Part 2: How to Predict Postseason Success in Baseball

Wouldn't it be nice to predict the next time your team will hoist the Commissioner's Trophy?

While Part 1 looked at driving in runs without hitting home runs, the second hypothesis has more to do with hitting the league's most elite pitchers in the postseason. Will this hypothesis lead to some statistically significant results?


Performance Against Top Pitchers

Hypothesis

Against top-line starters and relievers, it is very difficult to hit home runs, so my theory is that teams that have a more simplistic batting approach will have a better opportunity against these very good pitchers. Also, because a team is very likely to face great pitching in the postseason, I also hypothesize that teams that face good pitchers (I have categorized “top pitchers” as those who finish in the top 20 of ERA minus, or ERA-, as calculated by Fangraphs) more often and/or have more success against them (in terms of runs scored per nine innings) are more likely to have playoff success.

Results

By hand, I compiled the top 20 starting pitchers in terms of ERA- every year from 2003-2012, and then used Baseball Almanac to record every game these pitchers played against teams who made the playoffs that year. I compiled total innings, total runs scored, total games and runs scored (not just earned runs) per 9 innings for each team each year. The reasoning behind looking at all runs, and not just earned runs, was because runs of any kind are so hard to come by in the postseason, or when facing a top pitcher, and even if a run is unearned, most of the time the opposing team would still need to string together a couple of hits to allow that unearned run to score.

When I finished compiling data on team performances against top 20 pitchers, I ran individual regression analyses with PV being the outcome variable, and these new statistics being the predictors. However, no single statistic correlated to having a high PV. Even when using multiple predictors with the top 20 pitching stats, there was still no significant correlation.

Conclusion

Based on the results of my tests of these two hypotheses, I unfortunately did not find any significant regression models that could predict PV from any of these statistics, I was not hugely surprised by this outcome for a few reasons. Because I only looked at playoff teams in the past ten years (many of the statistics I used in these models were not compiled before then), my sample size was smaller than ideal to start with. Also, there is high multicollinearity among so many of these statistics. This means that it was it was difficult to interpret the individual coefficients.

Also, having too many predictors, or controlling for too many variables, makes it extremely difficult to find a model that is both significant, and that makes sense from a baseball perspective. There were a few interesting findings, such as how LDp is marginally correlated with playoff wins (but not correlated with playoff series wins), but for the most part, no major discoveries were made.

Possible Improvements

One of the changes I could have made included how I calculated the top 20 pitchers statistics. I chose the number 20 randomly, but I also compiled the top 20 pitchers regardless of league. In hindsight, I probably should have compiled the top 20 pitchers from both the American and National Leagues in each year. Also, maybe there is a better statistic than “runs per 9 innings” to gauge how well teams do against these top pitchers. Also, when my second hypothesis failed, I started to compile 28 new statistics from Fangraphs’s “high leverage situations” split. I originally tried this because essentially all playoff batting situations can be considered “high leverage.”

However, these statistics were compiled from late and close game situations, rather than ability to drive in runs without hitting home runs, which is what my two hypotheses were related to. My time might have been better spent looking at statistics with runners in scoring position. Those kinds of statistics would have been more relevant to my hypotheses, as driving in runners in scoring position is not only the most effective way to score off top pitchers, but it is also a skill that requires the batter to shorten his swing, and have a more simplistic batting approach. As I continue this research in the future, I will take into account all of these factors in my quest to find a formula for postseason success in Major League Baseball. 

Part 1: How to Predict Postseason Success in Baseball



Just how did the Red Sox get past the Rays and Tigers in 2013?

Introduction

“They got hot at the right moment.” “They’re just lucky they peaked in October.” “It was just meant to be.”

These are all things that have been said about recent World Series winners. Ever since Major League Baseball switched to its current three-division system (and after adding a second wild card in 2012), it has made it more difficult for teams with the best records to win it all. This is because probably more than any other sport, baseball’s playoffs are so much different than its regular season.

Baseball’s 162-game regular season is a marathon of endurance and mental toughness. On the other hand, the playoffs are a sprint, with the winner often times being a team that by all traditional metrics (such as wins and winning percentage) is inferior. However, it is extremely difficult to predict when such a team will go on a World Series run. Even though there are several metrics to measure a player’s overall value to his team (such as WAR, or Wins Above Replacement), there is not a lot when it comes to statistics or groups of statistics that can best predict postseason success.

Michael Lewis’s Moneyball introduced the importance of on-base percentage (OBP) to many baseball fans, but I have determined through a simple regression analysis that statistic alone does not correlate to team postseason success. The general consensus among fans, commentators, and analysts is that having dominant pitching, particularly starting pitching, is the key to advancing far in the playoffs.

