A Baseball Weblog

Monday, January 31, 2011

A posting note (not to be confused with a post-it note)

I've been doing a lot of work at Beyond the Box Score since November, and all of the articles I've written over there have also been cross-posted here at DBITL.  From now on, that will not be the case.  This site will become more "Yankee-centric" (even though it kind of already is) while my work at BtB will not have a focus on a particular team.  Though that certainly doesn't mean that I'll post only about the Yankees here and that my work at BtB will only be about other teams.  In fact, I'm working on an article for BtB on A.J. Burnett that should be ready in about a week.  No big deal; just keeping you up to date.       

Tuesday, January 25, 2011

Hanson's drop in strikeouts

Tommy Hanson had a successful second season in the big leagues last year, pitching 202 2/3 innings with a 3.33 ERA, 3.31 FIP, and 4.04 xFIP.  His contributions were good for 4.3 fWAR.  Throughout his minor league career, Hanson had been an extreme strikeout pitcher (10.7 K/9 in 389 innings); his 2010 rate of 7.68 was better than the league average, but was nothing particularly spectacular based on his minor league track record.  Looking more closely at his 2010 season, you can see that his ability to strike out hitters sharply declined by the end of the year.  The table below shows some splits for Hanson by month in 2010 (September stats include one October start):




ERAFIPxFIPK/9BB/9Whiff RateGB Rate
April2.172.943.5010.243.10.242.417
May5.094.154.697.903.31.224.356
June6.312.843.779.122.81.213.432
July2.432.473.469.711.82.240.482
August3.054.074.645.172.35.174.413
September 1.813.033.895.841.81.163.489


At the end of the year, Hanson experienced a sharp drop in his strikeout rate; he remained effective due to improved control.  Taking a closer look at his pitches gives us a slightly better idea of what changed for Hanson last year.  His repertoire consists of a fastball, a slider, a curveball, and a changeup.  I'm calling his fastball a four-seamer for now, but there are probably some two-seamers as well.  Hanson threw a fastball or a slider last year nearly 85% of the time, so I'll focus primarily on those two of Hanson's offerings.  Splitting up Hanson's whiff rate by the two pitches gives us a bit more insight as to what happened:


FF Whiff RateSL Whiff Rate
April.200.250
May.160.270
June.106.391
July.175.309
September.100.313
October .073.290

While the slider became more effective from the start of the season, the fastball's whiff rate was half as high in August and September as it was in April and May.  So right there, you have some explanation for the decrease in strikeouts - Hanson became unable to rely on his fastball for garnering swings and misses.  But why?  A possible explanation is that Hanson lost velocity as the season went on:


The error bars represent one standard deviation. 



I saw no detectable difference in pitch location (fastballs higher in the zone tend to be harder to hit) or movement, so I think it's quite possible that the loss of velocity caused more contact.  One last think I'd like to see is whether or not Hanson's pitch selection with two strikes changed as it became harder for him to put hitters away with fastballs.  There was a very slight change:  







MonthFF% with 2 strikes 
April.558
May.549
June.463
July.504
August .488
September .482


Basically, I can draw one conclusion from the data: a sharp increase in fastball contact, probably at least in part caused by decreased velocity, made it harder for Hanson to put hitters away, leading to more balls in play. 

Thursday, January 13, 2011

Rafael Soriano?

Well, I'm surprised.  According to Jon Heyman, the Yankees have signed Rafael Soriano to a three-year contract worth $35 million.  Yeah, $35 million.  Surely there will be better contract analyses of the situation, but I'm just thinking that even if you assume Soriano provides 2 WAR over the next three seasons (which is optimistic virtually unheard of for a setup man), it's still over-paying (assume ~$4.5 million per win).  Heyman adds that there's an opt-out clause after the first and second seasons, so conceivably he could look for a closing gig elsewhere if he pitches well in 2011.  


I'm also surprised because the Yankees already have two capable setup men in Joba Chamberlain and David Robertson.  And also because Brian Cashman recently said that the team would not forfeit their top draft pick, which they'll now have to do given Soriano's Type A status.  As good a reliever as Soriano is, I don't like this move right now.  With two rotation slots open and Andy Pettitte not pitching next year, this would be a perfect time to slot Chamberlain as the #4 starter, but I highly doubt that the Yankees will do that, given what Cashman has said this offseason.  Then again, Cashman also said they weren't going to give up their draft pick.  Nonetheless, I'm not getting my hopes up.  

