Matt Adams is Beating the Shift

Mike Petriello had a nice article yesterday describing what the Reds observed first-hand when they faced the Cardinals: Matt Adams has been beating the shift.  He's mostly doing this by taking the ball to left field, especially on outside pitches:

But just about no one really bunts against the shift, especially not power hitters like Adams. (So far as I can tell, he has never successfully done so in the regular season, though he has tried in spring training.) Instead, he’s just taken advantage of it in a simpler way. In all of 2013, he had 17 hits to the left side of the field. In 2014, in less than 10 percent of the season, he already has 11.

Interestingly, it seems to be because he's offering more at pitches on the outer half of the plate:

All these minor differences begin to add up. Adams is seeing a higher percentage of pitches on the outside of the plate, he’s swinging at more of them, and he’s making contact with more of those — and overall, he’s got a .425 BABIP across all pitches. So maybe this is small sample size noise, but considering how close he is to matching his 2013 opposite field hit total, maybe there’s some amount of signal to it. Adams probably can’t exert a huge amount of control over where the ball goes — this isn’t billiards — but he can exercise some choice on which balls he chooses to swing at. Understandably, more outside pitches lead to more balls the other way. (And raises questions about whether teams choosing to shift should be more in tune with their pitchers about where their pitches go.)

I tried to visualize this by going to Mike's preferred site, Baseball Savant, and looking at spray charts.  Below, I'm reporting all ground balls and line drives by Adams in 2013 (left) and 2014 (left).  Green squares are singles, purple squares a ground-outs, blue-green triangles are doubles, green triangles are line-outs.

...I dunno.  I think this is a case where I'd rather just have the total event numbers than try to extract it from this.  But if you compare what's happening in left field, I'd say that Adams has roughly half as many left field events as he did in 2013.  And far less than half of the events to right field.  I'm going to have to look around and see if I can find a better source for this info that will compile the numbers better.

In any case, I think the notion is interesting here.  I love that the Reds are going to invoke the shift on batters this year.  But if you're going to do it, you have to be aware of how you're planning to pitch the batter.  If the pitching approach is to pound a hitter outside, then it might not make sense not to do a dramatic shift.  Not a revolutionary idea, but one worth keeping in mind.

.....

Update: I'd forgotten this, but FanGraphs actually provides a Pull/Center/Opposite split.  Here are Matt Adams' career splits:

Direction	AB	H	%BIP	wOBA
To Left	78	32	24%	0.418
To Center	118	39	36%	0.353
To Right	131	55	40%	0.547

And Adams' 2014 performance:

Direction	AB	H	%BIP	wOBA
To Left	19	12	39%	0.643
To Center	9	1	18%	0.099
To Right	21	8	43%	0.434

So, we're seeing Adams hitting many more balls to left field.   But it's worth keeping in mind the sample size here; we're dealing with a total of 50 PA's, here.  The difference between this year's pattern and his career pattern is about 10 balls hit to right field.  That's not irrelevant, but we might not want to declare him to be a radically new hitter based on 10 balls hit to right field.

Also, just for comparison, here's Jay Bruce's career numbers:

Direction	AB	H	%BIP	wOBA
To Left	520	157	23%	0.345
To Center	720	235	32%	0.370
To Right	1012	395	45%	0.508

Bruce looks to be even more of a severe pull hitter than Adams--though Adams was at about these numbers prior to the season.  It's no wonder that pretty much every team plays him to shift.

Why Steve Smith should keep being aggressive

When a runner is tagged out at home to make the first out
in an inning, everyone blames the third base coach.
But he might have made the right decision, and just
lost the bet.
Photo credit: Keith Allison

In the aftermath of today's loss to the Rays, Reds third base coach Steve Smith beat himself up over his decision to send Joey Votto home with no outs.  Votto was tagged out, and ultimately the team lost 1-0.  His quote:

"I can't get a guy thrown out with no outs, and I understand that," Smith said. "I tried to be too aggressive at the wrong time. I got burned. ... I've got to play the game. I got aggressive. If the ball bounces one way or the other, then everybody doesn't know about it. It's no outs. I always assume a guy makes a great throw, and he did. I take full credit for that."

