Monday, September 29, 2014

2014 MLB Awards

The 2014 MLB season was certainly not the craziest on record.  Many division leads were large and have been held for quite some time, which has led to minimal intrigue in the last few weeks.  Most of the division winners were the teams we expected to win their divisions, so there hasn't even been much drama there.  That said, even the most by the book seasons of baseball are still season of baseball.  Since baseball is awesome, that means there is still so much awesomeness to celebrate, that one doesn't have the time to fully appreciate it all until it's over.  Given that the regular season is now over, I will perform that appreciation here, by picking the winners of the major MJB awards.

As usual, I am just picking the six awards I care the most about.  Everything else is either too far afield of my interests (manager, executive) or too silly (comeback player) for me to spend time writing about.  Also as usual, I will just list the number of players I feel like for each award, instead of fitting my picks within the framework of the BBWAA.  Perhaps when the BBWAA starts admitting people who occasionally post to a personal blog, I will change my tune.

*Since I don't write about baseball often, I always like to note that the majority of stats you see below come from Fangraphs.  For a good explanation of the stats used in this post, see the Fangraphs library.  For a good explanation of why we should use these stats, even though they aren't perfect, see here.


NL Rookie

1. Jacob deGrom - SP, NY Mets
2. Billy Hamilton - OF, Cincinnati

Let's get this one out of the way first.  The NL rookie crop was not particularly deep this year, which left just two real contenders for the award.  While I like Billy Hamilton and understand that he created a lot of value by being great at defense alone, I have to give the nod to deGrom.  Hamilton is lauded as a supreme baserunner, but was caught stealing on 30% of his stolen base attempts.  He also put up a wRC+ of just 79 (meaning he was 21% worse at batting than an average MLB batter), which shows that there are a few significant holes on the offensive side of his game.  Jacob deGrom, on the other hand, was one of the most pleasant surprises in baseball this year, which helps to give the Mets legitimate hopes for a top rotation next year.  deGrom struck out 25.5% of all batters faced in his 140 Major League innings, which is a higher rate than he achieved at any level of the minors.  That he was able to almost match Hamilton's fWAR in only about two-thirds of a season is as good of an argument as any for him to win the award.


NL Cy Young

1. Clayton Kershaw - LA Dodgers
2. Stephen Strasburg - Washington
3. Jordan Zimmerman - Washington
4. Zack Greinke - LA Dodgers
5. Adam Wainwright - St. Louis

I was strongly considering just putting Kershaw's name here, and leaving it at that.  I don't think anyone would blame me, given that this is what he has done this season:

27 starts, 198.1 innings pitched - an average of over 7 innings per start
1.77 ERA - First in the NL
31.9% Strikeout Percentage - First in the NL
4.1% Walk Percentage - Fourth in the NL (slacker)
51.8% Ground Ball Percentage - Eighth in the NL among qualified starters
1.81 FIP - First in the NL by over half a run
2.08 xFIP - First in the NL by nearly half a run
2.09 SIERA - First in the NL by over half a run
7.2 fWAR - A full two wins ahead of everyone else in the NL, in spite of the relatively low innings total

So yeah, Kershaw was a comfortable first in ERA and every reputable ERA estimator.  Sabermetrics often open up baseball-related debates to a wider swath of possibilities than the narrow views often held by mainstream sportswriters.  This is one time where that is not the case.  Kershaw was and is the best pitcher in the NL by a mile.  There is no debate.

All that said, there are some worthy runners up that I wanted to take a moment to recognize.  Stephen Strasburg finally put together a full season of awesomeness, going over 200 innings for the first time, while still maintaining an insane strikeout rate (27.9% - 2nd among qualified starters, behind you-know-who).  His xFIP (my preferred measure of pitcher quality) of 2.56 was a comfortable second in the NL, which shows that his overall quality as a pitcher absolutely matches the initial hype from his rookie season.  Jordan Zimmerman almost matched his teammate in excellence, attaining both a career high in strikeout rate and a career low in walk rate (3.6% - 2nd in the NL), which led to the best season of an already great career.  Unlike the two Nationals on the list, Greinke and Wainwright didn't quite have career years.  However, both continued to march along as top-ten pitchers, which is impressive enough to end up on my list.


