All products featured on WIRED are independently selected by our editors. However, we may receive compensation from retailers and/or from purchases of products through these links.
Our first GeekDad Sports fantasy baseball league is less than a week away from the first pitch. Twenty ownership teams of parents and kids are in the process of selecting their players, a prerequisite for submitting starting lineups for the 15-week season. Our rules of engagement specify a set of nine core statistics we'll use to determine weekly game results.
"It really is a hard mathematical problem, how to pick Rotisserie," says Wayne L. Winston, a professor of Decision Sciences at Indiana University's Kelley School of Business and the author of Mathletics. Along with his business partner and longtime friend Jeff Sagarin, Winston developed a statistics tracking and rating system (adjusted plus-minus) used by the Dallas Mavericks basketball team.
Winston says his hoops formula isn't perfect, due to the nature of the game. "If Kobe Bryant steals a ball and gets a dunk, the other guys on the floor get credit for it. Over the course of time, that should average out. The best thing about baseball, though, is that it's pitcher-hitter, with fielding layered in. That's why the math of baseball works so much better. It's an individual sport."
For Winston, win probability added is the best measure of a player's value. Positive contributions by a player increase the chance of his team winning the game. Before Bobby Thomson's famous home run, for instance, the Giants may have had a 30 percent chance of winning. After he hit the dinger, the odds became 100 percent, or a net gain of +0.7 for Thomson. Ralph Branca, on the other hand, generated -0.7 for the same pitch.
"You tell your kids when they play sports that they shouldn't care what they score. They should care how they helped the team," argues Winston. "These are the ultimate measures of how they helped the team."
Statistics are the foundation of a good fantasy league experience. While there are a number of complicated formulas that can tell us how good a player is, the trade-off for that level of precision is that the numbers become too abstract to be counted when sitting in the stands. How we keep score in the fake games should be as understandable as what we see in a real game.
In the GDS league, we have structured the season to pit teams against each other each week. The head-to-head battles span nine statistical categories, or "innings." These innings are won or lost based on the performance of the dozen players selected to start each week, with the better team scoring a fantasy run. After tallying the runs for these nine innings, a winner is declared, or a tie score is broken by comparing additional statistics beyond the initial nine.
Since we want the league to allow a variety of roster-building strategies to be successful, it is vital that the core stats be distinct enough to allow owners to create championships from speedy slap hitters or slow-footed sluggers, with workhorse starters or a deep and dominant bullpen. What follows is a rationale for the decisions our league made about which nine stats to track.
Pitching constitutes half our starting lineup and four of the nine fantasy innings. The statistics pitchers generate can be broadly separated into four groups: Production (includes metrics like appearances, innings pitched, wins and losses), Relief (saves, blown saves and holds), Effectiveness (earned runs, quality starts, walks and hits allowed), and Dominance (strikeouts, complete games and shutouts).
These groups reflect the different strengths of a pitcher's repertoire. No pitcher will be able to be great at all four categories, mainly because of the distinction between starting pitchers and bullpen pitchers. One can be dominant without being effective, and vice versa. One can be productive and look horrible doing it. For our league, we wanted to pick statistics to represent each of these groups while still being easy to follow in the stands or the game boxscores.
Wins + Quality Starts
The most important stat to a baseball team is the Win (W). At the end of the season, it doesn't matter how the club managed to get each win, as long as it accumulates a sufficient number of them to keep playing in October. Every game is guaranteed to assign a win to someone, and pitchers are the ones who get the statistical credit.
Wins are one way to tell if a pitcher is having a productive season, but even a poor performance can earn a W. That is why in 1985 Philadelphia sportswriter John Lowe came up with the Quality Start (QS): a starting pitcher completing at least six innings and allowing no more than three earned runs. There are several examples of great games not garnering a quality start (and quality starts awarded for mediocre performances), but overall this stat tends to reflect a stellar game. In 1992, David Smith calculated the average ERA for quality starts is 1.91, considerably lower than non-quality starts.
By tracking these two stats together, quality wins are worth double the cheap ones. This might make Sandy Koufax happy. In his final season — in which he went 27-9 with a 1.73 ERA — Koufax dubbed teammate Phil Regan "The Vulture" for picking up 14 wins in relief.
Saves
Our league has a requirement to include at least one reliever in the starting lineup. With a 20-team league, that most often will be a bonafide closer, someone whose main contribution to the game is the Save (SV). A good closer will get 2-3 saves in a week of MLB games.
