Glossary: Sabermetric
View Glossary Entries by
A pitcher's average on batted balls ending a plate appearance, excluding home runs. Based on the research of Voros McCracken and others, BABIP is mostly a function of a pitcher's defense and luck, rather than persistent skill. Thus, pitchers with abnormally high or low BABIPs are good bets to see their performances regress to the mean. The league average for modern pitcher BABIP is around .300.
Hitter BABIP is much more of a skill, based on how well they are able to hit and place the ball, along with their speed.
Equation:
BABIP = (H  HR) / (AB  K  HR + SF + SH)
Refers to the Baserunner State combined with the number of
outs in the current half inning, used to calculate Win Expectancy. For example, '2103' indicates two outs with runners on first and third.
Click here for a complete list of DRArelated articles.
DRA 2016
(Start here for three articles describing the new 2016 version of DRA: DRA 2016)
Where we once had one linear mixed model, we now have 24. Five of them are for hits (home runs, triples, doubles, infield singles, and outfield singles). Four are for notinplay events (unintentional walks, intentional walks, hitbatsmen, and strikeouts. For singleout plays, we separately modeled putouts at each position, which makes for 9 more models. And finally, we modeled doubleplays that began with 1 of the 6 infielders.
But how do you find the best predictors for each model? Last year we used the Akaike Information Criteria (AIC) and likelihoodratio tests to find the best combinations. These are timehonored methods for evaluating mixed models, but they also test only insample data and make their own assumptions about how the data might be organized.
This year we decided to step up our game in two ways. First, while we still kept an eye on AIC and likelihoodratio tests, we also moved to using 10fold crossvalidation, employing a random sample of half of the 2015 season as our testing ground. To test out of sample, we moved to using the Receiver Operating Characteristic Curve, commonly described in machine learning as the Area Under the Curve or AUC. Pioneered by WorldWar II radar operators, and then extended to statistics and other fields, AUC measures a model by the likelihood of returning false positives versus false negatives. A worthless score is .5, indicating that the outcomes of your model are essentially a coin toss. A perfect score is 1. Given the amount of random variation in baseball, our goal was to exceed at least .6 for each modeled event, and some models exceeded that threshold by a substantial amount.
(See these articles from 2015 for a history of DRA: http://bbp.cx/a/26195 and In Depth article)
DRAMinus ("DRA–")
As noted above, we've received multiple requests for a "minus" version of DRA, something that rates pitchers by how well they compared to their peers rather than by an amount of predicted runs allowed in a given season. Knowledgeable baseball fans are familiar with statistics like this. Common examples include wRC+ and ERA. The idea is to put an average player for each season at 100, and then rate players by how much they vary from the average. By rating every pitcher by how good (or poor) he was by comparison to his peers, we can make fairer comparisons across different seasons and different eras. These comparisons aren't perfect: We can't make baseball 50 years ago more diverse or force today's players to endure the conditions of 50 years ago, but metrics like DRA– allow comparisons of pitchers across seasons and eras to be much more meaningful.
Unlike cFIP (which measures true talent), DRA– (which measures true talent plus luck) will not have a forced standard deviation. The two numbers (which are otherwise both scaled to 100) can still be compared, but be mindful of that distinction. For both cFIP and DRA–, lower is better.
See: http://www.baseballprospectus.com/article.php?articleid=26613
Type the definition of the term here, or leave the text as it is if you don't want to add a new term.
Type the definition of the term here, or leave the text as it is if you don't want to add a new term.
Def Eff, or Defensive Efficiency, is the rate at which balls put into play are converted into outs by a team's defense. Def Eff can be approximated with (1  BABIP), if all you have is BABIP, but a team's actual Def Eff is computed with
1  ( H  HR ) / ( AB  SO  HR + SH + SF )
the Team Audit Standings use the latter formula.
Alternately, it will be seen some places (including some past editions of Baseball Prospectus) computed as 1  ((H + ROE  HR) / (PA  BB  SO  HBP  HR)), which comes out slightly lower in most instances.
