Trout’s a God & McCutchen’s a Fraud

I am compelled to ask because of confusion stemming from neo-sabermetricians writing when they provided reason for MVPs. Specifically, why was Mike Trout the AL MVP but Andrew McCutchen not the NL MVP? My seemingly odd query will become much less so below.

Writers had exemplified the power of WAR when describing Mike Trout’s value in the AL award race.  At the time that award winners were having the cockles of their hearts warmed over Trout, McCutchen was leading NL WAR. And yet no one was naming him MVP.  So why was WAR a justified primary argument for one league but not the other? 

To simplify, Trout had positively contributed in more areas for the Angels than Miguel Cabrera had for the Tigers, Robinson Cano for the Yankees, and Josh Hamilton & Adrian Beltre for the Rangers.  On September 25thKeith Law wrote Trout is the “rational” choice and not picking Trout is a “backlash against progress.”  But Keith Law did not apply this approach to the Senior Circuit.  On October 1st, Law had selected Buster Posey over McCutchen despite Posey having a lower WAR.  By October 2nd Jayson Stark, generally a balance between saber & tradition, used WAR for Trout’s AL candidacy. McCutchen was leading the NL, as noted by Stark, but pushed aside in consideration.  Spencer Schneider on October 9th went as far to solely site WAR for why Trout should win and Cano be selected second.

There were many regular season statistical parallels for McCutcheon and Trout.  Both led their respective league in WAR at time of press for most predictions.  Both played centerfield.  Both nearly had the same OPS (Trout’s .963 to McCutchen’s .953).  Both of their teams missed the playoffs.  Both were clearly the catalysts for their team’s success.  It is not a stretch to say McCutchen had less talent on his team  (no one will confuse Garrett Jones for Albert Pujols). 

This is not an article to question the value of WAR, nor is it making a case for McCutchen to win the NL MVP.  It is just an elaborate preamble for one final statement. In my eyes, it is indefensible to apply WAR to the AL but skirting the same standards for the National League. That’s called “cherry-picking”, folks. 

Ned’s Dead, Baby: How bad of a GM is Ned Colletti?

Ned Colletti has been the chief architect of the Los Angeles Dodgers’ roster since 2006. In this time, Mr. Coletti has undoubtedly sampled the fine and eclectic cuisine of that region. But perhaps he should sample some of the goods shipped east. How about a little Duncan Donuts? Without the emergence of two Red Sox clubhouse feuds, he would be on life-support as General Manager. Simply put: His track record has been that bad.

According to “Cot’s Baseball Contracts”, since he took over in 2006, the Dodgers are the only team in the NL West to average a $100M+ payroll ($107M plus a yearly average). Six out of 7 years Colletti has had a $100M+ payroll, which is extremely high for the mostly mid-market National League West.  Of the nine GMs who have been with a franchise as long as Colletti, only Brian Cashman ($203M+), Dave Dombrowski ($114M+) & Kenny Williams ($108M+) have averaged higher payrolls since 2006. Doug Melvin ($80M+) & Billy Beane ($61M+) deal with financial shortcomings from their respective ownership.  Those two also happen to be the only longer tenured GMs to never reach the World Series. Compared to a majority of teams in MLB, money has never been an issue for the Dodgers while Colletti has been in charge. This was even true after the last few years of the McCourt divorce proceedings.

But the issue is that Colletti has horribly misspent the money given to him by the franchise.  The following serves as a fine example of Coletti’s abysmal financial tenure. Here are free agents Colletti signed to multi-year deals prior to 2012: Juan Pierre, Jason Schmidt, Andruw Jones, Hiroki Kuroda, Rafael Furcal, Juan Uribe, and Matt Guerrier. None of the preceding players were traded for then resigned. Kuroda & Furcal were the only two that remotely worked out.  The others were, in a word, gross.  Pierre was traded, Schmidt was hurt, Jones was cut after 1 year, Juan Uribe rides the bench feigning an injury and seemingly lying in wait to be released, while Matt Guerrier has pitched in only 7 games this year. The pattern begins to emerge if you pay even the slightest bit of attention.