"I hate saves, but you have to do it or relievers don't mean anything [in fantasy leagues]," says Winston.
The Hold (HLD) was invented in 1986 by John Dewan and Mike O'Donnell to give middle relievers credit for their contribution to a win. The Hold is not an official statistic tracked by MLB, however, and the criteria to qualify varies from one service to the next. The inclusion of this stat could help deepen the bullpen options for a fantasy league and add value for quality setup players like Mike Stanton (266 career holds, the career leader). The top setup pitchers are awarded 30 holds a year, compared to 40 saves for the best closers. That could give too much depth to a position that doesn't need to be diluted. For the GDS, we're sticking with saves.
WHIP
One interesting baseball statistic — Batting Average on Balls In Play (BABIP) — describes the percentage of the balls hit into the field that become hits. According to Winston, BABIP appears random from one season to the next, an indication there is a lot of luck involved when a pitcher has a good year or a bad one.
"The way to evaluate pitchers, because there is a lot of luck involved in pitching, is the defense-independent statistics," explains Winston. "These are walks, strikeouts and home runs allowed. It's the best predictor of future success of a pitcher."
That is why the GDS went looking for a replacement for Earned Run Average (ERA), a cornerstone stat for pitchers since 1912. Sabermetricians tend to consider Walks plus Hits Per Innings Pitched (WHIP) as the single most important statistic for evaluating a pitcher's effectiveness. WHIP incorporates the first of those three defense-independent statistics, the walk, and it considers each pitcher-hitter outcome, even those coming after an error erased the third out (when ERA stops tracking runs). Pedro Martinez (0.737 in 2000) boasts the best single-season WHIP ever recorded, while Cleveland great Addie Joss has the best career WHIP (0.968). WHIP also wins out over the ERA because it is easier to calculate.
Strikeouts
The Strikeout (K), the next defense-independent statistic, is one of the most exciting outcomes for an at bat. There is something visceral about watching a pitcher throw the ball hard (or expertly) enough to force a strongman to swing and miss. One of Kerry Wood's 1998 rookie wins netted a record-tying 20 strikeouts and remains my favorite game to date. Crash Davis may think strikeouts are fascist, but the stat is easy to understand and count, making them exciting for fans to watch.
Strikeout artists are perceived as dominating pitchers. Other stats that share that cachet, but they occur too infrequently to be of much value in a weekly head-to-head league like the GDS. Home Runs Allowed (HRA), the third of the defense-independent stats, equates to about one-third of the total walks given up by major league pitchers and one-tenth the number of hits. As a result of its relative infrequency, HRA was relegated to an extra-inning tie-breaker category for us.
Two other metrics to consider are complete games (CG) and shutouts (SO). A century ago, only one game in nine did not end in a complete game. By the 1950s, the rate of complete games was down to about one-third of all starts. Tony LaRussa's use of his bullpen in the 1980s — which includes bringing in pitchers to throw just a couple pitches — combined with greater attention to arm abuse started a downward trend that now sees pitchers finish what they started about 3 percent of the time. In 2011, there were only 173 complete games out of over 5,200 opportunities.
Shutouts are even more sparse from week to week. The all-time career leader is Walter Johnson, a Senators pitcher from 1907-1927, who finished with 110 shutouts. Grover Alexander had a whopping 16 shutouts in 1876. Current leader Roy Halladay barely has more shutouts than seasons pitched. Although the shutout appears on the rise, it just isn't enough to be considered a practical statistic for weekly fantasy leagues.
As with pitching, hitting statistics will aggregate into a few general groups: Contact (includes metrics like batting average, hits, walks and strikeouts), Power (home runs, slugging percentage, and total bases) and Clutch Hitting (runs batted in and even at bats). For those lucky enough to get on base, a fourth group — Baserunning — has metrics like stolen bases and runs scored to help evaluate a player's ability to do something after the at bat.
These groups reflect the various strengths of a player's offensive repertoire. The most valuable players are the ones who can excel in all four areas, but rare is the hitter who can be the best at everything. Only thirteen players in baseball history have won the Triple Crown, by leading or being tied for first in batting average, RBIs and home runs. The last Triple Crown winner was Carl Yastrzemski in 1967, and the irascible Ty Cobb remains the lone Quadruple Crown winner by virtue of also stealing a league-best 76 bases in 1909.
