Monday, April 28, 2014

Alex Avila Was Clutch in 2013

What does one do when there's two days of no baseball (thanks a lot, rain)?  Look through FanGraphs' list of stats of course!  I was in the midst of comparing Alex Avila to the other catchers in the league (and he was right in the middle of the pack, where you'd expect a 92 wRC+ hitter to be).  When I got to the "Win Probability" tab, I noticed 2 things.  One, hey, there's a clutch stat!  And two, Alex Avila is sitting right on top at 2.09, making him the most clutch catcher in 2013.  Pretty neat, huh?

Looking deeper into FanGraphs' glossary, I got the formula to how they calculate this stat:



Clutch = (WPA / pLI) – WPA/LI

Avila had a WPA of 1.02, a pLI of 0.89 (1.02/0.89 = 1.15) and a WPA/LI of -0.95 (1.15 - -0.95 = 2.10).  There's probably a rounding issue here.

So what do these stats mean?  WPA is Win Probability Add.  It's a measure of the change in win expectancy from the time a batter enters that batter's box until the end of play.  Without getting too technical, there's a win expectancy chart that shows the percentage of how likely a team is going to win the game at any particular moment in the game.  For example, say a team has a 45% chance of winning a game.  Then a batter hits a HR.  Now the team has a 75% chance of winning.  The difference (0.75 - 0.45 = +0.30) gets credited to the player.  You may have already seen this in chart form:


Source: FanGraphs


pLI is the player's Leverage Index in all situations.  It attempts to quantify the "pressure" aspect of the game by factoring in the inning, the score, the number of outs and the number of runners on base.  Think of bases loaded in the 9th inning of a tie game as a really high leverage situation and a 13-0 blowout in the 5th inning a really low leverage situation.  Neutral is given a 1, "high leverage" is considered to be 1.5 and above and "low leverage" is below 1.  Avila's 0.89 is right around 1, which is to be expected.


WPA/LI is context neutral wins and is calculated on an individual play basis and then added up during the season.  During the game, some players will see more at bats in higher leverage situations than others, so this stat puts everything on a level playing field.  Sometimes a walk is just as valuable as a HR, so this stat will show that.


The subtraction of these two will result in showing only high leverage situations.  So what we are seeing is how much better a player performed in high leverage situations compared to how he performed in low and medium leverage situations.  FanGraphs also gives splits on these situations, so we can see them in more traditional stats:



2013
AVG OBP SLG wRC+
Low Leverage
0.218
0.317
0.324
80
Medium Leverage
0.225
0.316
0.383
93
High Leverage
0.290
0.324
0.645
148

And it's clearly noticable that Avila performed much better in high leverage situations than in low and medium leverage situations.  The difference between them was better than the difference of any other catcher last year.  And better than any other Tiger batter last year.  In fact, Avila was the most clutch player of anyone with at least 300 plate appearances last year.    

One last thing to keep in mind - there is no predictability value here.  Just because a batter preformed well in high leverage situations one year, doesn't mean he'll continue to perform that way every year.  This is why several people claim there is no such thing as a clutch hitter.  To prove this, look at how Avila has performed in high leverage situations in the past:

High Leverage AVG OBP SLG wRC+
2011
0.224
0.333
0.408
85
2012
0.295
0.360
0.477
127
2013
0.290
0.324
0.645
148

While Avila performed pretty well in High Leverage situations in 2012, in 2011 he was pretty poor.  If you remember, 2011 was Avila's big breakout year (.295/.389/.506, 140 wRC+ overall).  


FanGraphs says that this is a good storytelling stat.  And that's all this is, a good story.  

Wednesday, April 23, 2014

Phil Coke Is Now Throwing a Cutter

Last night the Tigers took a big lead into the 9th inning, leading the White Sox 8 to 3.  Manager Brad Ausmus thought this was a good time to get struggling pitcher Phil Coke some work as he had a little cushion.  It looked like Coke would get through the inning unscathed, striking out the first 2 batters.  But then he gave up a double, single and finally a home run to lefty Adam Dunn, surrendering 3 ER to an already bloated ERA.  His ERA went from 8.10 to 13.50 and his WHIP went from 1.80 to 2.25.  Even though he went 4 appearances without an ER to his name, Twitter exploded:

And there were many more just like that, some were not that nice.  

Curious on what pitch he threw, I went to Brooks Baseball and saw that it was a cutter.  In fact, he has thrown 17 cutters this year.   Now I don't remember Phil Coke ever throwing a cutter before, so I went back to his previous years and saw that he has only thrown 3 cutters before this year, all in 2010.  Pitch f/x data was still in in the early days then, so it is reasonable to assume that those pitches were merely misclassified.  

Over at Bless You Boys, a fanpost was made by J_the_Man suggesting that Coke throw less of his fastballs or even abandoning his sinker altogether.  Coke has evidently taken this a step further and switched to throwing a cutter instead.  Coke has thrown 15 of his 17 cutters in his last 3 appearances, so it's a fairly recent switch.  He still threw his sinker, but much less frequently.