Colletti has always had money, but has invested little in the draft.  Between 2007-2011 the Dodgers draft budget totaled 26th in the league.  If you want to argue it was a result of the McCourt divorce, take a look at the major league payroll, operating at a high budget.  Or look to Zach Lee, drafted 28th overall in 2010 and given a signing bonus of $5.25M under McCourt’s watch. Who has Colletti drafted since 2006 that is contributing to the Dodgers?  Clayton Kershaw is a beast, winning the CY Young award last year.  But not too much can be spotted beyond him. Of those being used regularly on the 2012 pitching staff, Colletti can claim credit for Scott Tolleson.  Offensively, Dee Gordon was the opening day SS and played awful, but it is too early to write him off as a non-contributor in the future.  There is not another contributor that he drafted in the lineup, which seems almost impossibly bad. Matt Kemp, Chad Billingsley, Kelley Jansen & AJ Ellis were drafted before Colletti took over; Luis Cruz, Ronald Belisario & Andre Ethier were not drafted by the Dodgers.  After seven drafts, Colletti has failed to build a strong team.  There is only one superstar and one middle reliever who have played on a regular basis after seven years.  Amateur talent evaluation has been a weakness of the Colletti regime since day one, and it appears little is being done about it.

Andre Ethier is the only player on the Dodgers from the first major league roster Colletti constructed in 2006, meaning he has turned over the other 24 positions.  He has had a consistently lavish payroll, of which has been the envy of his divisional rivals in the NL West.  Yet in his seven years of massive financial advantage, the Dodgers have only been to the playoffs three times. He is closing in on a decade of draft picks with only two regular contributors to show for it. That is an epically bad track record.

While absorbing the context of his reign, one has to wonder how Ned Coletti has kept himself afloat.  Well, Andre Ethier was an absolute steal of a trade for one. And Luis Cruz has been a solid free agent for ½ a year, as has Ronald Belisario.  He overpays (by trade or sign) for medium talent that is enough to be productive. On the current team, he has brought in players like Ted Lilly, Chris Capuano, Aaron Harang, and Mark Ellis. These players are usually categorized as “good, but not great.”

Now in steps the Boston Red Sox to “fire” star players.  The Red Sox could not deal such talent to another AL team.  To further ruin leverage for the Red Sox, only a few NL teams are big-market enough to afford Boston’s headaches.  In 2008, the Dodgers, Mets, Cubs & Phillies were above .500 and classified as big-market.  The Cubs & Phillies were set with high-profiles corner outfielders, leaving the Mets & Dodgers.  If Boston wanted to alienate a fan base, they could have traded Manny to New York.  In 2008, Colletti absorbed Manny Ramirez after Manny was “just being Manny” and was no longer allowed to play for the Red Sox.  And just the other day, Colletti absorbed the monstrous contracts of Adrian Gonzalez, Carl Crawford and Josh Beckett.  In 2012, The Los Angeles Dodgers were the only club compatible for The Boston Red Sox.

And so it seems like The 2012 Los Angeles Dodgers have already constructed their tagline: Texting, chicken & beer, golf, and Hanley.

The new Dodgers ownership clearly has multiple blank checks to sign. But I guess everything boils down to this. Is it wise to have a GM write the checks whose only success has been to absorb expensive clubhouse cancers? 

Common Output in terms of consistency

This is a study in offensive consistency, trying to maximize streaks while realizing offensive explosions are exceptions not standards.  Common Output (“CO”) games are runs scored most often, in a consecutive range.  2-6 runs are scored more often than any other range.  An issue has been “Why include 2-run COs?”  They are clearly games of lower winning percentage.  And if the league average in runs scored is ~4.5/game, then a balanced CO of +/- 1.5 runs to make the range covering 3-6 runs makes sense. 

Except not including 2 runs would be make this invalid.  While teams lose a majority when scoring 2, and the league average is 4.5 runs, SB & HR leaders averaged 2 runs games MORE than 6 runs; or 7 runs or 8 runs for that matter.  Because this is about consistency, gauging 2-run results is a necessity in determining CO.


Common Outputs <1995 - 2011>

SB leaders

99 COs

61% of games

.461 winning percentage

 

HR leaders

95 COs

59% of game

.448 winning percentage

 On winning percentage, this equals 2 wins extra per full season for SB leaders.  SB leaders had 4 more CO games a year, and this would seem to be inconclusive.  But this is all about the percentages!  Remember that SB & HR teams average 2-run games more than 6-run games, 7-run, 8-run, and higher.  SB teams since 1995 have been more consistent compared to HR leaders.

What about non-Common Outputs? 

Non-Common Outputs <1995 - 2011>

SB leaders

63 non-COs

39% of games

.542 winning percentage

 

HR leaders

67 non COs

41% of games

.624 winning percentage


HR leaders average four games a year of non-CO and have a much higher winning percentage compared to SB teams.  But that is not consistent, as these games vary in runs scored between 0-1 & 7+.  Winning a higher percentage on a non-linear plot is volatile, and volatility is chaos.  Is it better to score less overall but score consistently or to have chaotic outbursts of guaranteed wins with guaranteed losses?


