Saturday, June 29, 2013

The Tigers Shut Down After the 6th Inning

While watching the Tigers this year, there seemed to be one trend that kept coming up: the Tigers seemed to do nothing after the 6th inning.  Rather than chalk it up as selective memory, I decided to see if the numbers supported this.  

Here is how the Tigers offense ranks in the American League in the first 6 innings:

Stat Value AL Rank

First in every single category except walks and strikeouts, which they are still in the top 5.  There's a good argument to be made that the Tigers have the best offense in the American League during the first 6 innings.

And here is their pitching ranks in the first 6 innings:

Stat Value AL Rank
Runs Allowed

Most of this is due to their awesome starting pitching, which has already been well-documented as being the best in the American League. 

Now let's see how much they fall in the 7th inning and later.  First the offensive numbers:

Stat Value AL Rank
Runs 80 15th
AVG 0.227 13th
OBP 0.310 8th
SLG 0.319 15th
HR 14 T-14th
BB 95 5th
K 209 6th

Walks and strikeouts are middle of the pack, but everything else is near the bottom.  This is a complete 180-degree turn from the first 6 innings.

Here are the pitching ranks in the 7th inning and later:

Stat Value AL Rank
Runs Allowed 114 13th
ERA 3.98 13th
WHIP 1.30 12th
BAA 0.247 12th
HR 18 T-4th
BB 90 9th
K 266 1st

Strikeouts are still #1, but mostly everything else is near the bottom.  Most of this is due to the bullpen, which again has been well-documented as being bad this year.  

Offensive and pitching stats are taken from Baseball-reference as of games played through 6/28.

One sabermetric stat is the Pythagorean Expectation.  It shows how many games a team should win based on the amount of runs scored and runs allowed.  Right now the Tigers have a Pythagorean W/L record of 46-32, which is 3 wins better than their real W/L record of 43-35.  The formula is:

Win%= runs scored^2/(runs scored^2 + runs allowed^2)

Using the first 6 innings, the Tigers have a Pythagorean Win% of .701 which translates to between 113-114 wins over 162 games.  However, in the 7th inning and later, the Tigers have a Pythagorean Win% of only .330 which translates to between 53 and 54 wins over 162 games.  

This could be a small sample anomaly, but there's just one thing that needs to be asked, what happened to playing 9 full innings?

Friday, June 21, 2013

Comparing Miguel Cabrera to Ty Cobb

An interesting debate sprung up on the official Detroit message board asking if Miguel Cabrera has earned the title of "Best Tiger Ever" (I assume he meant of the positional players).  That honor has traditionally been held by Ty Cobb.  He's the Tigers all-time leader in batting average, runs and RBI while in the top 5 in OBP and SLG.  Advanced stats like him too as he's the Tigers top leader in wRC+, OPS+, fWAR, and rWAR.  Looks like both traditionalists and Sabermatricians can agree on Ty Cobb being the best Tigers player ever.

However, is Cabrera on his way to surpassing Cobb?  There are several counter-arguments to the above stats.  First of all, Cobb is done as a player; his stats are final.  Cabrera is right in the middle of his career; who knows what he'll do in the future?  Also, both have played in completely separate eras.  While the advanced stats above kind of neutralizes this, I think we can take it a step further.  

For this comparison, I'm going to take both players 5 best consecutive seasons, which just happens to be exactly 100 years apart (1909-1913 for Cobb and 2009-2013 for Cabrera).  Now this is kind of unfair to Cabrera because he's right in the middle of his 2013 season, so obviously his numbers are going to change after the season. Although Cabrera does have more PA, 3047 to Cobb's 3005, so I think it'll work alright.

First the traditional stats:

Ty Cobb
Miguel Cabrera

Now the advanced stats:

Player Years PA wOBA wRC+ OPS+ fWAR/150 rWAR/150
Ty Cobb 1909-1913
Miguel Cabrera 2009-2013

For the next set of stats, I'm using the same methodology as OPS+.  OPS+ is calculated as follows:

OPS+ = (OBP/lgOBP + SLG/lgSLG - 1) * 100

OPS+ adjusts for both the park and league and is put on an easy scale of 100 being average.  So if a player has an OPS+ of 110, he was 10% better than league average and someone with an OPS+ of 90 was 10% below average.

