastball checked in at almost exactly the same speed at it did last year. It will be interesting to watch Hudson's next few starts to see if that usual sink reappears.
Cutter or slider?
The next interesting thing to look at is the group of pitches to the right of Hudson's fastball. It is mostly made up of what Pitch f/x calls sliders, but two cut fastballs and a regular fastball are mixed in. Hudson's slider is very hard and lives right on the edge of being a cut fastball. John identified it as a slider, but my classification algorithm called it a cutter and it appears Pitch f/x is having some trouble making up its mind. The difference between a cutter and a hard slider is very small but from now on I am going to group those two cut fastballs in with the sliders and just call them all sliders.
To look at how Hudson located his pitches, I am going to break them down by the handedness of the batter so you can get a better idea of how Hudson was working Sunday. We will start by looking at the fastballs he threw to right-handed batters.
Again, negative x means the ball was closer to a right-handed batter. It is pretty clear what Hudson's game plan was against the Nationals' right handers: Keep the ball middle away with an occasional fastball busted inside. Absolutely nothing was thrown on the inner part of the plate. This seems to have been a very good plan: Time and time again, the right handers tried to pull the ball, mainly resulting in weak grounders to the right side. The only hit allowed to a right hander was Austin Kearns' single through the right side of the infield that plated the Nationals second run. Maybe in part due to his lack of sink, it appears that Hudson was having some trouble keeping the ball down in the zone, with many pitches belt high or above. For the most part he got away with that though.
To left-handed batters, Hudson again tried mostly for the outer part of the plate. This is no surprise: Many right-handed pitchers work lefties this way. When he did go inside, it was mostly up and in instead of more underneath the hands as he did to the right handers. The lone hit on this chart was by Christian Guzman, who somehow managed to poke the first pitch he saw, a fastball that was off the plate, into right field in front of Jeff Francoeur. Sometimes as a pitcher you throw the pitch you want in the location you want but the hitter manages to get a hit anyway.
I am not going to break Hudson's sliders down by handedness because he threw only one slider to a left hander all night. Unfortunately for Hudson, it really was his only mistake of the night, a slider on the inner part of the plate to Nick Johnson, who promptly doubled to right, scoring the game's first run. As you can see, the theme continued with sliders as Hudson tried to throw all of these away to right handers.
It looks like Hudson didn't have great command with his slider Sunday, with most of the pitches missing badly. Despite this, Hudson didn't walk a batter, which shows how good a pitcher he is. The slider he threw to Johnson was the fourth pitch of the at-bat; he had started Johnson with two fastballs away and then a fastball up and in. He must not have expected Johnson to be looking in and tried to sneak a slider by. This was a good piece of hitting by Johnson, who wasn't fooled on the pitch.
For a second pitch to left handers, Hudson threw four changeups. All were down and away, and he produced two more pulled grounders with the pitch. It is hard to say with such small statistics but it looks like Hudson had pretty good command of his change Sunday and maybe should have thrown a few more to lefties.
It is pretty clear, looking at the data, that Hudson didn't have his best stuff Sunday night. Many sinkerballers say they need regular work to keep their command and this certainly could have been an issue with Hudson at the beginning of the season, pitching during a pretty cool night. Despite that, it appears that Hudson's game plan was right on the mark. Without great sink on his fastball, he still produced 12 groundball outs to six flyball outs. Even with poor command on his slider, he managed not to walk anybody. The Braves didn't get the win Sunday, but they sure got a good performance from Tim Hudson.