NL MVP

1. Jonathan Lucroy - C, Milwaukee
2. Clayton Kershaw - SP, LA Dodgers
3. Andrew McCutchen - OF, Pittsburgh
4. Buster Posey - C, San Francisco
5. Giancarlo Stanton - OF, Miami
6. Anthony Rendon - 3B, Washington
7. Carlos Gomez - OF, Milwaukee
8. Russell Martin - C, Pittsburgh

The NL MVP race is quite odd this year.  In spite of the lack of a slam-dunk candidate, the field is still very strong.  Certain Cy Young winner Clayton Kershaw will probably run away with the award in reality, and the last two winners (McCutchen and Posey) both had amazing seasons that were basically of the same quality as the years they won.  Young guys Giancarlo Stanton and Anthony Rendon had their first true superstar seasons, while veterans like Carlos Gomez and Russell Martin turned in fantastic efforts.

What makes the NL MVP race the most odd, however, is that I am picking none of those players to win the award.  Instead I am going with the guy who has sneakily become the best catcher in baseball, Jonathan Lucroy.  His traditional numbers likely won't wow you (13 HRs, 69 RBI), but his batting line of .301/.373/.465 would be spectacular even if he wasn't a catcher, and his 6.3 fWAR (4th in the NL) displays his overall value quite well.  Of course, that gaudy WAR total doesn't even include his typically awesome framing numbers, which adds 2-3 wins of value to his resume.*  Even though the dark art of measuring this additional impact of catching is in its infancy, there's a lot of evidence to support the claim that Lucroy is among the best as helping his pitchers out.  You can't go wrong championing Clayton Kershaw for the MVP award, but I think a full evaluation of what it means to be valuable supports Lucroy as just as worthy of a recipient.

*There's a lot of debate as to where the value of framing should be assigned.  Of course, credit should go to the catcher, since he's the one doing it.  That said, if a pitcher knows he's dealing with a top catcher, he's more likely to work the corners knowing that the extra strikes are there for the taking.  If a pitcher can execute that game plan successfully, he should surely share in some of the credit.


AL Rookie

1. Jose Abreu - 1B, Chicago White Sox
2. Masahiro Tanaka - SP, NY Yankees
3. Dellin Betances - RP, NY Yankees
4. Mookie Betts - 2B, Boston
5. Marcus Stroman - SP, Toronto
6. Colin McHugh - SP, Houston
7. Kevin Kiermaier - OF, Tampa Bay
8. George Springer - OF, Houston
9. Danny Santana - SS, Minnesota

Tanaka was well on his way to winning this (and maybe other awards), and then got injured.  Thus, Jose Abreu ends up winning the award handily.  I was tempted to still give the honor to Tanaka given how freaking good he was (a 2.58 xFIP shows how amazing he could be if he could just keep the ball in the yard a little more), but Abreu's 36 homers and .383 OBP is such an impressive combination over a full season that I couldn't bring myself to do it.  I agree with others that it's a little odd to give this award to players that established themselves in other leagues, but they are technically rookies, so they end up in the top two spots. 

You may have noticed that Mookie Betts garnered the #4 spot on this list ahead of many players who played a full season or close to one.  While I strongly consider playing time in the MVP and CY Young, it matters less to me in the Rookie of the year award.  The main reason for this is that rookies can't really control when they're called up and when they will play.  Even if you're raking in AAA, you might be held down for several reasons: lack of room on the MLB roster, not wanting to insert a rookie into a pennant race immediately, or waiting until the super-two deadline to add an extra year of team control.  Given that you lose your rookie eligibility after just 45 games played or 50 innings pitched, I like to make sure to recognize guys like Betts who haven't played as much as others, but have clearly shown that they are a fantastic player with a long career ahead of them.