Here is an example of the Defensive Efficiency spectrum based on the 2011 season:
Excellent  Tampa Bay .735
Great  Texas .722
Average  Toronto .710
Poor  Pittsburgh .700
Horrendous  Minnesota .693
(from http://bbp.cx/a/26195)
Under baseball’s scoring rules, a wild pitch is assigned when a pitcher throws a pitch that is deemed too difficult for a catcher to control with ordinary effort, thereby allowing a baserunner (including a batter, on a third strike) to advance a base. A passed ball is assigned when a pitcher throws a pitch that a catcher ought to have controlled with ordinary effort, but which nonetheless gets away, also allowing a baserunner to move up a base. The difference between a wild pitch and a passed ball, like that of the “earned” run, is at the discretion of the official scorer. Because there can be inconsistency in applying these categories, we prefer to consider them together.
Last year, Dan Brooks and Harry Pavlidis introduced a regressed probabilistic model that combined Harry’s pitch classifications from PitchInfo with a With or Without You (WOWY) approach. RPMWOWY measured pitchers and catchers on the number and quality of passed balls or wild pitches (PBWP) experienced while they were involved in the game.
Not surprisingly, we have updated this approach to a mixed model as well. Unfortunately, Passed Balls or Wild Pitches Above Average would be quite a mouthful. Again, we’re trying out a new term to see if it is easier to communicate these concepts. We’re going to call these events Errant Pitches. The statistic that compares pitchers and catchers in these events is called Errant Pitches Above Average, or EPAA.
Unfortunately, the mixed model only works for us from 2008 forward, which is when PITCHf/x data became available. Before that time, we will rely solely on WOWY to measure PBWP, which is when pitch counts were first tracked officially. For the time being, we won’t calculate EPAA before 1988 at all, and it will not play a role in calculating pitcher DRA for those seasons.
But, from 2008 through 2014, and going forward, here are the factors that EPAA considers:
 The identity of the pitcher;
 The identity of the catcher;
 The likelihood of the pitch being an Errant Pitch, based on location and type of pitch, courtesy of PitchInfo classifications.
Errant Pitches, as you can see, has a much smaller list of relevant factors than our other statistics.
In 2014, the pitchers with the best (most negative) EPAA scores were:
And the pitchers our model said were most likely to generate a troublesome pitch were:
t want to add a new term.
FIP is a component ERA inspired by the work of Voros McCracken on defenseindependent pitching statistics, but has become more widely used because of the ease of computation  it requires only four easilyfound box score stats, uses only basic arithmetic operations and has four easilymemorized constants. It was conceived of by both Tom Tango and Clay Dreslough, the latter of who called it DefenseIndependent Component ERA.
At Prospectus, we are including hit batters in the walks term. The constant we use is both league and season specific  in other words, a pitcher in the American League will have a different FIP constant than a pitcher in the National League. This differs from the presentation of FIP on sites such as Fangraphs, which use one constant for both leagues in each season.
Here is an example of the Fielding Independent Pitching spectrum based on the 2011 season:
Excellent  Roy Halladay 2.17
Great  David Price 3.36
Average  Tim Stauffer 4.00
Poor  Carlos Zambrano 4.56
Horrendous  Bronson Arroyo 5.68
Offensive Winning Percentage. A Bill James stat, usually derived from runs created. In EqA terms, it could be calculated as (EQA/refEQA)^5, where refEQA is some reference EQA, such as league average (always .260) or the positionaveraged EQA.
For Pythagenpat, the exponent X = ((rs + ra)/g)^.285. Although there is some wiggle room for disagreement in the exponent, that equation is simpler, more elegant, and gets the better answer over a wider range of runs scored than Pythagenport, including the mandatory value of 1 at 1 rpg. See here for more.
Type the definition of the term here, or leave the text as it is if you don't want to add a new term.
Speed Score (SPD) is one of five primary production metrics used by PECOTA in identifying a hitter's comparables. It is based in principle on the Bill James speed score and includes five components: Stolen base percentage, stolen base attempts as a percentage of opportunities, triples, double plays grounded into as a percentage of opportunities, and runs scored as a percentage of times on base.