Altogether, his cutter has resulted in 8 at bats being ended.  2 were strikeouts, 1 was a single and then the home run.  That's a batting average against of .250, but a slugging against of .625.  We're dealing with small samples for sure.  Looking deeper, both of the strikeouts from the cutter were from last night's game, so he obviously felt good throwing it.  It just didn't work on Adam Dunn.

Coke has thrown 13 of the 17 cutters to left-handed batters and only 5 sinkers to left-handed batters.  7 of the 8 at bats that have ended on the result of the cutter have come against left-handed batters.  This is where the transition is taking place, using the cutter instead of the sinker to get left-handed batters out.  Adam Dunn is a left-handed batter.  The pitch selection shouldn't be in question here, if it can indeed be a successful pitch to get them out (however his curveball has bee historically good against left-handed batters, so maybe it was a bad pitch selection).  The only at bat that ended on a cutter to a right-handed batter was a strikeout, though, so maybe it can be a decent pitch to right-handed batters too.

Here is a chart of his pitch location in/out of the strike zone from the cutter:



The home run he gave up was in the middle right of the strike zone, so it was a bad location as he did hang it.

And here's a chart of the whiff rate of his cutter:



As expected, the best results come from when he's keeping it down (and away from left-handed batters).  

Part of me wants to keep Coke around for little while longer to see how his cutter develops.  It is a new pitch for him, so some struggles should be expected.  Although I am also in the same boat as everyone else in that I'm about done with Coke.  His primary job is to get left-handed batters out and he simply did not do that last night.  Then I go back to small sample sizes and it was just one pitch on an attempt to correct something that hasn't worked the last couple of years.  I'm just very confused right now.  

Friday, April 18, 2014

Al Aburquerque Is Being Aggressive in the Strike Zone

Al Alburquerque has only faced 24 batters so far this season, so small sample caveats apply.  However, he's only given up 1 walk for a walk rate of 4.2%, which is a much better rate than he has shown in the past (15.9% in 2011; 15.1% in 2012; and 15.5% in 2013).  Based on his plate discipline numbers, there is evidence to suggest a change in approach to attack the strike zone more aggressively this year:


Year F-Strike% Zone%
2011 51.7% 43.5%
2012 54.7% 41.4%
2013 56.8% 41.8%
2014 66.7% 59.7%

His percentage of pitches in the strike zone is up almost 20% from past seasons and even his first pitch strike percentage is up about 10% from last year.  

One of the reasons for this could be his pitch selection.  Alburquerque's best pitch is his slider and while he's always thrown it at a high rate, he's throwing it even more often than in the past (according to Brooks Baseball):

Year 4-Seamer Sinker Slider
2011
38.1%
4.8%
57.1%
2012
5.8%
31.4%
62.8%
2013
5.2%
29.4%
65.4%
2014
16.9%
9.1%
74.0%

This is concerning from an injury standpoint, but again it is a small sample.  

Visually, this chart shows where in the strike zone he is pitching:




Compare this to the 3 years before:



The bottom right corner is still a hot spot for Alburquerque (mostly his slider) but it can been seen in the 2014 chart there are more hot spots in the strike zone.  An encouraging sign for someone prone to walking a lot of batters.

Of course when a pitcher is throwing more strikes, there is the potential for hitters to make more contact:


Year Contact% Z-Contact% O-Contact%
2011
58.8%
74.4%
38.6%
2012
58.5%
66.7%
47.5%
2013
62.4%
80.3%
41.4%
2014
81.1%
92.0%
58.3%

Alburquerque has a fairly high BABIP (.333), but his line drive rate is only 15.8%, down from 25.4% last year, so batters aren't hitting him as hard as last year.  So even though hitters are making more contact, they aren't making hard contact, which is a good sign.  Therefore, there might be some regression here as the season goes on:

Year BABIP LD% GB% FB%
2011
0.250
13.6%
56.8%
29.6%
2012
0.222
11.1%
63.0%
25.9%
2013
0.312
25.4%
40.4%
34.2%
2014
0.333
15.8%
42.1%
42.1%

Visually, we can see where he's been given up the hits in the strike zone:


Alburquerque has given up 7 hits and almost half of them, 3, have been right down the middle.  Maybe he's being to aggressive in throwing strikes?  It's one thing to throw strikes, but you don't want to put in on a tee for batters.  
   
Another thing that Alburquerque is doing that we're not used to seeing are low strikeouts, only 4 or 16.7%.  In 2011 his strikeout rate was 36.8%; in 2012 it was 34%; and in 2013 it was 31.8%.  The low strikeouts are making his Fielding Independent Pitching (FIP) look not as good as in the past:

Year K/9 BB/9 HR/9 FIP ERA
2011
13.92
6.02
0.00
2.08
1.87
2012
12.15
5.40
0.00
2.19
0.68
2013
12.86
6.24
0.92
3.72
4.59
2014
6.75
1.69
1.69
4.55
5.06

Alburquerque showing better control and command of his pitches is a good sign.  If he can keep it up, while still maintaining a low line drive rate and increase his strikeouts to where they were the last 3 years, he can possibly be one of the best relief pitchers in the league.  However, it's only been a little more than 5 innings, so it could just be a blip on the radar.  It's something to keep an eye on, though as the season progresses.