Win distributions for innings pitched and ERA

This article by Dave Allen got me thinking about something really simple.  If you took all games by starters over the past ten years and grouped them by a particular number of innings pitched, how many wins do they end up with?  It gave me an excuse to do a distribution:




Since 2001, there has only been one pitcher to win 20 games in fewer than 200 innings.  That was Pedro Martinez in 2002, and he just barely made it (199 1/3 innings).  The pitchers that won the fewest games with at least 200 innings pitched were Tanyon Sturtze in 2002 and Wayne Franklin in 2004, who both won only 4 games.  


In addition, here is the win distribution for the specific criteria that Allen mentioned in the article, 200 innings pitched with an ERA under 4:




Like Allen mentioned, the average pitcher that meets these requirements wins an average of 15 games.  Randy Johnson won 24 in 2002, which was the same year that both Curt Schilling and Barry Zito won 23.  At the "unlucky" end of the spectrum are Brandon Webb in 2004 and Matt Cain in 2007; they both had only 7 wins.

Monday, January 10, 2011

Venters' rookie campaign

The Braves' bullpen was definitely one of the strongest in the league last year.  Their collective ERA was second to the Padres and Giants, and they trailed only the Padres in FIP, xFIP, and wins above replacement.  Veterans Billy Wagner and Takashi Saito were a huge part of this, as they both posted ERA/FIP/xFIPs under 3 over a combined 123 1/3 innings of work.  Another dominant reliever out of the Braves' bullpen last year was rookie Jonny Venters, who made his major league debut against the Rockies during Ubaldo Jimenez's no-hitter on April 18th.  The 25-year-old southpaw threw 83 relief innings last year and turned in 1.7 WAR, on par with elite closers Mariano Rivera, Neftali Feliz, and Rafael Soriano.  It would appear that Bobby Cox gained confidence in Venters throughout the year as he let him pitch in more high-leverage situations as the season went on.  The graph below illustrates this in the form of cumulative player leverage index (in which a "neutral" situation is 1).




The next question would be: what did Cox like so much about Venters?  Before I look at his stuff with PITCHf/x, I will present a table that shows Venters' 2010 performance and, in the form of a percentile, how it related to other pitchers with a minimum of 1,000 pitches thrown.  





PerformancePercentile 
ERA*1.9597
FIP*2.6995
xFIP*3.1095
K/9*10.0895
BB/9*4.2314
GB Rate**.678100
Whiff Rate**.359100
Chase Rate**.35295
RV/100**-1.5095
xRV/100**-1.1093

*BIS data
**Gameday data

Other than suspect control, Venters has the qualities of an excellent relief pitcher, producing a ton of whiffs, swings out of the strikezone, strikeouts, and groundballs.  Venters led the majors in groundball rate and was second in whiff rate to Carlos Marmol.  Venters' repertoire is conducive to excelling in both of these categories, as last year he primarily used a mid-90s sinker and a mid-80s slider with great movement.  Venters also throws a four-seam fastball; he showed a changeup last year, but ditched it in May.  The charts below show the spin deflection from the catcher's perspective for each pitch and the velocity distribution for each pitch.









Venters' sinker drew rave reviews from teammate Billy Wagner, who said, "There's never been anybody in the game that's had a left-handed sinker like that.  I mean nobody."  The pitch had a pfx_z (vertical spin deflection) value a bit greater than 3 inches, which means that it dropped about a half a foot more than the typical fastball.  And since it sits 94-95, there's not a whole lot of time for hitters to react to pitch's break.  


Next up are some results for Venters' pitches, excluding the changeup that was only thrown 13 times.  It's awesome if you know what all of these metrics mean, and if you don't, that's okay too - use this old post as your glossary.  