AL Cy Young

1. Felix Hernandez - Seattle
2. Corey Kluber - Cleveland
3. Phil Hughes - Minnesota
4. David Price - Tampa Bay/Detroit
5. Chris Sale - Chicago White Sox
6. Jon Lester - Boston/Oakland

This is the most insanely loaded ballot for any award in my four years of doing this.  Max Scherzer basically matched his Cy Young season, and doesn't even make the list because a typical Cy Young season apparently isn't superlative enough.  Jon Lester had a career season, and essentially pitched the A's into the playoffs (2.35 ERA in 11 starts).  Chris Sale put up the best rate stats in the AL (30.4% strikeout %), and only takes fifth because injuries are stupid.  David Price led the league with 248 innings and a 23.1% K%-BB% (the best back-of-the-envelope pitcher quality calculation).  How on Earth is he only fourth?  And finally, Phil Hughes only put up the best strikeout to walk ratio in the history of the sport.

In spite of all those worthy contenders, the race essentially comes down to two men: Corey Kluber and Felix Hernandez.  Both were near perfect this season, so any attempt to separate the two is going to come down to some nit-picking.  Felix took the league crown in xFIP (2.51 to 2.57 for Kluber), while a low home run rate gives Kluber the edge in FIP (2.35 to 2.56).  Kluber struck a few more people out, while Felix limited walks slightly better.  Felix induced more ground balls, while Kluber was able to keep the ball in the yard in spite of more fly balls.  The pitiful Indians defense (29th in the majors in UZR) strongly suggests that Kluber had to go it alone when he was on the mound, but Felix wasn't exactly pitching in front of a bunch of gold glovers either.  In the end, I am forced to go to a  Bayesian tiebreaker, and give my vote to the person who I thought was the best prior to the season.  Since Felix has been doing this for almost a decade, he gets the nod. 


AL MVP

1. Mike Trout - OF, LA Angels
2. Felix Hernandez - SP, Seattle
3. Corey Kluber - SP, Cleveland
4. David Price - SP, Tampa Bay/Detroit
5. Jose Bautista - OF, Toronto
6. Michael Brantley - OF, Cleveland
7. Josh Donaldson - 3B, Oakland
8. Adrian Beltre - 3B, Texas
9. Ben Zobrist - 2B/SS/OF, Tampa Bay*
10. Alex Gordon - OF, Kansas City

Mike Trout will most likely actually win the MVP this year, which means I won't have to make an impassioned plea for him as I usually do.  What is interesting about this is that Mike Trout MVP season is easily his worst full season.  His walk rate dropped from 15% to 12%.  His strikeout rate climbed to 26%, and his 184 Ks led the league.  His batting average dropped below .300 for the first time.  His fielding numbers were decidedly below average, according to both UZR and DRS.  He only stole 16 bases, after stealing more than 30 each of the past two years.

And yet, in spite of all of this, he still deserves to run away with the award.  Basically, he was so good in 2012-2013 that he could survive a mild downturn in production and still be the best player in the world.  The increase in strikeouts came with an increase in power, and his career-high 36 homers, helped power him to a 167 wRC+ (just ahead of Victor Martinez for best in the AL).  He played 157 games, which shows a durability that some other candidates lacked.  Playing a slightly below-average center field still makes you one of the best defenders in the game.  Baserunning is more than just steals, as Trout added roughly half a win on the basepaths.  And finally, if you care about context-dependent stats, Trout led the majors in both WPA and RE24.  Mike Trout will not always put up 10-win seasons, but he just might always be the best player anyway.

*There are too many players here to write about everyone, but I did want to mention that Ben Zobrist has been the second most valuable player in baseball over the past six years.**  That is insane.  Ben Zobrist is and probably always will be the most underrated player in the game.