Beginning in 2006, BP has developed a proprietary version of Speed Score that takes better advantage of playbyplay data and ensures that equal weight is given to the five components. In the BP formulation of Speed Score, an average rating is exactly 5.0. The highest and lowest possible scores are 10.0 and 0.0, respectively, but in practice most players fall within the boundary between 7.0 (very fast) and 3.0 (very slow).
(from http://bbp.cx/a/26195)
Our hypothesis is that basestealing attempts are connected with the pitcher’s ability to hold runners. When baserunners are not afraid of a pitcher, they will take more steps off the bag. Baserunners who are further off the bag are more likely to beat a force out, more likely to break up a double play if they can’t beat a force out, and more likely to take the extra base if the batter gets a hit.
Takeoff Rate stats consider the following factors:
 The inning in which the basestealing attempt was made;
 The run difference between the two teams at the time;
 The stadium where the game takes place;
 The underlying quality of the pitcher, as measured by Jonathan Judge’s cFIP statistic;
 The SRAA of the lead runner;
 The number of runners on base;
 The number of outs in the inning;
 The pitcher involved;
 The batter involved;
 The catcher involved;
 The identity of the hitter on deck;
 Whether the pitcher started the game or is a reliever.
Takeoff Rate Above Average is also scaled to zero, and negative numbers are once again better for the pitcher than positive numbers. By TRAA, here were the pitchers who worried baserunners the most in 2014.
And here were the pitchers who emboldened baserunners in 2014:
Unintentional Walk Rate (BB) is one of five primary production metrics used by PECOTA in identifying a player's comparables. It is defined as (BBIBB)/PA.
The probability of winning the current game, given some
information about how many runs each team has scored to a certain point in the game, how many outs there are, whether there are runners on base, and the strength of each team. Keith Woolner outlined a method for computing Win Expectancy given all of these parameters in BP 2005.
A correction made to raw runs when converting them to a standard league to preserve their win value. Define an average team from season games played, league runs per game (9 innings or 27 outs, depending on whether you are using pitcher or batter data), and appropriate adjustments (park, team hitting/pitching, difficulty). "Team" is the effect of replacing one player on the average team with the player we are analyzing. Calculate the pythagorean exponent from (average + team) / games as your RPG entry; calculate winning percentage using the modified pythagorean formula. Now, go backwards, solve for "team" runs, given the winning percentage, an average team that scores 4.5 per game, and a pythagorean exponent of 2.00.
(from http://www.hardballtimes.com/fipincontext/)
Because cFIP is on a 100 “minus” scale, 100 is perfectly average, scores below 100 are better, and scores above 100 are worse. Because cFIP has a forced standard deviation of 15, we can divide the pitchers into general and consistent categories of quality. Here is how that divides up for the 2014 season, with some representative examples:
Representative Examples, 2014 Season 

cFIP Range 
Z Score 
Pitcher Quality 
Examples 
<70 
<2 
Superb 
Aroldis Chapman (36/best), Sean Doolittle (49), Clayton Kershaw (57), Chris Sale (63) 
70–85 
<1 
Great 
Zach Duke (72), Jon Lester (75), Mark Melancon (75), Zack Greinke (82) 
85–95 
<.33 
Above Avg. 
Hyunjin Ryu (87), Francisco Rodriguez (88), Johnny Cueto (89), Joba Chamberlain (90) 
95–105 
.33 < 0 < +.33 
Average 
Tyson Ross (95), Sonny Gray (96), Matt Barnes(99), Brad Ziegler (104) 
105–115 
>.33 
Below Avg. 
Brian Wilson (106), Tanner Roark (107), Nick Greenwood (111), Ubaldo Jimenez (112) 
115–130 
>1 
Bad 
Edwin Jackson (116), Jim Johnson (120), Kyle Kendrick (124), Aaron Crow (125) 
130+ 
>2 
Awful 
Brad Penny (130), Paul Maholm (131), Mike Pelfrey (132/worst), Anthony Ranaudo (132/worst) 