The hard part is figuring out the park factors.  Luckily, Baseball-reference already does this, not only for OBP and SLG, but also for AVG too.  League stats are easily accessible too.   So essentially  we can figure out AVG+, OBP+, SLG+, OPS+ and ISO+.  I also want to calculate BB+ and K+ using walk rates and strikeout rates, but since neither of those are balls in play, I don't think the park effects are necessary.

Player Years PA AVG+ OBP+ SLG+ OPS+ ISO+ BB+ K+
Ty Cobb 1909-1913 3005 148 136 162 198 114 108 63
Miguel Cabrera 2009-2013 3047 127 127 142 169 115 148 77

So now we have stats of both Cobb and Cabrera with about equal playing time, right in the prime of their careers, taking into account park effects and adjusted to the league.  A player wants to have a low strikeout rate, so having a K+ of less than 100 is a good thing.

Aside from the walk rate, it looks pretty clear that Cabrera hasn't reached Ty Cobb's level yet.  However, he still has many years of playing time to try to catch him.  Cabrera currently has an OPS+ of 188 for 2013 - the highest of any year of his career.  If he can sustain that for several more years, it could get close.  

Wednesday, June 19, 2013

Should Andy Dirks be Replaced in LF?

Andy Dirks has been disappointing this year, hitting .238/..291/.338 before today's game.  And he's been especially awful his last 17 games, hitting .183/.216/.197 from May 26th - June 17th.  This is after hitting .322/.370/.487 in 344 PA in 2012, leading many people to believe he can be an above-average everyday left fielder.  He hit .274/.354/.397 against left-handed pitching, proving that he didn't need a platoon partner.  However, he has failed to sustain his .365 BABIP, which means that 2012 couldn't been a fluke all around.  He's only hitting left-handers at a .182/.250/.273 clip so far this year.  

James Schmehl at MLive recently posed the question that Andy Dirks' job could be in jeopardy when the Tigers need a starting pitcher on Thursday.  When Anibal Sanchez went on the disabled list, instead of recalling a pitcher right away, the Tigers elected to call up Avisail Garcia again.  This is indication that the Tigers might at least be thinking about sending Dirks down to AAA; they currently have 11 pitchers on their roster and Leyland normally keeps 12.  If Andy Dirks does become the odd man out, there are 3 viable candidates to take his place, Matt Tuiasosopo, Avisail Garcia and Nick Castellanos - all have their positives and negatives.

Matt Tuiasosopo:

Why he should be the everyday LF:  He's been hitting excellent this year, .333/.465/.551 before today's game.  He's been hitting for average, power (.217 ISO), and patience (17.4% BB rate).  His 183 wRC+ demands more playing time.    

Why he shouldn't be the everyday LF: Before this year, Tuiasosopo was a career .176/.234/.306 hitter and hadn't seen ML pitching since 2010.  At age 27 it's hard to believe that he's all of a sudden figured everything out.  Some of his stats do suggest a regression.  His BABIP is at an unsustainable .417 as it's hard to believe he can keep a 29.4% LD rate.  His HR/FB ratio is at 21.4% but his career rate is only 13.1%.

Avisail Garcia:

Why he should be the everyday LF:  He's one of the Tigers top prospects that has some success against ML pitching.  In his limited time in the majors, he's hit .296/.339/.391 while adding power in 2013 with a batting line of .279/.315/.441 with 2 HR.  In his first 10 games at AAA, he hit .467/.489/.578, suggesting that he has nothing more to learn in the minors.  The Tigers keep adding him to the roster whenever there's a spot open, even though Leyland hasn't played him as much as most people would like.