I think Steve Smith isn't looking at this the right way.  The fact that Votto was thrown out does NOT mean that he made the wrong decision in that spot.  He may have just lost on what was a good bet.

A few years ago, Pizza Cutter wrote an article titled "Why All Third-Base Coaches Should Be Fired."  In it, he noted that, based on success rates, third-base coaches seem to be far too conservative.  He noted that in sacrifice fly situations, runners that attempted to run home were successful 96% of the time.  That's despite the enormous gains that a team receives when a runner scores.

Let's try to repeat Pizza's math for the situation with Votto.  Pizza didn't explain his math, but I think the following is right.  We have three possible outcomes:

1. What happened: Joey goes for it, gets thrown out.  That leaves the Reds with a runner on second (Phillips) and one out.  The average run expectancy in that case is 0.637 runs
2. Joey goes for it, but scores!  That leads to 1 run (Votto's), plus Phillips went on the throw and thus is now at second base with no outs.  That gives us an extra run expectancy of 1.0499, for a total of 2.0499 runs.
3. Joey holds at third on the single, giving the Reds 1st & 3rd with no one out.  Average run expectancy there is 1.6423 runs.

So, we want to find the break-even percentage of success at which point Votto should be sent.  That percentage, x, would make the average value of sending the runner equivalent to having him stay at third.  Therefore, any success rate over that percentage means you should send him, even if there is some chance that he is thrown out.  (warning, a little algebra ahead...you can skip if you'd prefer to do so)

I solved this equation for x: 

Runs_Hold = Runs_Success*x + Runs_Failure*(1-x) 

Which would be:
1.6423 = 2.0499*x + .637*(1-x) 

So:
1.6423 = 2.0499*x + 0.637 - 0.637x 

...simplying...
1.0053 = 1.4129x
x = 0.71

What this means is that if Steve Smith is more than 71% certain--let's say 75%--that Votto will score on that play, he should send him.  Even though there were no outs in the inning.  Yes, one out of four times, he'll get burned--just as he did tonight.  But he should be playing the percentages.  The Reds stand to gain more over the long haul by sending runners in that situation whenever the coach thinks they will succeed three out of four times.  The one in four chance might be that "great throw" that Smith cites in his statement.

It's probably worth mentioning that this play comes only days after Billy Hamilton's amazing feat of scoring from third on a pop fly barely past the second baseman's position.  Smith presumably was in on that, and looked like a genius.  This time, he got burned.  But I hope he continues to be aggressive on sending runners home in the future.

Why outfielders should throw to the cutoff man, part II

Last sunday, I highlighted a neat article by David Kagan at the Hardball Times that demonstrated the importance of understanding ballistic flight paths when explaining why outfielder should hit the cutoff man.  In my post, I also noted that accuracy on shorter throws by the infielder is also likely important, a point that +Joel Luckhaupt expanded upon when noting that the outfielder's throw is also more accurate when traveling shorter distance.

I thought I'd do some back-of-the-envelope calculations (seriously, I found a spare envelope and sketched this out) to illustrate the point.  Let's assume that any given fielder has a certain throwing accuracy, which we can measure in degrees.  To visualize this, put your nose up to the circle at the center of this protractor:

Our target is at 90 degrees.  If a player's accuracy is + 2 degrees, that would mean that, under stressful, rushed conditions, that player would be able to routinely make a baseball travel within a "window" that is between 88 and 92 degrees.  It's a very small window, but Major League Baseball players are pretty amazing people.

The longer the distance traveled, however, the more a minor angular error in the trajectory he fires a baseball  will result in a ball sailing away from his target.  If we take an outfielder who is 270 feet from home plate, lobbing a baseball at a runner heading home, small angular errors can result in the ball missing the plate by quite a bit:

At 270 Feet
Angular Error (Plus/minus Degrees)	Misses Home Plate By (Feet)
0	0.0
0.5	2.4
1	4.7
1.5	7.1
2	9.4
2.5	11.8
3	14.2

This is where the cutoff man comes into play.  Let's assume (as David did in his article) that the cutoff man is 90 feet from home plate, rather than 270 feet.  Let's further assume that the average MLB fielder's throwing accuracy, under pressure, is +2 degrees.  This is a guess, but it seems sort of reasonable.  Here's what happens (graphic is showing a blimp's-eye view of the field):

The infielder and outfielder have the same throwing accuracy in my diagram: both are +2 degrees.  But the infielder is so much closer than his throws don't have the opportunity to deviate from their target as much.  With some simple trig, we can estimate that the infielder's throws all come in within 3 feet of the plate (roughly the length of a player's arm + glove), whereas the outfielder's throws might come in up to nine feet from the center of home plate.  In other words, the outfielder's throws will be three times as wild as the infielder's, simply because he is three times as far away.