#%Swing RateWhiff RateZone RateChase RateWatch RateRV/100xRV/100 
SI778.581.478.293.355.365.315-0.49-0.67
FF242.181.372.233.393.211.379-1.880.43
SL292.218.428.656.332.431.577-4.05-4.08

1332
.445.359.354.352.384-1.50-1.10
    


GB RateFB RateLD RatePU Rate
SI.748.112.119.021
FF.447.237.237.079
SL.667.190.143.000

.678.146.146.029

A few things in particular caught my eye.  Nothing is in the strikezone a whole lot, not even the four-seamer.  Overall, the four-seamer looks like his weakest offering, and it still generates a whiff rate well above average for its pitch type.  The sinker and the slider are absolutely filthy.  Unsurprisingly, given the metrics I showed on Venters earlier, the sinker and slider are both spectacular at getting grounders and whiffs.  The sinker was hit into play 143 times last year, and three-quarters of the time, it endangered a worm.  When swung at, it was missed nearly 30% of the time, which is well over the league average for sinkers/two-seamers.  In a much more limited sample of 38 balls in play, the slider induced a grounder two-thirds of the time, which is still spectacular.  However, the thing that most caught my eye was the whiff rate on Venters' slider.  Over the 125 swings that batters took against Venters' slider, they came up empty 82 times, which makes for an insane .656 whiff rate.  Anything over 50% is elite, and I can only think of a few pitches that are in the category of Venters' slider.  Ryan Madson's changeup was just about as nasty, garnering a .645 whiff rate over 138 swings; last year, Brandon League's splitter posted a mark of .683 over 183 swings.  The high watch rate would mean that on the off-chance that the slider is in the strike zone, batters take it for a strike more often than not.  The slider looks like an elite pitch in every way.  


Before I move on, I'll give you a "bonus graph," if you will.  It shows the plate locations, in feet and relative to the batter's strikezone, for Venters' sinker and four-seamer (the lines have been smoothed to better accentuate the trends).  As you can see, the sinker usually crosses the plate at the bottom of the zone, while the four-seamer is more often than in the middle third.



The next thing I'd like to look at is Venters' pitch selection based on count leverage.  He likes to throw the sinker a lot in any count; the only time he'll throw a lot of four-seamers are in 2-0 and 3-0 counts.  With two strikes, you'll probably see either a sinker or a slider.  The chart shows the pitch selection based on each count, and the tables break it down by batter handedness. 


vs LHBFirstAheadBehindFull
SI.251.447.352.450
FF.272.152.455.300
SL.228.385.182.250

vs RHBFirstAheadBehindFull
SI.283.588.724.737
FF.367.111.148.132
SL.154.272.089.132
CH.000.023.019.000

Hopefully this has given you a pretty good idea of what Venters has to offer.  He'll usually go after hitters with a viscous sinker in the mid-90s, and if he chooses to, he can also put them away with one of the best sliders in the majors.  When hitters aren't missing against him, they're pounding the ball into the ground.  His control isn't great, but he makes up for it by keeping balls out of play and limiting home runs.  Overall, he looks like an elite reliever, and he will be given the chance to compete (along with Craig Kimbrel, whom I briefly profiled last week) for the Braves' closing job this spring. 

Thursday, January 6, 2011

Yankee bunters

I've been all into this bunting stuff for the past month or so.  I think it's fair to say that I've learned some things, though not as much as I may like.  Briefly, I would like to examine some Yankee hitters with the bunting metrics I've been using.  Data are from 2008-2010; use this as your glossary.



Attempt%Bunt AttemptsFair%Hit%Out%Sac%Bunt RunsBunt Runs / 100
Brett Gardner.05273.507.243.378.378-1.731.16
Derek Jeter.01345.489.364.091.5451.917.76
Curtis Granderson.01441.463.211.368.421-1.290.37
Ramiro Pena.05631.323.400.100.5000.254.32
Francisco Cervelli.04731.613.158.211.632-0.601.61
Nick Swisher.00721.619.231.231.5380.063.83


Watching Yankee games on television, I've always assumed that Gardner is a below average bunter, or at least that he's below average at getting the ball in play.  Turns out that he's right at the league average on this front, and he also turns a higher amount of fair bunts into hits than does the average player.  His speed helps with that.  Jeter and Pena look like they're good at avoiding bunt outs, though Pena hits way too many foul bunts.  Cervelli is great at getting the ball in play, and is rarely asked to do more than sacrifice.  Swisher bunts extremely infrequently and has also shown the ability to get the ball in play.  Granderson looks like an average bunter all around.