I agree that the most important variable on a playoff team is their starting pitching, but pitching alone doesn’t win you the World Series either. The 2013 postseason saw the Boston Red Sox in the ALCS beat the Detroit Tigers, a team that had what was considered to be the most dominant starting rotation in baseball. This was after they beat another team with excellent pitching, the Tampa Bay Rays, in the previous series. In a sport that has metrics to measure everything from speed on the base paths to the strength of an outfielder’s arm, there is no accepted metric that can accurately and consistently predict postseason success based on regular season performance. My goal was to see if I could find such a measure.

This is not a simple task. In an October 2013 article for ESPN’s Grantland, Rany Jazayerli wrote, “Trying to find the magic formula for postseason success has been the sabermetric community's version of trying to turn lead into gold: Many have tried, but none have entirely succeeded.” I first came up with the idea for this project after angrily watching the New York Yankees over the past decade consistently be one of the best teams in the league, but then lose in the postseason (often in in the division series).

Most fans and analysts pointed to the Yankees’ lack of quality starting pitchers post-2003 to why they couldn’t win in the playoffs after winning four of five World Series from 1996 to 2001. However, the Atlanta Braves, led by their dominating pitching trio of Greg Maddux, Tom Glavine and John Smoltz, had even more trouble in the postseason, winning only one World Series title from 1992 to 2005, despite winning the NL East title in all fourteen years. It amazed me how these teams could consistently dominate their respective divisions and leagues for 162 games, only to come out flat in a five or seven game series. It made me wonder if there were hints in a playoff team’s regular season statistics that could predict a successful postseason run.

For this research, I have defined postseason success as “playoff value” or PV. A PV of 1 means losing in the division series, 2 means losing in the Championship Series, 3 is losing in the World Series, and 4 is winning the World Series. Therefore, in order to find statistics that can predict postseason success, I ran hundreds of linear regression models, with the outcome variable PV, and with many different predictors.

Ability to Drive in Runs Without Hitting Home Runs
 

Hypothesis

For my research, I decided to focus mainly on regular season batting statistics of playoff teams from the past ten years (2003-2012). I did this for a few reasons. First off, as previously mentioned, it is widely accepted that good pitching beats good hitting in the playoffs. However, I think this only holds true when looking at conventional measures of “good” hitting, such as batting average and runs scored. Instead, it could be more important to look at team batting patterns and tendencies. It is my hypothesis that teams that have more simplistic batting approaches, or those that emphasize contact and putting the ball in play and deemphasize over-swinging to try to hit home runs, will be more successful in the postseason. The reasoning behind this is that the pitchers in the postseason are so dominating (the number of off days in the postseason means that teams usually only use three or four of their best starters), a team might only get one or two chances a game to get a rally going or drive in runs. And because the top pitchers in the playoffs, are usually less likely to give up home runs, it is important that when given the proper opportunity, teams are able to drive in runs without hitting home runs.

Results

I started by using the stepwise regression function in R in which, I predicted PV from the original 38 statistics I gathered. These statistics ranged from simplistic, such as hits and home runs, to advanced, such weighted on base average (wOBA) and weighted runs create plus per 600 plate appearances, to contact-based, such as groundball percentage and home run to fly ball ratio. The stepwise function took all possible predictors and entered and removed them from the regression model until all predictors in the model had a p value of less than .1.

The stepwise function gave me the following: PV ~ H + HR + BABIP + GBFB + LDp + HRFB + BUH + Swingp + Contactp. What this meant was that playoff value could be predicted by the combination of hits, home runs, batting average on balls in play, ground ball to fly ball ratio, line drive percentage, home run to fly ball ratio, bunt hits, swing percentage and contact percentage. After finding the summary of this model, I discovered it was statistically significant, as it had a p value of .038.

I was not surprised by a few aspects of the formula, as teams with higher LDp (line drive percentage) and GBFB (ground ball to fly ball ratio) stats usually mean they have more simplistic hitting approaches, as higher rates of hitting line drives and ground balls means that they aren’t over-swinging or trying to only hit home runs as much. However, it is very difficult to interpret these individual coefficients, due to the multicollinearity of the model.

This multicollinearity is caused by the high correlation between the variables in this model. For example, teams that usually have more hits are going to have more home runs, and a higher Batting Average on Balls in Play. After trying several other models that included variables that I thought would be significant (such as contact percentage, line drive percentage and zone contact percentage) I was still unable to find another model that was statistically significant, so I came up with another idea.

Be sure to check back tomorrow for Part 2 of Andrew's analysis.

Sportvision and the Future of Sabermetrics

During the 1970s and 1980s, Bill James revolutionized baseball through his collection of Baseball Abstracts. His unique perspective of evaluating players and discovering their true impact on their team's chances of winning was the beginning of a movement that would shake the very foundation of the sport.

Since the last Baseball Abstract was published in 1988, baseball sabermetrics have only continued to become increasingly popular and crucial to the ways franchises construct their teams. They even developed a presence in pop culture through the release of the Michael Lewis’ best-selling novel, and the later Hollywood film, Moneyball.