In a highly functional lineup, it takes nine hitters having a defined role & executing to avoid inconsistency.  A HR is an instant run, and the most powerful weapon of a lineup.  This easily explains why HR-teams score more than SB teams but it does not explain why SB leaders have a more consistent winning percentage of Common Outputs.  Why are National League teams not stealing more, or peppering a lineup with isolated Stolen Base threats?  If the Stolen Base is providing a template of consistency, why are teams not strategically placing & dedicated one or two more Stolen Base leaders to counter-balance the inconsistency of HR scoring?

HRs are expensive and inconsistent. Future posts will focus on the individual cost of Stolen Base threats vs Home Run threats, and the common occurrences of each.

From 1995 - 2011, NL team leaders in STOLEN BASES had a better winning percentage than NL team leaders in HOME RUNS when scoring anything but 2 or 5 runs a game.  The top graph compares winning percentages based off runs scored per game, while the bottom graph compares how often on average those amount of runs were scored per game.  
Common Outputs (CO) is now referred to scoring 2, 3, 4, 5 &amp; 6 runs as they were the most runs teams would score per game for both sides.  SB teams averaged a CO 99 times/year and won .461% of those games; HR teams CO was 95 per year with a .448% winning percentage.  SB teams are hitting a common output more often, and have a better winning percentage.  Based off this trend, NL teams have a better chance of winning when sporting a threat on base paths than a threat of power (interestingly enough, HR-leading teams have a higher overall winning percentage because of their propensity to score 8, 9 or 10+ runs a game).
CO games are the most common, the scoring output that seemingly every team can reach often.  The difference is that SB teams are winning CO games more and should be the focus of NL teams.

If SB teams win more in the most common outputs, how can stealing more lead to (1) more CO or (2) an increase of runs?

From 1995 - 2011, NL team leaders in STOLEN BASES had a better winning percentage than NL team leaders in HOME RUNS when scoring anything but 2 or 5 runs a game.  The top graph compares winning percentages based off runs scored per game, while the bottom graph compares how often on average those amount of runs were scored per game.  

Common Outputs (CO) is now referred to scoring 2, 3, 4, 5 & 6 runs as they were the most runs teams would score per game for both sides.  SB teams averaged a CO 99 times/year and won .461% of those games; HR teams CO was 95 per year with a .448% winning percentage.  SB teams are hitting a common output more often, and have a better winning percentage.  Based off this trend, NL teams have a better chance of winning when sporting a threat on base paths than a threat of power (interestingly enough, HR-leading teams have a higher overall winning percentage because of their propensity to score 8, 9 or 10+ runs a game).

CO games are the most common, the scoring output that seemingly every team can reach often.  The difference is that SB teams are winning CO games more and should be the focus of NL teams.


If SB teams win more in the most common outputs, how can stealing more lead to (1) more CO or (2) an increase of runs?

It was the best of stats, it was the worst of stats…

The best of baseball statistics take performance on the field to reach conclusions.  Power used to be determined by Home Runs — as time progressed and extra base hits became a by-product of power, Slugging Percentage developed to tell a more thorough story.  While Home Runs & Slugging Percentage are rooted in different times of baseball rationalization, they are conclusive.  Conclusive does not introduce variables statistics into equations, or leave to chance interpretation.  The past few years, equations Wins Above Replacement (WAR) & Fielding Independent Pitching (FIP) are being used to conclude the true value of a player.  The problem is the use of variables statistics (usually static numbers posing as conclusive) into equations rather than purely conclusive stats.


Variables!  Why are equations that utilize variables being consumed as holy?  You should use only absolutes whether sabermetrics or raw — homeruns, OBP%, ERA — stats that are not open to interpretation or conjecture.  It’s a matter of How to apply those stats vs what are the stats.  

WAR assigns historical data to develop an augmented reality of a players worth.  WAR does not utilize conclusive statistics exclusively, thus it cannot be relied upon as THE central argument in determining a players worth over another.  WAR uses the following  adjustments in determining the defensive value of one position over another.  

+1.0 wins C
+0.5 SS/CF
+0.0 2B/3B
-0.5 LF/RF/PH
-1.0 1B
-1.5 DH

FIP focuses solely on what the pitcher is responsible for minus his defense.  The equation is below:

(HR*13+(BB+HBP-IBB)*3-K*2)/IP … then uses a variable number to round into an ERA-type.