Therefore, not only is hitting the cutoff man allowing a more direct route (lower launch angles negate the time it takes the infielder to catch, transfer, and throw again), and not only does doing so help prevent the hitter from advancing to second base on a throw home, but it also means that the ball will usually arrive in a better location for the catcher to make a play on the runner than if it was thrown all the way from the outfield.

Why outfielders should throw to the cutoff man

A Cardinal Sin of outfielding is to throw home directly, rather than bypassing the cutoff man.  I've always wondered, however, if there are times when that really is the optimal decision for an outfielder.  Allowing the infielder to catch, grab, and then throw the ball must take time, and that's all time that the ball could be advancing toward home plate.  I figured air resistance might be negating some of that gain, but I wasn't sure that it would be enough to matter.

Therefore, I really liked David Kagan's article at Hardball Times last week on the physics of hitting the cutoff man.  David's piece shows that, in fact, one of the major advantages of hitting the cutoff man is that it is a straighter path to get to home plate.  When you throw directly home, you have to use a high launch angle because you have to overcome gravity every step along the way.  Throwing to the cutoff man is a more direct route, and the infielder then has a nice, short throw, again with very little need to elevated launch angle. His diagram shows this best:

Once one also includes factors like air resistance that leads to decreasing velocity during flight (blue line), the total flight time of a ball thrown directly home from 270 feet was 2.75 seconds, compared to just 2.39 seconds.

Now, when hitting the cutoff man, the infielder still needs to catch, transfer, and throw the ball, which must take another half-second to a second (this would be a fun thing to measure...I'll be watching for that in future ballgames).  Therefore, if the infielder is major league baseball-level efficient, we're looking at nearly equal travel times.

Consequently, the benefits of hitting the cutoff man seem to be as follows:

1. Nearly equal time to plate, because the gains of bypassing the infielder are mostly offset by higher launch angle, and thus distance traveled, that are required to get the ball to home plate from the outfield.
2. Improved accuracy by infielder on throw, because slight errors in trajectory angle make less of a difference when you're closer.  I'm guessing this is a really big deal.
3. Ability to prevent the batter from advancing to second base.

Postcript: I'd like to formally apologize to my infielders in 8th grade little league when I bypassed the cutoff man and tried to throw out a batter running to third base.  I didn't succeed, but we might have if I'd thrown to the cutoff man!

Bunting for Sluggers and Sprinters

As much fun as it is to see Jay Bruce hit monster
home runs, sometimes he is just as well-served
when he tries for the bunt single.
Photo credit: Trev Stair

The bunt has a potential to be an important weapon for the Reds this season.  Billy Hamilton, the Fastest Man in Uniform, will try to use the bunt to get the ball on the ground so that he can use his legs to steal first base--especially when batting from the left side.  It can be a big part of his game, much as it was for Willy Taveras.  Obviously, the hope is that he can be a better player than Willy Taveras, but there was a time when he was in the running for both the retrosheet bunt hit record as well as the single-season bunt success record.

Taveras was an especially skilled bunter, but Jeff Sullivan did a piece at FanGraphs this week that demonstrated how hard bunting can be.  It is clearly something for which players can have a lot of success, but he found that the average rate at which bunt attempts are even bunted fair is only about 50% of the time.  And that's for "frequent" bunters; other players are down near 46% of the time.  And that's just bunting the ball fair, not successfully getting on base.  If the bunt is going to be useful for Hamilton, it's clearly a skill he'll have to develop (and he hopefully already has).

But with the increased prevalence of the infield shift, other Reds hitters may see reason to use the bunt.  Jeff Sullivan wrote a follow-up looking at how pull hitters, and especially left-handed pull hitters, can use the bunt to beat the shift.  The idea is so tantalizing: if the infield defense is cheating on you, giving up the entire left side to get four infielders on the right side, then you can get an "automatic hit" just by dropping a bunt toward third base.

