Why Barndon, and everyone else here, needs to stop ignoring Sabremetrics

Here's one wagon you SHOULD jump on!

Well, I read your post Brandon, and I would like to respond. 

First of all, I want to address your Toy Story metaphor.  I don't think it is very accurate, so I'll propose something I think is closer.  For a long time, Isaac Newton's laws of motion were viewed as the canonical "stats" when it came to describing physics.  Over time, however, it became obvious that they weren't as precise as was needed.  This is when new "stats" like special relativity and quantum mechanics were developed.  Newton's laws still give you a rough idea of what is going on, so they certainly still have use, but in instances where you want to become more precise, you have to turn to the more advanced forms.

The traditional statistics certainly tell you something.  Average tells you how good a player is at getting hits.  OBP tells how good a player is at getting on base.  Fielding percentage tells you how many times a player will make an error.  But, do any of these traditional stats tell you how good a player is in overall offense?  Sure, you can try and look at their average, their OBP, their steals, and their Slugging %, but how do you value what you see?  Is a player who hits .300/.350/.500 with 40 steals and 20 HR better than a player with .280/.375/.555 with 5 steals and 25 HR?  I don't know, do you?  All these traditional stats, but they can't tell us the most important question, how good are either of these players?  Which one is better?

Let's break it down a bit.  What is the goal of offense?  It's to score runs, correct?  Well, how much better is a double than a single?  To find the answer, you need to know how many runs, on average, a double gets you, and how many runs, on average, a single gets you, and find the difference.  Records of baseball games have been kept for over a hundred years, so we have a gigantic sample size on which to base statistics.  In the above example, we can see that on average, a double nets you 1.08 runs, while a single nets you .77 runs (both are runs above making an out, btw).  Knowing the run value of anything you can do on offense (IBB, BB, 1B, 2B, 3B, HR, SB, CS), to find how many runs a player was worth on offense, all you have to do is multiply everything they did by the average run value, and add it all up.  Now we know how many runs a player's offensive was worth, and this is something concrete.  When we take this number and divide by plate appearances, we get a stat that is known as wOBA.  This stat encompasses everything you can do on offense, so it allows you to compare players with vastly different styles of play.  None of the traditional stats allow you to do this.

Here is another caveat:  How do we know how players compare that play in different ballparks?  The classic example of this is Matt Holliday.  For three years, from 2006 to 2008, Holliday had a ridiculous wOBA of .418.  For reference, Arod's wOBA for that three year span was also .418.  By looking at his traditional stats, many people in baseball began saying that Holliday was one of the best hitters in the game.  They didn't take into account, however, that he plays in Coors field 81 games of the year, which is a big-time hitters park.  When Holliday went to the A's in 2009, he was now spending half his games at Oakland Coliseum, a pitcher's park.  He is now with St. Louis, who also play in a pitcher's park.  His numbers have come back down to earth, showing us that while he's still a great player, he's not as good as Coors made him look.  To account for this, we can look at years of data and see how much a certain stadium inflates or deflates the numbers of players, both hitters and pitchers.  A stat that does this is wRC+.  It takes the base of wOBA, and adjusts it based on the parks a player was playing in.  Obviously, hitters that play in Colorado will see their numbers deflated, and hitters that play in San Diego will see their numbers inflated (the opposite would be true for pitchers).  To make the numbers easy to understand, wRC+ is also normalized so that a score of 100 is league average, anything below that is below average, and above that is above average.  What wRC+ tells us in our original example is that Matt Holliday was 49% more productive than the average player from 2006-2008, while Arod was 59% more productive.  So while their numbers looked very similar, Arod was the better player because he put up the same numbers in a less hitter friendly park.

Another great example in favor of sabremetrics is Ichiro.  People look at his gaudy batting average and hit totals, and come to the conclusion that he is a superstar offensive player.  I've even heard people on this site say he's one of the top hitters of this decade.  What are the facts?  Ichiro hits a lot of singles.  That's it.  He doesn't hit many extra-base hits, and he doesn't take walks.  Is he a good hitter? Sure.  Is he great?  No.  wRC+ has him at a career 122 hitter.  That's certainly a good player who will make some All-Star games, but a 122 wRC+ is not superstar level.  Arod's wRC+ in his down year this year was 127, and he's a career 152 player.  Pujols, as another example, is a career 173 hitter.  Those are superstars.

As you have all heard me say numerous times, defense is very hard to quantify.  That doesn't mean that we shouldn't try our best, however.  Defense has been a misunderstood aspect because our eyes often fool us.  People see Jacoby Ellsbury making diving catches and think he's a great center fielder, but they don't realize that he makes routine plays look much harder because he takes bad routes to balls.  This is why it's just not wise to trust our eyes when it comes to defense.  Let's do the same thing we did before.  What is the purpose of defense?  It's to prevent the other team from scoring runs.  How do you do that?  Getting to balls hit in your area of responsibility (range), not making mistakes (errors), limiting baserunners from taking extra bases for outfielders (arm), and turning double plays for infielders (DP).  UZR makes an attempt to account for all of these factors, so as to quantify the overall defense of players.  You can certainly make the argument that certain aspects are undervalued or overvalued, and there is definitely a problem with batted ball types being treated the same, but until someone comes up with something better, UZR and Total Zone rating are the two best defensive stats we have.  With enough sample size (generally accepted to be 3 years of data), we can get a rough picture of how defenders compare.  UZR should not be taken as gospel like wRC+, but if there is a large gap between two players, you can assume that the one is better than the other.

The argument that these stats are too complicated is absolutely ridiculous as well.  None of us hand-calculate stats.  Do you count up all a player's hits or total bases and divide by their at bats or plate appearances to get your stats, or do you just look at a site where AVG and SLG are already calculated and listed for you?  That's the same thing I do.  I don't have the linear weights for wOBA memorized, I don't need to know that.  I know why wOBA is a more trustworthy stat and I understand it, but I don't need to know exactly how to calculate it, it's listed on Fangraphs for me.

Finally, none of us sit there watching the game, thinking about how a player's wRC+ will go up or how they can improve their UZR by reaching balls.  Just because we chose to trust stats that are more precise in their evaluations of players doesn't mean that we watch the game any different or enjoy it any less than you.  You can feel free to keep using Batting Average as a comparative tool among hitters, but someone who uses a more inclusive statistic will always trump you, just as an electronics company who chose to ignore quantum mechanics would always produce inferior products than one who manipulated the laws of QM.  If you chose to ignore the better stats that have been created, you aren't less knowledgeable, but you are less informed, so you put yourself at a disadvantage.