Harry Pavlidis passes his time as a pitch F/X specialist at The Hardball Times, Beyond the Boxscore and Cubs F/X. He brings today a look at Daniel Bard and his developing slider. Due to his fantastic work, we can now see the two subtle differences between Bard’s evolving slider.

Much ado has been made about Daniel Bard’s fastball and his future as closer for the Red Sox. Management doesn’t think he’s ready – or at least his breaking ball isn’t.

Bard spoke to The Providence Journal about his breaking pitch back in August:

When it’s ineffective, the pitch resembles a curveball, slower and with more loop. It’s easier to whack around, and the lack of contrast makes it easier to sit on his fastball….The pitch was a problem early in the season, but then Bard tried a new grip and aimed lower in the zone….But Bard may have turned another corner with the pitch. In his last few outings, Bard has used a new approach, leading to a tighter, harder breaking pitch, sitting right between 87 and 89 mph….”I’m holding it like a cut fastball, throwing it like a cut fastball, and it comes out like a slider, so whatever works,” Bard said.

What Bard was talking about his working his way around the cutter/curveball continuum. Cutters can blur into fastballs, but they go the other way into slutters, sliders, slurves and, finally, curves. Bard seems to be finding slutter to be a better option than a slurve.

Theo Epstein had this to say about Bard’s readiness to close:

He’s still tweaking his breaking ball. He’s got a good breaking ball but it probably isn’t where it will be eventually. This is somebody who is still really a work in progress.

I won’t attempt to answer the “real” question — is Bard the best option to close in 2010 — but we can look at the breaking ball via PITCHf/x.

Daniel Bard PITCHf/x

Note: I am using my own pitch classifications, not Gameday’s.

Bard is a four-pitch pitcher. In addition to the slider(s) he threw in 2009 he also mixed in a few 90 mph change-ups that moved like his two-seam fastball but 8 mph slower. Bard only threw the change-up to a few left-handed batters. He also was about two fastballs per slider against righties but three-to-one against lefties. Not too surprising, as a lot of righties will avoid throwing a breaking pitch down and in to a left-handed batter. Bard’s primary weapon was the 98 mph four-seam fastball.

Pitch	# Thrown	Avg MPH
Four-seam fastball	537	98
Slider	209	85
Two-seam sinker	79	98
Change-up	16	90

Bard’s slider is slow. Most pitchers run their sliders and change-ups at the same speed. The eight MPH gap from the heat to the change is solid, so it’s the slider speed that stands out to me. Often, sliders are a couple MPH faster than a change-up In effect, he’s throwing it at the speed other hard throwers get on their curveballs. But that’s on average. The real story in Bard’s slider is the variations. Before diving into the slurve to slutter journey, I want to point out something else interesting about his slider, and nearly ironic.

When Bard talked to PoJo, he was about three weeks into a slider phase-out of sorts. In July and early August, Bard threw between 30% and 40% sliders. By the time October rolled around, he was throwing them less than 20% of the time. Meanwhile, his four-seam fastball went from 65% to 70%. When the playoffs rolled around, Bard only used the slider about 1 out of 6 pitches.

Here are a few samples of four-game runs for Bard:

Games	Fastball	Slider
May 13-22	59%	19%
June 5-14	63%	23%
June 28-July 9	56%	39%
August 18 -26	61%	27%
October 2-11	67%	15%

This pattern doesn’t suggest confidence in the slider, on the part of Bard, his catchers and the team. “Hey, let’s try this pitch out some more” became “Hey, these games really count, so ditch it.”

Was the Slider Really a Bad Pitch?

I think it got tagged once in a while, but was otherwise solid.

Note: One thing we’ve discovered via PITCHf/x is a tendency for batters and umpires to play, in effect, to a two-foot plate. I’ll be using that “standard” for any strike zone metrics below.

Here are a few of my favorite stats, just for Bard’s slider and four-seam fastball.

	Fastball	Slider
In Strike Zone	.510	.593
Balls:Called Strikes	2.98	1.46
Swing Rate	.531	.354
Whiff Rate	.277	.446

For a fastball, that’s not a lot of strikes. For a slider, that’s a ton. The whiff rates on both pitches are impressive. The lack of swings in the zone (.427) against the slider go a long way towards the B:CS ratio.

Nothing wrong with the slider yet.

	Fastball	Slider
SLGCON	.581	.500
Ground Ball Rate	.378	.600
Line Drive Rate	.230	.000
HR per FB+LD	.077	.250

Bard gave up five home runs in all of 2009. Four made it into the PITCHf/x data, one was off a slider. Yes, Bard allowed no line drives and just four fly balls against his slider. Mark Teixeira pulled one that sat middle-in 357 feet down the line on August 8. It would’ve been a home run in just seven Major League parks (source: Hit Tracker).

I imagine there are hundreds of pitchers who would live to throw an 85 MPH breaking pitch that finds the zone like Cliff Lee and misses bats like Rich Harden. This Bard fella does, but I guess that’s not good enough to close in Boston.

