Baseball stats get even more specific
The statistical revolution came to baseball long ago, and these
days it’s as much a part of the game as cowhide, sunflower seeds
and functionally illiterate Bill Plaschke columns. And despite much
fretting to the contrary, the statistical revolution is nothing to
fear, loathe or run from.
enhance baseball. Any baseball argument is informed by the
numbers (even if you’re using the wrong ones), and any
understanding of this great game is made better and deeper by an
understanding of the numbers.
So, with those dearly held principles in mind, let’s take a
walking tour of some of the more illuminating stats that the
revolution has produced.
BABIP (Batting Average on Balls in Play)
A conceptual knowledge of BABIP is essential to, among other
things, projecting the performances of pitchers. Broadly speaking,
a pitcher exerts the most control over whether a batter walks,
makes contact, hits the ball in or out of the park and hits the
ball on the ground or in the air. What becomes of a fair batted
ball that isn’t a home run, however, is largely a function of luck
Most often, a pitcher’s BABIP will fall somewhere between
.290 and .300. If his BABIP deviates wildly from this (and his
career norms don’t justify such deviation), then you can probably
expect him to return to the .290-.300 range the next season.
By extension, when forecasting pitcher performance, it’s best
to look at his strikeout rate, walk rate, home run rate and
groundball-to-fly ball ratio, then make separate estimations of the
defense that’ll be supporting him. A traditional pitching measure
like ERA, of course, misses such vital nuances.
When it comes to grasping modern statistical analysis in
baseball, there’s no getting around the idea of “replacement
level.” Replacement level is, for many useful measures, the
baseline to which all players are compared. The name says it all
Imagine that a team under tight budget and talent constraints
loses a regular contributor to injury or surprise retirement or
some other unforeseen phenomenon. To replace him, they’d turn to a
prospect who’s in need of further seasoning. Or they’d sign a
scrap-heap free agent. Or they’d pick up someone off the waiver
wire. Or they’d trade for a B-lister who’d require little in
All of that is the essence of the replacement player —
what you could summon up on short notice and with few resources.
The replacement player’s level of performance varies from year to
year, but it’s always measurably worse than the league average.
That’s the case whether it’s a replacement-level position player or
The concept is important because it’s not the hypothetical
average performer to which players should be compared; it’s the
contingency plan — the replacement level — against
which they should be measured.
Think of this as a souped-up version of that old standard,
ERA+ is simply ERA adjusted for the effects of the pitcher’s
home park and the league environment. That is, ERA+, unlike plain
old ERA, accounts for the fact that pitching in, say, Dodger
Stadium circa 1968 is ludicrously easier than pitching in Coors
Field circa 2000.
Also, ERA+ is scaled to 100, which means that an ERA+ of 100
equals a league-average ERA after you adjust for park effects.
Anything more than 100 is that many percentage points better than
the league average (again, adjusted for the pitcher’s home yard),
and anything less than 100 is that many percentage points worse
than the league average.
with a mark of 202, which means his career ERA is, on a
park-adjusted basis, 102% better than the league average (wow). On
the downside, Brad Lidge in 2009 posted an ERA+ of 59, which means
his ERA was 41% worse than the league mean after correcting for his
Here’s where BABIP comes into play for the pitcher. FIP
stands for “Fielding-Independent Pitching,” and, as the name
suggests, it attempts to decouple pitching and defense. It takes
into account strikeouts, unintentional walks, hit batsmen and home
Also, FIP yields a number that
looks like ERA, so it’s in a statistical language the
everyday fan can understand. It’s useful in that it allows you to
evaluate a pitcher in a vacuum, without the good fortune or
misfortune of his defensive help getting in the way.
You’ve probably encountered OPS, which is shorthand for
on-base percentage plus slugging percentage. As quick-and-dirty
measures go, OPS is a good one, but OPS+ constitutes an
Like ERA+, OPS+ takes its foundational metric (OPS in this
instance), adjusts it based on park and league conditions and
scales the final result to 100. Also like ERA+, OPS+ is a useful
tool when you want to eyeball production and compare across years
and even eras.
For comparison’s sake … the top career OPS+ is
href="http://www.baseball-reference.com/leaders/onbase_plus_slugging_plus_career.shtml">Babe Ruth, 207
Ruth, 207. Albert Pujols’ OPS+ in 2009 was 188.
