Concepts from Moneyball

I love baseball statistics, always have and always probably will.  There's just something to them that rings true to me.

At their core, baseball statistics are a measurable way of calculating the worthiness of individual players and a way that baseball games of the past can be recreated in some fashion.  Due to the nature of the statistics, they are inherently less obscure in telling individual outcomes than say, football statistics are.  In football the stat line you will see about say, a running back, will look like 23 carries, 104 yards, 1 TD, 1 Rec. 5 yards.  On average that means 4.52 runs per carry which doesn't tell you that he had one carry for 87 yards and the rest averaged .77 yards per carry.  It's harder to recreate games in football through the box score so usually, without watching the game, its harder to tell what exactly happened without further analysis.

Baseball box scores read simpler and easier to understand what happened in an individual game basis.  Traditional stats (RBI/H/Runs/AVG/SB for hitting Wins/ERA/WHIP/Ks/Svs for pitching, also known as the 5x5 rotisserie statistics for fantasy baseball) are easy to comprehend (except WHIP which is a fantasy made up stat but we will get into specifics later).  You see that say, Ryan Braun went 2 for 3 with 2 RBI's 1 run and 1 HR and you already have a good idea what happened.

Traditional stats are however, misleading and in many ways bad indicators of true performance both past and future, and this is the over-reaching concept of the book Moneyball.  I liked the movie, as far as movies go, but to get a true understanding of the concept one would have to read the book.  The movie spent a lot of time showing the emotional side of book and did it in a great manner.  Unfortunately it had the opportunity to get others to think more critically about baseball and it's base statistics and why they are bad indicators of true performance.

This article will attempt to show the general themes of the book, broken down by stat used or concept used by Billy Beane and Paul Depodesta during their 2002 season.  Whether or not they are trying to use these concepts anymore is a different story (the downside to the book I believe is that their concepts made their way into other teams so the players they wanted were not longer cheap which is what they need to compete)

Walks


Walks in baseball are highly underrated, even today, but less so than they used to be.  Walks have a lot of advantages to them and basically zero disadvantages.  Many people wrongly associate walks are solely on the pitcher giving them up, which is wrong depending on the player.  Plate patience plays a larger role in the number of walks a batter has over the course of the year than anything else.  There was an interesting chapter of the book entirely dedicated to Scott Hatteberg based almost entirely on walks and plate discipline. Walks do the following for the batter and their respective team:

     a) It's not an out.  This is the most important concept to note on why a walk is a great outcome.  It's not an out and has zero possibility of becoming an out on it's own.  An average baseball player will have an out something like 67% of the time they head to the plate (league average OBP is somewhere around .330 i believe but it fluctuates yearly).  Walks are not subject to luck like hitting a ball in play does.

     b) It makes the pitcher throw more pitches.  Generally speaking, Starting pitchers are the best pitchers on their team.  Followed by closers (an overrated position), then set up (8th inning guys), then long relief.  Long relievers are who you want to get to.  In modern baseball, starters are generally out of the game around 100 pitches.  (Earlier in the year it's lower, later in the year it sometimes means more innings).  The sooner you get the starter to that mark, the sooner you get to the worst pitchers on the team and the higher chance you have of scoring more runs (and thus winning)

     c)  Taking pitches forces the pitcher to give you better pitches to hit.  The better the count is in the hitters favor, the better the likely outcome of not producing an out.  Everyone knows this already, but it still bears pointing out.  3-1 is a better place to be in while batting than 1-2.  Guys who walk more get to 3-1, 2-0, 2-1 more than other hitters because they consistantly force the pitcher to make good pitches over the plate.

     d)  The more you walk, the more often you are going to hit mistakes rather than swing at the pitches the pitcher wants you to swing at.  This is all about selectivity and getting pitches you can drive rather than just making contact and hoping for the best.  Line drives give more hits on average than grounders which on average give more hits than flyouts.  Selectivity why Pujols or Prince Fielder seem to crush the ball when they hit it (and thus produce Homeruns by the truckload)

Pitching


Moneyball spent an entire chapter discussing Chad Bradford.  While most people think this is about as exciting as watching paint dry, it had a very key concept written into it that most people either fail to understand, or don't ever even consider:

On average, a pitcher cannot control where the ball ends up if he allows a ball to be hit into play.

In other words, hits other than homeruns are at least partially luck based.  This is the only conclusion that makes sense when you look at the careers of some of the most widely known pitchers who have had some of the widest variance in ERAs.  Greg Maddox went from 2.22 to 3.57 from '98 to '99.  From 2.62 to 3.96 from '02-'03.  Why did this happen?  Luck.

1998: 251 innings, 201 hits, 13 HR given up 45 walks, 204 Ks
1999: 219 innings, 258 hits, 16 HR, 37 walks, 136 Ks

Same pitcher, different outcomes, nothing noticeably different about mechanics.  This is the kind of thing that led to two stats being created that I will discuss about in later posts.  BABIP and FIP(xFIP).

To put them simply for now, BABIP is Batting Average on Balls In Play.   Basically over a long enough sample size batters will hit approximately .300 against every pitcher on balls that are put into play (all plays in fair territory that are not homeruns).  Both hitters and pitchers can be counted the same way using the same numbers without correcting for any park factors.  If a pitcher is getting lucky (Like Maddox in 1998) his BABIP will be significantly lower than .300 until he has his eventual correction (1999) which brings his career average closer to the .300 mark.

BABIP doesn't tell the whole story because it takes the defense behind the pitcher into account.  This is where FIP was created (they didn't call it FIP in the book if i remember correctly but it was basically FIP).  FIP is Fielding Independent Pitching.  It takes everything about fielding out of the equation and gives you the stats that the pitcher can control.  Meaning Walks (non intentional including HBP), Ks, and Homeruns and puts it on an ERA scale.

In that season Billy Beane used FIP to secure Chad Bradford, basically a quad-A player who the White Sox didn't trust to be in their bullpen because he didn't throw a fastball in the 90's and had a weird delivery (submariner).  Even though Bradford allowed very few homeruns because of his delivery (in the Pacific Coast League in the worst pitching park in AAA) and walked few batters, the White Sox did not trust the results he was getting.

Player Evaluation


Billy Beane has never had a lot of money to throw around at players like other teams (Yankees, Red Sox, now the Angels and Rangers for instance) so he had to construct his teams utilizing concepts that other teams valued very little.  Namely in hitters, walks and in pitchers FIP.  Think of who the big name high dollar players are and what they bring to teams.  ARod speed and power.  Giambi (at the time) pure power player.  What teams undervalued was not getting outs for hitters and teams overrated ERA stats and throwing velocity in pitchers (particularly high school pitchers).

A lot of Billy Beane's draft picks over the years have been not so good, and people will look at that as a damnation of his method of drafting, but fail to recognize how cheaply he does it and how seemingly random drafts are anyways.  When you pick a player out of high school or college you have very little to actually go on besides what his statistics tell you for how good of a player he is going to be, that is the Billy Beane, baseball player story.  "natural talent" is random, statistics are definite.





Well that's all I can think of for now on what to write about my takeaways from Moneyball for now.  If I think of anything else I'll write either an addendum or a separate blog entirely about a different concept.