Tuesday, January 4, 2011

Laying it down, part 2

Last week, I used some different metrics to quantify bunting performance in 2010.  In this post, I would like to build on what I found after researching the first part of this post.  The clear issue with using data from only one year is that it becomes susceptible to small sample sizes, and since bunt attempts occur on such a small percentage of swings, that issue is magnified in this case.  In order to get a better feel for bunting skill, I've set up the statistics I used last week for 2008 and 2009 data.  Before I get into go on, I'd like to revisit the league averages that I showed at the beginning of last week's post. 




Attempt%Fair%Foul%Missed%
2008.020.501.425.074
2009.020.509.421.070
2010.019.505.416.080

.019.502.423.075



Hit%Out%Sac%Double Play%
2008.189.285.515.011
2009.177.288.528.007
2010.188.286.517.009

.184.286.520.009

The league-wide samples stay pretty consistent year to year, which is good for establishing baselines for these metrics.  However, the samples that I used in last week's post (20 bunt attempts or 10 fair bunts) did not correlate very well year-to-year.  Notable R^2 are as follows: hit% - .514, out% - .226, sac% - .528, Bunt Runs/100 - .178.  The one that surprised me the most, however, was coefficient of determination for fair bunt%.  That was only .065, which I found strange considering fair bunt% appears to be a distinct skill.  Bear in mind that even though we're looking at three years' worth of data, it's still a small sample size.  In order to get players that were qualified bunters in back-to-back years, I had to eliminate all but 82 of the individual seasons.  I was particularly puzzled by the low correlation for fair bunts, which I figured would be a detectable skill even in the limited sample.  


Since we appear to have some of the SSS blues, I think the best way for us to cheer ourselves up is to open up the leaderboards to include 2008 and 2009 data.  Well, maybe that's just me.  Anyway, this section is going to contain a lot of tables - I'll have leaders and trailers for all of the metrics I discussed last week, along with some commentary when I feel it is necessary.  Also, the minimums are now at 50 bunt attempts and 25 fair bunts.  I'll begin with attempt percentage, for which I won't show any trailers (there 55 qualified players who haven't attempted a bunt in the three years).  Oh, also, the swing minimum for attempt% is 1,000. 

RankNameSwing#Attempt%
1Willy Taveras1625.118
2Carlos Gomez2178.114
3Julio Borbon1049.104
4Emilio Bonifacio1666.093
5Erick Aybar2420.092
6Nyjer Morgan2181.091
7Juan Pierre2386.086
8Gregor Blanco1293.081
9Alexi Casilla1290.078
10Luis Castillo1652.076
   
My love for the fair bunt% statistic is slightly diminished after seeing how poorly it correlated in my data, but I still think that it's an important statistic to look at.  Below are the 10 leaders and trailers for fair bunt%; Chris Young (of the Diamondbacks) is really in a league of his own. 

RankNameAttemptsFair Bunt%
1Ramon Santiago65.769
2Joey Gathright78.692
3David Eckstein78.679
4Clayton Kershaw62.677
5Zach Duke64.672
6Scott Podsednik68.662
7Cole Hamels53.660
8Franklin Gutierrez51.627
9Ryan Dempster80.625
10Elvis Andrus103.621




1xChris Young51.157
2xRajai Davis78.295
3xCorey Hart69.333
4xOmar Infante55.364
5xMelky Cabrera65.369
6xB.J. Upton73.370
7xEmilio Bonifacio155.381
8xEmmanuel Burriss70.386
9xBrendan Ryan75.387
10xTed Lilly66.394

The next set of leaderboards are for hit%, out%, and sac% out of fair bunts.   