The hitter equivalent of WHIP may be Runs Created (RC). It is an attempt by Bill James to answer the question of how production at the plate translates into the runs that win games. Although the technical formula can be quite involved, the idea generalizes to the product of a hitter's ability to get on base and hit for power over the number of chances he gets to make contact. Tom Tango, author of The Book, updated the stat to turn it into a comparative measure against league average: Weighted Runs Created (wRC). ("He's really smart," admires Winston, "smarter than Bill James, I think.") RC has been shown to be quite accurate in predicting a team's run total, and therefore is considered a good indicator of individual contribution.
There are two practical problems with Runs Created. First, it's all formula. It would be challenging for the average fan to watch a game of baseball and track RC. Second, like its cousin On-Base Plus Slugging (OBS), it is too comprehensive for a fantasy baseball league that wants to separate different aspects of the game. Combining contact and power may be a great way to evaluate talent, but we are looking to compare those kinds of stats separately.
On-Base Percentage
Many sports stats junkies have come to dislike the uber-stat OBS because its formula equates On-Base Percentage (OBP) with Slugging Percentage (SLG), despite the former being almost twice as important to a team's run production. OBP takes into account hits, walks and times hit by a pitch to calculate the rate at which the hitter contributes a baserunner for the team. OBP gets the nod over Batting Average (BA) as a core fantasy league stat because it isn't only about contact, but also reflects the patience of a hitter to draw walks and his fearlessness when crowding the plate.
OBP will reward power hitters, too, who draw intentional walks for fear of giving up home runs. Even more than batting average, on-base percentage offers a wide range of values for players, making it a great stat to help differentiate fantasy team performances.
"I like on base and slugging," says Winston. "They became so popular with Moneyball. You can find those stats on any baseball card now."
Home Runs
Slugging is calculated by dividing total bases (worth one for singles, four for homers) by the number of at bats, simple math that isn't difficult to figure out from the bleachers. "Slugging percentage is a good thing to teach kids," says Winston.
The only other time in history when slugging spiked as high as it is now was during the heyday of Babe Ruth. As power numbers steadily increased since the 1960s, the league's overall slugging percentage has plateaued around .425. How much steroids or contemporary ballparks have contributed to these most recent power numbers, only time will tell. The biggest knock against using SLG is that it wouldn't win a popular vote.
Ever since the Babe called his shot, the common fan considers the home run to be the promise of every at bat. The iconic sign of a power hitter is his ability to knock a pitched ball out of the park, and it takes a special kind of player to do that consistently. Even as the home run is becoming more frequent — for the past decade and a half, major league teams averaged just over a homer a game — the league leaders are still typically in the 40s and 50s.
The exception: Four years between that wonderful and infamous 1998 season and 2001, when Barry Bonds hit 73. During that span, Roger Maris' 37-year-old record was broken a half dozen times by three players (Bonds, Mark McGwire, and Sammy Sosa). 1998 set a MLB attendance record of 70,601,147 fans, starting a surge that grew to 79,447,312 in 2007. At the end of that year, the Mitchell Report was released detailing abuse of performance enhancing drugs over that time. Attendance took a dive as a result of the scandal, but the increased presence of the homer clearly was a hit with fans.
"You've got to do home runs, because that's what people care about," admits Winston.
Stolen Bases
During periods of increased power in baseball history, stolen base totals decrease. In baseball, a baserunner can advance to an open base without the ball being hit. If successful, a Stolen Base (SB) is awarded; if not, the runner is charged with a Caught Stealing (CS). There are even some situations where an official scorer can declare "defensive indifference" and not give credit for the steal. Steals are often the mark of quick players who tend to lack power.
"Stolen bases in real life doesn't mean that much," says Winston. "It makes it more fun to draft, because you can take that fast guy."
There are other measures of ability on the basepaths, but each brings its own headache. Incorporating caught stealing into the equation as a penalty could bring up the possibility that there would be a negative total at the end of the week (My own preference in fantasy leagues is to deal only with positive numbers). Using a percentage of steals per attempts would bias in favor of those who don't attempt many. Since statistical studies have shown that a runner needs to be successful at least two-thirds of the time to avoid hurting his team, sabermetricians came up with Stolen Base Runs (SBR) as another attempt to equate running to scoring. Each successful steal is worth .3 runs while being caught stealing takes away .6 runs. While accurate, SBR is too abstract. In the GDS, we stay simple and count steals.