(Value Over Replacement Player)
Simply put, VORP tells you how many more runs a hitter
produces relative to a replacement-level hitter. This means “runs”
in the theoretical sense — i.e., how he contributes to the
scoring of runs with his homers, doubles, singles, stolen bases,
VORP also compares hitters to their positional peers. For
instance, shortstop is a more demanding position than third base,
which is a more demanding position than right field. And so on.
When comparing players across positions, you
must take into account positional scarcity. In the
circumscribed universe of Major League Baseball, it’s easy to find
a first baseman who can hit. It’s manifestly difficult to find a
catcher who can hit. VORP reflects this fact.
(Ultimate Zone Rating)
The problem with fielding errors (and, by extension, fielding
percentage) is that they make little account for range. That is,
they reward or penalize fielders for making (or not making) the
routine plays, but they don’t consider the batted balls that were
never reached in the first place. In other words, you can’t make an
error on a ball you never got to, but the ball you never got to
To cite an extreme hypothetical, you field a soft grounder
straight at you and so does 1985 Ozzie Smith. Then 1985 Ozzie Smith
ranges deep in the hole after a sharp bounder, splays his arms out,
dives as far as he’s able, snags the ball in the last fraction of
webbing and nicks the runner at first. You … barely even lean
in the direction of the ball as it zips past you for a base hit.
According to traditional fielding measures, you and 1985 Ozzie
Smith are equally adept glove men.
On top of that deficiency is the fact that the scoring of
errors is a highly subjective process, and it’s also one as prone
to homerism and “superstar treatment” as anything an NBA ref could
come up with. And that brings us to UZR.
UZR makes note of errors, and it also, by dividing the entire
field up into 64 different fielding “zones,” evaluates a fielder’s
range. It does all this by calculating what percentage of balls hit
into a fielder’s zone are converted into outs and then making a
host of contextual adjustments (e.g., park, ball speed, pitcher
tendencies, etc.). Then that figure is compared with the league
average for the position.
You need at least three years’ worth of UZR data to make
sensible valuations of fielders, but even with those limitations
UZR is substantially better than the numbers usually bandied about
in discussions of defense.
(Defensive Efficiency Rating)
DER is a wonderfully simple statistic (wonderfully simple in
conception, anyway) that gives you a nifty snapshot of team
defense. Simply put, it’s the percentage of balls in play (i.e.,
fair balls that aren’t home runs) a team’s defense turns into outs.
Home parks and other conditions can create some noise in the
numbers, but it’s still measurably better than judging teams based
on error totals or fielding percentage.
(Win Probability Added)
This handy little widget tells you how a hitter or pitcher
changes his team’s chances with any given play. In other words,
it’s what a clutch statistic
should be. Unlike flawed stats like game-winning RBI or
saves, WPA accurately gauges how crucial a situation is and then
determines how the player changes the complexion of the game with
his performance at that given moment. It’s expressed as a
percentage. In the case of a positive figure, the player increased
his team’s chances of winning by that much. In the case of a
negative number, he harmed his team’s chances by that much.
For instance, when Bucky Dent homered for the Yankees in the
top of the seventh in the one-game playoff with Boston in 1978, he
increased their chances of winning by 46 percent. Thus, his WPA on
that one swing was 46 percent.
Total player value
WAR (Wins Above Replacement)
Here we have a stat that provides a complete picture of
player value. For hitters, WAR for the most part takes into account
on-base percentage and a modified version of slugging percentage,
with OBP weighted more heavily in the sausage-making process. As
well, WAR uses UZR to evaluate the defensive contributions (or lack
thereof) of those same hitters. The output, then, is an expression
of what a player contributes, in wins, over and above the
aforementioned replacement level, to his team.
WAR also works for pitchers. In their instances, the WAR
calculus includes FIP, the replacement level and home and league
href="http://www.fangraphs.com/leaders.aspx?pos=all&stats=bat&lg=all&qual=y&type=6&season=2009&month=0" target="_blank">Ben Zobrist (!) led all position players with a WAR of 8.6
target="_blank">Ben Zobrist (!) led all position players with a
WAR of 8.6, and Zack Greinke led all pitchers with a WAR of
So, viva la revolucion and all that.