Of all the Reds hitters, Jay Bruce would seem to be the best candidate for this approach.  And he clearly does so: Sullivan reported that Bruce has made 30 bunt attempts during Sullivan's 2008-2013 window.  Of his 30 attempts, only 9 went fair, which means the other 21 attempts went as strikes (fouls or whiffs).  But on those 9 fair bunts, 5 were successful: that's an 0.556 clip!  The interesting thing, however, is that Sullivan found that even after successful attempts, defenses were sometimes slow to adjust out of the shift...which means, if you can execute, you can continue to get on base with it at a high clip.

(Also, as an aside, the second animated gif in Sullivan's piece of Hank Conger's "successful" bunt hit is required viewing)

Just this morning, David Laurila posted a great quote from Brandon Moss on the advantages of using the bunt against the shift.

“If you’re able to get one down, it works two ways,” explained Moss. “Either it shifts the defense back over and opens up holes on the pull side, where I hit the ball, or they keep doing it and it’s a free hit. I just have to make sure I get it by the pitcher.”

It's classic game theory, and it works for both pull hitters and speedsters: if the defense knows you won't bunt, they don't have to defend against it, which means they can play back and even shift if necessary.  If they know they have to defend against the bunt, however, they will have to think twice about the shift.  And for a guy like Hamilton, they may have to cheat in on the grass, which gives his ground balls an even better chance of sneaking by infielders when he does swing.

Price on bullpen use

Sean Marshall has historically had good success against
both righties and lefties.  Photo Credit: spablab

I thought this was an interesting comment from Bryan Price on bullpen usage:

"I've said to myself I don't know what it's like to be an everyday player professionally," he said. "But I do understand what it's like to be a pitcher and what these guys go through on days they don't pitch. 'Is so-in-so available?' Well, we got him hot three times yesterday and he pitched the two games before.

"It wasn't really a day off. Sometimes, as a manager (who was a position player), you may see that the guy got loose. But they see it as a day off. It's not. It's completely different. If you don't respect that, you end up over-utilizing your guys – even if they pitch in 65 games.
...
"There are going to be times when you bring in a right-hander to get that right-hander out based on the history of the pitcher that's in there at the time. But, to me, it just doesn't bode well when you match up. I don't like situational match-up pitchers because they beat up you bullpen. I've never liked it. I'd much prefer to have guys who get lefties and righties out.

"You never want to lose a game because you feel like you have the wrong matchup in," he said. "But if you want the right matchup every single game, you're going to blow out your bullpen."

It's going to be interesting to see whether his stance evolves over the season.

I think he makes a lot of good points, and important ones.  There's no question that platoon splits are real, and you absolutely do gain when you use righty-on-righty and lefty-on-lefty matchups.  I imagine that there are other kinds of match-ups that are worth exploiting, like using sinkerballers, breaking-ball specialists, or power arms against hitters that are particularly vulnerable to those pitcher types.  But I also think he's right about the importance of considering how often a guy is asked to get up and get warm as part of a reliever's usage load.

This is going to sound ridiculous (it's akin to pontificating about how to fly an airplane based on experience playing a flight simulator video game).  But my favorite baseball "video game" is Out of the Park baseball, where you play the role of the general manager and manager.  In terms of reliever usage, what I've taken to doing when deciding what reliever to bring in is to consider at least the next three batters that they will face.  Mostly, I consider handedness, but there's always some weighting for quality as well.  Ideally, there will be a rhythm to the opposing lineup, such that there is an optimal time to switch between a lefty and righty reliever that is spaced by at least three hitters.  Best case scenario, I can plan several innings in advance.  This lets me play matchups, at least in a coarse sense, but also avoid overtaxing my bullpen by using relievers for only a batter or two at a time...except in emergencies, of course!

I think (hope?) that's what Price is talking about here.  Yes, Sean Marshall will get to throw against righties.  But one should still try, at least generally, to use your top lefty reliever against the opposing team's top lefty hitters.  It's much better than blindly following a "Broxton in the 7th, Marshall in the 8th, Chapman in the 9th" plan.