RBIs
Runs Batted In (RBI) gives credit to the batter whose at-bat results in a run being scored (double-plays and errors excluded). The stat has been around since the earliest days of the National League and is one-third of the coveted Triple Crown. RBI totals are often visible factors in determining annual MVP awards and Hall-of-Fame consideration. The HoF Standards formula doesn't start giving credit for RBIs until a player reaches at least 900 in his career. Harold Baines has the most career RBIs for someone eligible but not (yet) in the Hall of Fame, with 1,628.
RBIs can also serve to prop up below-average hitters (Joe Carter comes to mind) and might be better considered as a percentage or WAR (Wins Over Replacement) kind of stat. Despite criticism that the RBI is too dependent on the lineup position and teammates surrounding a player — a batter with runners on base has more opportunities to drive in runs — it remains an easily comprehended and reasonable indicator of clutch hitting.
That brings the GDS to eight core statistics, four categories for pitchers and four for hitters.
Since this is a baseball league, a ninth "inning" would be aesthetically pleasing. It also reduces the likelihood of tied games each week. Ideally, the ninth and final core statistic for our fantasy league scoring would reflect defense. Fielding, however, is the most subjective area of baseball.
The error (E) is a stat that keeps track of fielding mistakes, like bobbled or poorly thrown balls. It is the most common of several judgment calls made by a game's official scorer, the designated recorder of everything that took place in the game. What constitues an error in the boxscore depends on who is watching. Two different official scorers could judge the same fielding attempt and come to opposite conclusions about charging a player with an error.
Errors are controversial for other reasons. It is easy for a poor fielder with limited range to avoid being charged with an error that a quicker or better positioned fielder would receive. Errors are contextual to the position, as well. Shortstops and third basemen are more likely to be error prone due to more fielding opportunities and longer throws to first. Errors can be as rare as triples, making it an impractical choice to become a core fantasy statistic.
If readily available, Assists and Putouts might be a consideration. They may not reflect the true quality of fielder, but these stats do translate to defensive time spent touching the ball. As with errors, these are biased toward certain positions. First basemen and catchers tally significantly more putouts than their third-base teammates. Assists are credited about once for every three putouts, with the middle infield and third basemen earning the lion's share through ground outs. The practical problem with fielding stats is that there is more difference between the positions people play than the top players at each position.
According to Winston, the best stat for defense is the Ultimate Zone Rating (UZR). For every ball hit into play, the fielder gets points based on whether his position is likely to get to the ball and if he was successful fielding it. These base metrics are converted into runs, which in turn can be converted into wins.
"That will make kids appreciate that the traditional metric is really stupid," says Winston.
Winston singled out Derek Jeter, who won the Gold Glove at his position last year, as a lousy fielder. Jeter's UZR was -6.5, despite only making 12 errors as a shortstop. On the other hand, Brett Gardner had a 25.2 UZR for the Yankees in 2011. At 10 runs to a win, Gardner's fielding translates to an extra two-and-a-half wins for his team. "Even if he cost his team -2.5 wins as a hitter," Winston explains, "that makes him an average player. It's easy to interpret, but you don't want to do the math with 9-year-olds."
Without a reliable defensive stat to use, our league will use an additional offensive category: Runs Scored (R). Annual leaders finish between 120-130 runs per season, well shy of Ruth's modern era record of 177 in 1921 (Billy Hamilton has the actual record, 196, with the Phillies in 1894). Though teams cross home plate just under five time a game these days, it is still off the pace of the Golden Age of baseball, with Ruth and the Yankees in the 1920s.
The run is a stat that combines good base running with an ability to get on base. It doesn't exclude power hitters, either, who get a noticeable percentage of their runs by hitting the ball out of the park. Runs Scored is arguably the only stat that matters to a team, making it a good choice to include in fantasy play, too.
"Anyone who knows the math of baseball — and has a law degree to know how the salaries work — can be a GM," says Winston. We presume the law degree is optional for fantasy managers.
The inaugural 2012 GeekDad Sports fantasy baseball season is sponsored by FanStar Sports, helping to give a voice to the sports geeks in our community. Since 1991, FanStar Sports (http://www.fanstar.com) has provided services to thousands of fantasy leagues. The Arizona company provides fantasy league hosting for baseball, football, basketball, hockey and NASCAR. The administration and presentation of each league are highly customizable, able to accommodate a wide range of unique needs.