RankNameFair BuntsHit%
1Ichiro Suzuki37.568
2Lastings Milledge25.480
3Jacoby Ellsbury38.474
4Gregor Blanco60.450
5Angel Pagan36.444
6Jose Reyes40.425
7Rafael Furcal48.417
8Alexi Casilla60.417
9Michael Bourn99.404
10Carlos Gomez117.385


RankNameFair BuntsOut%
1Josh Anderson26.615
2Emilio Bonifacio59.525
3Joey Gathright54.519
4Juan Pierre107.505
5Tony Gwynn39.487
6Rafael Furcal48.479
7Carlos Gomez117.479
8Emmanuel Burriss27.444
9Nyjer Morgan100.440
10Willy Taveras107.439


RankNameFair BuntsSac%
1Braden Looper25.960
2Jamie Moyer29.931
3Ricky Nolasco28.893
4Roy Oswalt34.882
5Hiroki Kuroda30.867
6Barry Zito35.857
7Ryan Dempster50.840
8Clayton Kershaw42.833
9Livan Hernandez28.821
10Derek Lowe37.811

Unsurprisingly, all of the sacrifice leaders are pitchers.  The first position players to appear on the list are Daric Barton (.760), Yuniesky Betancourt (.724), and Jamey Carroll (.720).

Like I did last week, I will end with a glance at the best overall bunters with linear weights - this includes a weighting of their hits, sacrifices, and outs, and also takes into account missed and foul bunts.  Again, I will present in a counting form and in the form of bunting runs / 100 pitches.  However, in order to (hopefully) make it more intuitive, the rate stat will be scaled to the league average bunt as opposed to the league average event.  Over the past three seasons, the average bunt has been worth -3.53 runs per 100 pitches, so that will be what I consider "average," or 0.  Onto the best and worst bunters of the past three years:

RankNameBunt Runs
1Ichiro Suzuki5.41
2Alexi Casilla4.67
3Gregor Blanco4.32
4Jacoby Ellsbury3.87
5Angel Pagan3.77
6Michael Bourn2.70
7Rafael Furcal2.16
8Erick Aybar1.98
9Cliff Pennington1.94
10Gerald Laird1.84



1xJuan Pierre-9.54
2xRyan Dempster-6.57
3xUbaldo Jimenez-6.42
4xBronson Arroyo-6.08
5xZach Duke-5.88
6xMike Pelfrey-5.69
7xTed Lilly-5.54
8xDerek Lowe-5.44
9xOrlando Hudson-5.35
10xChad Billingsley-5.00


RankNameBunt Runs / 100
1Ichiro Suzuki9.97
2Angel Pagan8.99
3Jacoby Ellsbury8.31
4Alexi Casilla8.19
5Gregor Blanco7.65
6Cliff Pennington6.26
7Rafael Furcal6.07
8Coco Crisp5.78
9Gerald Laird5.67
10Jose Reyes5.37



1xMike Pelfrey-7.00
2xJeff Suppan-6.33
3xUbaldo Jimenez-6.20
4xZach Duke-5.66
5xDerek Lowe-5.54
6xTed Lilly-4.87
7xJair Jurrjens-4.86
8xCole Hamels-4.70
9xRyan Dempster-4.69
10xClayton Kershaw-4.22

As usual, the trailers include a lot of pitchers, who don't tend to get a lot of bunt hits.  The first position players that appear on the list are Chris Young (-2.54), Brendan Ryan (-2.42), Tony Gwynn (-2.29), Yuniesky Betancourt (-1.70), and Juan Pierre (-1.13).  Pierre has appeared a lot in these two posts, typically as a trailer in some category.  Based on the data for these three years, he doesn't have the ability to be a productive enough bunter to offset his great number of bunt attempts.  In fact, of the ten players that topped the attempt% list I showed at the beginning of this post, Pierre was the only player to grade out as a below-average bunter.  One other note - I'm skeptical of the bunt runs values for Pennington and Furcal since Pennington had a bunt double and Furcal had two.  Bunt doubles are essentially flukes, and since doubles are worth a lot more than singles are, they skew the run value totals.  



With that, I'll put an end to this venture into bunting.  There are more questions that I'd like to investigate (team bunting statistics and the impact of leverage on bunting as two that come to mind), and most importantly, I think we just need more data.  For the time being, I would like to recognize Ichiro Suzuki as the best bunter of the past three years.