I calculated, using two flavors of linear weights (hit type vs. batted ball type), rv100s of -2 and -3 for Bard’s slider. That’s a nasty pitch, especially when the -3 (more runs saved, which is good) is based on batted ball types. Bard may start giving up line drives, but less hits should come off his grounders as time goes by.

Maybe the performance isn’t predictive. Maybe the inconsistency in the pitch is enough to scare the Red Sox. Enough to cast doubt on Bard’s ability to continue getting exceptional results? We’re talking about a pitch that saved two (or three) runs per 100 times thrown in 2009. Small samples and everything, but something compelling must be weighing on the front office minds in Boston. One wonders where Theo thinks the slider can be eventually. Statistically, it will be hard to go up from where it is, in some regards.

Back to the Continuum

The concern expressed by the Red Sox is, or was, getting Bard’s slider to be less slurvey. Throwing the ball like a cutter is one thing, but the slider action the hitters experience is really all that matters. Still, based on the technique we’ll say the goal for Bard is to throw something more like a slutter.

The slurve will be the slowest, with the least back-spin and most sweep. That will feel like a sinking pitch to a hitter. The slutter will be the hardest thrown, with the most back-spin, moving the vertical sink/rise closer to the fastball. Less sweeping action will also be evident in a slutter.

This chart shows the speed of a pitch against the angle of spin on the ball. It is one of the best ways to classify and look at pitches, even more than the typical PITCHf/x spin movement graph. While the overlap of the sinker and fastball looks significant in this aggregated view of all of Bard’s pitches, when viewed game-by-game the split is quite clear. In any case, the focus is on the slider.

As you can see, the sliders produced by Bard ranged from 78 to 90 MPH. The variance in spin axis leads to a range of 9 inches lateral, or sweeping, motion caused by spin. The sliders also picked up anywhere from seven inches of sink (as in more than gravity alone) from top spin to 5 inches of “rise” (again against gravity) from back spin. That’s a full foot, for those of you keep score at home.

Some of this variance is attributable to differences in PITCHf/x systems form park-to-park. Variance exists within parks, too, making the problem space a little bigger and murkier. But the variance in the spin axis data is smaller — some of the “movement” issues are related to release points being off, and the spin data controls for that a bit. But there is also a wide range of speeds on the fastball (shown above), so my favorite approach is to gauge slider speed and movement relative to the same game’s fastballs.

This same game/it’s relative approach is especially helpful for Bard’s case. What we can look for are sliders that are closer to the fastball in speed and movement and assume those are the “good” sliders. The further away from the fastball, slower with more sink and sweep, the worse.

I chose a very crude approach to split good slider games from bad slider games. Using spin movement (pfx_x and pfx_z) along with MPH, games where the average slider where close to the average fastball were good, those that were not close were bad. I said it was crude. How crude? I simply summed the absolute values of those three differences and split the games into two groups, leaving out the three where Bard threw no sliders.

Once I had the games grouped into good and bad (only in reference to the slider), well, I was surprised. Someone smarter than me should use a more sophisticated approach, but my crude split worked. “Good” and “Bad” game fastballs didn’t look much different, but the sliders did — they were a little faster with a little less movement.

type	quality	#	MPH	pfx_x	pfx_z
Fastball	bad*	258	98.0	-4.9	10.4
Fastball	good*	271	98.3	-4.5	10.2
Slider	bad	90	83.4	7.1	-1.5
Slider	good	119	85.6	6.0	-0.2

Oddly, the good game sliders were more likely to be strikes.

type	quality	#	B:CS	ISZ
Fastball	bad*	258	3.3	0.523
Fastball	good*	271	2.7	0.498
Slider	bad	90	1.8	0.556
Slider	good	119	1.2	0.622

And the good game sliders missed more bats, yielding more swings to boot (especially chases out of the zone).

type	quality	#	Swing	Whiff	Chase
Fastball	bad*	258	0.531	0.226	0.415
Fastball	good*	271	0.524	0.331	0.360
Slider	bad	90	0.267	0.333	0.150
Slider	good	119	0.420	0.500	0.356

The good game sliders also ended up with better outcomes.

type	quality	#	nkSLG	GB%	FB%	PU%	LD%	HR/FL%
Fastball	bad*	258	0.643	36%	31%	12%	21%	14%
Fastball	good*	271	0.533	37%	30%	7%	27%	0%
Slider	bad	90	0.625	63%	25%	13%	0%	50%
Slider	good	119	0.417	58%	17%	25%	0%	0%

Using rv100 (hits and outs) and rv100E (batted ball types), the good game sliders blew the doors off the bad game sliders.

type	quality	#	rv100	rv100E
Fastball	bad*	258	0.48	-0.31
Fastball	good*	271	-0.71	-0.50
Slider	bad	90	0.00	-0.92
Slider	good	119	-3.56	-4.62

The fastballs have been shown for comparison, and while the “good slider game” fastballs fared a little better than their “bad slider game” counterparts, the difference is subtle compared to the gap between the sliders themselves.

As I said, this was a crude approach to the problem. But it certainly seems to support the Red Sox point of view. It also requires further study. Who’s game?