Why he shouldn't be the everyday LF:  There is one stat that stands out that needs improving with Garcia - his walk rate.  It's at 5.5% so far this year and it was regularly in the 3-4% range all through the minors.  At 22 years old, he needs to play everyday as he continues to mature as a hitter and Leyland has been criticized for playing Don Kelly more over Garcia.  If Garcia isn't getting the playing time in the majors, then he needs to get them in the minors. 

Nick Castellanos:

Why he should be the everyday LF: He's the Tigers #1 prospect and he's been playing like one so far this year.   He's hitting .293/.366/.477 so far this year at AAA and he's been especially great lately.  After hitting .237/.284/.391 in his first 169 PA, he's hit .359/.455/.580 in his last 156 PA.  He's showing that he has nothing more to prove in AAA.

Why he shouldn't be the everyday LF:  He has no ML experience.  That could be dangerous, giving a rookie an everyday role in the middle of a playoff race who has been known to struggle after getting promoted after every step in the minors.  The Tigers might not want to take the risk of having him lose confidence and like Garcia, Castellanos should be getting everyday playing time as he matures as a hitter.

It's very likely that the Tigers will send down Avisail Garcia to make room for Jose Alvarez on Thursday.  At the very least, Dirks might lose some playing time to Matt Tuiasosopo without him actually taking over the position.  As bad as Dirks has been on the hitting side, advanced stats show that Dirks has been excellent on defense so far this year as Patrick O'Kennedy of Bless You Boys has pointed out.  This has given Dirks a 0.8 fWAR and a 0.9 rWAR, which means he's on pace for about a 1.5 WAR.  While that's below average, it doesn't really warrant a trip to AAA either.  

Besides, the Tigers have more pressing needs, such as the bullpen and the Jose Valverde situation.  

Update:  Avisail Garica has indeed been sent back to AAA to make room for Jose Alvarez.   

Thursday, June 13, 2013

Applying Leverage Index to the Closing Situation

Yesterday, Jose Valverde blew his third save in twelve chances and his second in six games.  In his last six games, Valverde has pitched 5 1/3 innings, giving up 10 hits, 7 earned runs, 5 home runs, 6 strikeouts and 1 walk for an 11.81 ERA and 2.03 WHIP.  His ERA went from 0.75 on May 27th to 4.15 after yesterday's disaster.  All in all, Tiger fans across the blogosphere and Twitterville are questioning if the Jose Valverde closer experiment has failed and if a change is imminent.   

Bless You Boys: Tigers Closer Jose Valverde on the hot seat
Pepper Baseball: Enough is Enough for Papa Grande...Right?
Motor City Bengals: Is Jose Valverde Dancing His Way Out of Town?
Walk Off Woodward: A Plea to Tigers Brass: End the Jose Valverde Experiment
Tiger Snark: No More Potato, Please
Motown Sports: Just How Long.......
Detroit Forum: ABBA Grande, The Dancing Queen

To answer this question, I decided to look at how Tigers pitchers performed in high leverage situations:

1 Jose Alvarez 1 2 2 0 0 0 0 0 0 0 0 1 .000 .000 .000 .000
2 Anibal Sanchez 10 27 24 7 4 2 0 0 0 0 1 6 6.00 .167 .185 .250 .435
3 Al Alburquerque 7 22 18 4 3 0 0 0 0 0 3 9 3.00 .167 .273 .167 .439
4 Drew Smyly 10 30 27 1 5 0 0 0 0 0 3 8 2.67 .185 .267 .185 .452
5 Joaquin Benoit 16 48 43 2 9 2 0 0 0 0 4 13 3.25 .209 .277 .256 .532
6 Justin Verlander 12 54 48 14 11 5 0 0 1 0 4 18 4.50 .229 .296 .333 .630
7 Luke Putkonen 3 10 9 1 3 0 0 0 0 0 1 2 2.00 .333 .400 .333 .733
8 Max Scherzer 11 55 46 13 13 2 0 1 0 0 6 14 2.33 .283 .352 .391 .743
9 Doug Fister 12 76 69 25 22 5 1 1 0 0 3 7 2.33 .319 .347 .464 .810
10 Darin Downs 11 28 24 8 6 2 0 1 0 0 4 7 1.75 .250 .357 .458 .815
11 Jose Valverde 10 38 33 7 8 0 0 3 1 0 5 9 1.80 .242 .342 .515 .857
12 Bruce Rondon 1 5 4 1 2 0 0 0 0 0 0 0 .500 .400 .500 .900
13 Phil Coke 9 36 30 8 10 3 0 1 0 0 4 6 1.50 .333 .412 .533 .945
14 Rick Porcello 9 37 33 13 13 1 0 2 0 0 4 6 1.50 .394 .459 .606 1.066
15 Jose Ortega 5 22 17 5 5 0 0 2 0 0 4 5 1.25 .294 .429 .647 1.076
16 Octavio Dotel 4 6 5 3 2 1 0 0 0 0 1 1 1.00 .400 .500 .600 1.100
17 Brayan Villarreal 4 12 6 3 2 1 0 1 1 0 5 4 0.80 .333 .636 1.000 1.636
Team Total 57 508 438 117 118 24 1 12 16 4 52 116 2.23 .269 .343 .411 .754
Generated 6/13/2013.