**Your daily "Trout is in another stratosphere" reminder: Trout is 10th on that list, in spite of not becoming a full time player until halfway through that time period.

Tuesday, September 16, 2014

College Football Thoughts

At the end of week three, we find ourselves 20% of the way through college football's regular season.  This means we've heard about 20% of the ridiculous arguments that college football fans love to make.  Here to refute those arguments is me.

Argument 1: The Big Ten is awful and that means Michigan State won't make the Playoff

This is a good one to start with, because it shows that many of these arguments have a bit of truth to them.  It is quite obvious that the Big Ten is not having a good season, with all but three teams having already lost a non-conference game.  Combining those on-field results with general anger from many fan bases paint a pretty bleak picture about the rest of the season for the conference. 

In spite of the Big Ten's self-evident terribleness, the second part of that argument doesn't hold water.  The general argument I've read is that the lack of potential marquee wins will eliminate Michigan State from contention.  While I agree that the Spartans' upcoming schedule doesn't feature much in the way of top opponents, there is more than one way to collect a big win.  If Ohio State turns out to be about the 20th best team, would a three touchdown victory by the Spartans be any less impressive than a three point victory over a top-5 Buckeye team?  Would it perhaps be more impressive?  The current spread between Sagarin's top team (Oregon) and 20th best team (Kansas State) is only 11 points on a neutral field.  While the top teams carry an air of invicibility about them, they're really not orders of magnitude better than other teams.  If Michigan State can navigate their remaining schedule in dominant fashion, they should have a strong argument for inclusion in the Playoff, regardless of what other pratfalls befall the Big Ten.

The main message of all of this is to not fall victim to the conference fallacy.  Michigan State appears to be a step above the rest of their conference, and shouldn't be overly punished for other's down years.  It's both cliché and true that they can only play the teams laid out before them.  If they play those games in a manner that makes them appear to be a top four team, they should be in the playoff no matter how stained their conference's reputation is.*

*All of this assumes that the Big Ten remains at least at the level of the AAC or higher, which they currently are (scroll down here).  If instead you're playing something like Marshall's schedule, then you pretty much need to dominate every snap.


Argument 2: How can Georgia be ranked above South Carolina?  Stanford over USC?  Human polls are terrible, and shouldn't start until later in the season

One thing you can always count on in college football is people complaining about rankings that don't perfectly correspond to the head-to-head matchups that preceded them.  This season is giving us a couple of great examples with some of the marquee matchups from the early weeks.  If you paid attention to the polls this week, you may have noticed that the AP has Georgia one spot ahead of South Carolina, and Stanford one spot ahead of USC.  This isn't just a human poll phenomena either.  If you look at some of the major computer polls, you'll see Georgia and Stanford ranked higher than the teams they lost to in many of them.

While the poll guidelines do say to consider head to head results, I think those generally get overrated when people compare teams.  Both UGA-USC and Stanford-USC were three point games that could have easily gone either way (the former was literally decided by a chain link, and the latter was all kinds of crazy).  I accept that we can use the final results of these games as tie-breakers; since the season is so short, we really don't have much of a choice but to treat it as small sample theater.  That said, for evaluate purposes games between great teams that go down to the wire should generally be thought of as ties.  Do we really think that the USC team that gave up almost 500 yards rushing to Boston College is better than the Stanford team that has destroyed its non-USC opponents thus far?  Are we completely forgetting week one, when Georgia looked as impressive as anyone, and South Carolina couldn't stop anything?  If we're being honest with ourselves and evaluating the whole of what these teams have accomplished, it's hard not to agree with the ordering of these teams in the polls.