Think about this for a second — minus the defense!  This is impossible because FIP assumes defensive players as static robots who stand in the same position every time.  As examples the umpire who likes a high-strike zone dramatically alters how a pitcher throws pitch-by-pitch…the defensive positioning of a SS for a pull hitter…weather blowing in at 15 mph preventing home runs.  You cannot assign a singular defense statistic to attribute how the pitcher should be graded because he always plays with variables.  Thus you have to rely on conclusive data that does not introduce variable statistics, game-over-game, year-over-year.  ERA & WHIP & BABIP are paint a picture of consistency & conclusiveness over time that (1) minimize variables of umpiring, defense, et al

Can WAR & FIP provide use in painting a picture of value?  Sure if WAR & FIP are understood to be using variables.  They cannot be catalysts as absolutes in value or worth from one player to the next.  I aim to not use any variable statistics in my analysis.



*****
« UPDATE on March 31»
FIP angers me when it used to qualify or place value on all pitchers.  It should focus solely on power pitchers with the emphasis on Strikeouts and de-emphasis on “pitching to contact” pitchers such as Greg Maddux or Matt Cain.  Both named pitchers could strikeout a batter but focused on limiting scoring opportunities in the quickest manner possible.  Why throw 130 pitches over 6 innings when you can throw 110 over 8 innings with limited amount of hard hits?  If FIP were used to focus on grading pitchers who average 8 or 9 strikeouts per nine innings then we can have a possible value equation.  Relief pitchers would benefit from this perhaps.  FIP is not an absolute nor does it grade different type of pitchers on an equal playing field.



Pat Burrell vs. JD Drew

Is it weird to think JD Drew made almost $30,000,000 more than Pat Burrell?”  

I posted that comment in an MLBTradeRumors.com article upon the news of Pat Burrell’s retirement.  Burrell, like Drew, was a top 2 overall pick in the amateur draft.  Burrell, like Drew, had played in the majors for over a decade predominantly in the National League. Both had played for championship teams.  Yet this statement was met with much scorn.  The crux of my argument centered on (1) Burrell was noticeably healthier, and (2) the expectations of Burrell, as Burrell was constantly thought of to hit 3rd or 4th or 5th to drive in massive amounts of runs.  Burrell played in 140+ games 7 seasons to Drew’s 2 seasons.  Burrell had 7 seasons of 80+ RBIs & 9 seasons of 20+ HRs.  Drew has (being that he is still active) 2 seasons of 80+ RBIs & 5 seasons of 20+ HRs.  Easy enough, so how had Burrell received $30,000,000 less?  The answer has to be beyond JD Drew’s agent Scott Boras, considered the best agent in all of baseball.

Sabermetrics tells another story, and people were happy to point it out: Burrell has a career 21.9 WAR in 1640 games; Drew has a career 47.6 WAR in 1566 games.  Metrics showed Drew played tremendously better defense & was a much better baserunner.  I did not understand how a highly calculated system could be more wrong; how do “advanced” metrics show such appalling numbers for a 5th place hitter for two World Series teams?  How do “advanced” metrics show Burrell played poorly when he did his job year-in year-out?  Burrell was expected to have copious amounts of RBIs and he did.  Drew seemingly bounced around to whatever team gave him the most money; Drew didn’t even sign when he was the #1 overall pick!  How can advanced stats show Drew has had THAT much of a better career when no team has relied upon him to be a centerpiece?  No lineup built around or asking Drew to be an important cog won the championship — what are “advanced” metrics telling us??


* David O’Brien (Atlanta Journal Constitution): "I’d have to go with Burrell’s career over J.D. Drew’s, because of health. Drew has played more than 140 games only twice."

* Jayson Stark (Senior Baseball Writer for ESPN): "Tough call. At least Burrell was always into it!"

* Mychael Urban (Radio host): "Drew’s career isn’t over, is it? But if I had to make a call now, Pat’s had a better run. By a mile."


And this was a turning point for me in using “advanced” metrics.  Burrell hit 3rd, 4th or 5th 70.8% of the 1529 games he started.  Drew hit 3rd, 4th or 5th 50.5% of the 1393 games he has started.  Thus, Burrell was relied upon more.  Both are power hitters expected to drive in runs.  Raw stats, in this isolated argument, are showing Burrell as a better player while metrics are showing Drew far superior.  The true answer, of course, is subjective but if lineup placement is any indication, Burrell had a bigger role & impact. 

Keeping in mind the old adage “Numbers can be deceiving”, metrics need to be used carefully when calculating impact just as raw stats need to — neither tell the whole story nor should be used as an end-all-be-all.  Burrell & Drew have played predominantly in the National League.  What is the importance of lineup placement in the National League?  How can raw stats & metrics be used to create/understand the most consistent-scoring lineup possible when a National League team has to deal with the impending out of the pitchers spot? 

Why did Pat Burrell hit in the middle of the lineup 70.8%, 20% more of the time than JD Drew if all metrics indicate JD Drew the superior player?  How good of an agent is Scott Boras?