Leverage Index attempts to quantify pressure based on the inning, score, outs and baserunners; high leverage is high pressure situations, low leverage is low pressure situations.  From FanGraphs:
During the course of a game, some situations are more tense and suspenseful than others. For instance, we know that a one-run lead in the bottom of the ninth inning is more suspenseful than a one-run lead in the top of the third inning. Batting with two runners on and two outs in the eighth inning is filled with more pressure than batting in the same situation in the second inning. Leverage Index (LI) is merely an attempt to quantify this pressure so we can determine if a player has been used primarily in high-leverage or low-leverage situations.
From the table above, Baseball-reference used a Leverage Index of 1.5 to determine high leverage situations.  Al Alburquerque, Drew Smyly and Joaquin Benoit have been the Tigers best relievers in high leverage situations.  Alburquerque has been sent down to AAA because of control issues, so that leaves Smyly and Benoit as the best options in high leverage situations.

One of the problems of just using the best reliever for high leverage situations as the closer is that sometimes a situation will occur earlier in the game that is more high leverage than the ninth inning.  It's because of this that people argue that the closer position is overrated.  And I agree.  I want Drew Smyly to come in the game in a crucial moment instead of waiting until the 9th inning.  I also want Drew Smyly to stay in the game until it's finished, essentially being a multiple-inning closer.

Drew Smyly is a starter by trade, so limiting him to just 1 inning doesn't make sense.  He also isn't going to start anytime soon since the first time the Tigers needed a spot starter, they called upon Jose Alvarez, saying that Smyly wasn't stretched out enough.  Leyland has primarily used Smyly in the middle innings and based on the leverage situations, his role can be labeled as "mop up."   122 of his 152 (80%) batters faced have come in low or medium leverage situations despite his great numbers in high leverage situations.  Arguably, small samples need apply.  30 batters is hardly anything concrete.  Also, using Smyly like this will cause him to miss 1-2 days between appearances.  

However, Joaquin Benoit has been just as good as Smyly in high leverage situations and 44% of his batters faced have been in such situations.  Having a plan where both Smyly and Benoit alternate the closer role, each going both the 8th and 9th (and sometimes the 7th) is what I think, the best solution.  Putting a relief pitcher in the game and not taking him out until either a) he's tired or b) ineffective is a strategy that I've always believed in (unless of course that pitcher has severe split numbers.  I'm looking at your direction, Phil Coke).  Changing pitchers until one of them blows the game is a fear I never want to see happen.

The last time a Tigers starting pitcher failed to go at least 6 innings was Rick Porcello on May 23rd.  That means the Tigers have gone 19 consecutive games where the starter has gone at least 6 innings.  In fact 48 of the 64 games (75%) the Tigers have played, the starter has gone at least 6 innings.  The need for a "long man" in the Tigers bullpen just isn't there.  The best way to maximize Smyly's talents is to use him late in the game.  

The Tigers already have relievers capable of closing.  Jim Leyland just needs to use them properly.