The second part of this argument is a little tricker, but even more important than the first.  I touched on this a bit in a post from last season, but I thought I would address this particular point a little more right now.  Typically, when a highly-ranked team loses early in the season, the chorus of complaints arises: Why was that team so highly ranked?  We don't know anything?  Why even do rankings early in the season? I agree that our knowledge in the early part of the season isn't perfect, but I would argue that our relative lack of knowledge about teams in September is actually less of an issue than out adherence to ordering teams by number of losses in November.  The famed college hoops analyst Ken Pomeroy has studied this in his sport, and found that pre-season polls generally predict post-season success better than late season polls.  FiveThirtyEight also uses the pre-season poll in its model for predicting the tournament

The basic reason why this works is that, when freed from the sometimes misleading small-sample win-loss record, analysts are able to better pick which teams are truly the best.  We will probably never be able to fully predict which young offensive line will jell or which freshman QB will end up winning the Heisman, but pre-season predictions still end up doing a decent job of telling us which teams will be good and which will be bad.  That our predictions and our models aren't perfect doesn't mean we shouldn't use them*; rather it means we should strive to continue improving what we know, while acknowledging the usefulness and limitations of our current knowledge.  In the end, this means that we shouldn't take early season polls overly seriously, but we also shouldn't stop trying to do the best we can.

*If you're going to read just one hyperlink from this post, make it that one.  A good summarization of how one should treat current models in pretty much any field.

Argument 3: Why didn't Georgia run the ball with Gurley on first and goal?

Background on this in case you weren't watching: Georgia was down three late in their game against South Carolina.  Following a Gamecock interception, they had a first down at the four yard line.  The first play was a pass, after which they got called for intentional grounding.

I understand the general idea behind this sentiment.  Todd Gurley is awesome, and is most certainly Georgia's best offensive weapon.  With three plays to go for the win, it makes sense to spend at least one of them letting Gurley try to make something happen.  That said, I think this complaint is a bit silly for a few reasons:

It smells of post-hoc fallacy - Quite simply, It seems like fans are only angry because this didn't work out.  Earlier in the game, Georgia had trusted goal line plays to both their QB Hutson Mason and to their seldom-used fullback, and both resulted in touchdowns.  Had the late-game play worked as well, we would have heard nothing of this.

Georgia rushed and passed about equally well - If we credit sack yardage to the pass game (this is one of the few areas in which I prefer the NFL, as they actually do this), then we get the following numbers for Georgia: 34 rushes for 222 yards, 26 passes for 186 yards.  While I don't have anything more advanced available at the moment (success rate would come in handy here), taking the simple averages shows that the two methods of collecting yardage were roughly equivalent.  Todd Gurley has a lot of hype, and we remember all of the times he's come up with amazing plays, but he isn't perfect, and the Georgia pass game seems to be coming along well enough to trust it in important situations.

Play calling is about game theory - This is probably the most important counter-argument.  Here's what the Head Ball Coach had to say about it:


The field of game theory is a complicated area, but the basic practical implication is this: In order to attain the best long-term outcome, you often need to make a sub-optimal decision in the short-term.  Even if we make the concession that running the ball is the best option for the Georgia offense, that doesn't mean they should run the ball every play.  Because of this, it's tough to judge play-calling on individual plays, except in certain circumstances (if a coach calls a boneheaded play on say 4th and 1, feel free to call for his firing all you want).  Rather, one needs to take a more overarching view to truly judge the quality of play calling.  Per my count from the last section, Georgia ran the ball about 60% of the time against the Gamecocks, which most would probably agree is a good balance given their relative abilities. 

In the end, you can still question Georgia's late game play calling all you want.  I'm just not convinced there was an easy right or wrong answer in that situation.  What I find really silly is that this line of thinking has seemingly put the blame on Mike Bobo, when the Bulldog defense was once again the main reason Georgia lost a big game.  Bobo hasn't been perfect during his near decade as the Bulldog's offensive coordinator, but he's overseen multiple quarterbacks that went on to the NFL, and has led an offense that has been the program's greatest strength over the past few seasons.  Let's pump the brakes a bit before getting aboard the #firebobo train.


Argument 4: The SEC is Overrated

Nope