As sabermetrics have proven, through the success of teams like the Oakland A’s, to be effective in terms of evaluating the value of players, the precise statistics used have continued to evolve. Over the last ten years, sabermetrics have moved from the days of Bill James’ Runs Created, Win Shares and Range Factor statistics to even more complex formulas with more specific aims.

For example, the original Runs Created statistic that was developed by James has now been superseded by Weighted Runs Created plus (wRC+), a new equation which compares a player's On Base Plus Slugging (OPS) against the league average and then accounts for ballpark factors and run-scoring environments.

While sabermetrics have continued to become increasingly refined and specific, all of these detailed new statistics still only evaluate results. These advanced statistics such as Wins Above Replacement (WAR), Expected Fielding Independent Pitching (xFIP) and Skill Interactive ERA (SIERA), as effective as they are, are results-based. They fail to answer the question of why these results occurred.

Introduce Sportvision, a company whose technologically advanced cameras have been placed in Major League Baseball stadiums since 2006. Most fans probably already know Sportvision from the K-Zone cameras featured prominently by TBS and Fox this postseason. While Sportvision cameras might be enhancing the fan experience, their real value lies in the data they collect for teams to analyze.

Sportvision has developed services called Pitch F/X and Hit F/X, which track and record data from every single Major League Baseball game. For example, Pitch F/X tracks the velocity, horizontal movement, vertical movement and location of each pitch thrown. This data allows teams to analyze which pitch was most effective for a given pitcher and why that pitch was effective. A team could also analyze the value of velocity compared to location or movement.

Hit F/X takes a similar approach in analyzing batters. Instead of focusing on the results of each at bat, Hit F/X tracks the contact point, speed of the ball off the bat, elevation angle and field direction of each batted ball.

Sportvision, recognizing the value of this data, has created SCOUTrax, which uses the data from Pitch F/X and Hit F/X, as well as a third creation of theirs, Command F/X, to create heat maps and charts to better display the data to fans.

This new technology has opened the door for an endless number of new ways to evaluate the effectiveness of players, as well as help teams develop their own players. 

For example, teams will be able to see the value in an added half inch of movement to a pitcher’s fastball compared to extra velocity. Or, the team will be able to see that, although a particular hitter might not have had the best statistical year, he actually hit the ball particularly hard a high percentage of the time and should have fared better.

As the Sportvision data continues to be analyzed further, look for the development of future sabermetrics that are completely process-driven. Websites such as www.fangraphs.com have already started developing these types of statistics using Pitch F/X, such as Pitch Type Linear Weights, which attempts to determine a pitcher’s run expectancy per a given type of pitch.

Yet, the data from Sportvision is still in its youthful stages and needs to be further refined. For example, cameras in certain stadiums may read pitcher’s velocity or movement slightly differently, thus altering the way the data matches up. This issue might only create minor variances, but it is essential that it be fixed in the near future so a totally uniform data set can be collected.

Despite these minor deficiencies, Sportvision still holds the potential to greatly affect the way coaches, player personnel directors and baseball operations professionals develop players as their careers progress. As sabermetrics continue to evolve and become more refined, expect the data collected from from Pitch F/X and Hit F/X to be a central focus and have a profound impact on the game.

A Statistical Dissection of the 2012 Oakland A’s: Billy Beane’s Finest Work

Billy Beane 

The Hollywood blockbuster, Moneyball, ended much like any other feel good film, as protagonist Scott Hatteberg belted a home run that sent the small market underdog Oakland A’s into the postseason. However, unlike most silver screen dramas, the story of the Oakland A’s did not end with a “happy ever after” attached at the end.

After using Billy Beane’s innovative scouting to build teams that posted above .500 records from 1999 to 2006, the A’s could not finish above that mark for the next 5 years. It seemed as though the rest of the league had caught on to the A’s strategy, and that Billy Beane had lost his touch. Using their economical advantage, combined with the sabermetric scouting originally developed by the A’s, other larger market teams seemed to once again hold a competitive advantage over Oakland.

At the onset of the 2012 baseball season, virtually all baseball fans could not imagine the Oakland A’s making much of an impact in an AL West division dominated by the superstar heavy Texas Rangers and Los Angeles of Anaheim. However, 2012 turned out to be Billy Beane’s most masterful work yet as the Oakland A’s finished 94-70, winning the division with a payroll of roughly $99 million less than Angels and $65 million less than the Rangers. So how did this happen? How did the A’s once again use undervalued talent to reach the postseason?

When opposing pitchers stared at the Oakland A’s lineup card during 2012 season, it is hard to imagine any of them shaking in fear. The A’s did not feature any hitter batting over .300 or who drove in over 85 runs, and only had three players finish with more than 20 home runs. However, the A’s use of platooning, a declining trend in Major League Baseball, to account for their deficiencies, led to the underrated effectiveness of their lineup.

At first base for example, Brandon Moss is the regular starter against right-handed pitching, while Chris Carter usually gets the start against left-handers. Moss put up stellar numbers against right handed pitching all season, hitting 19 home runs, while driving in 44 runs and slugging .643 in 207 at bats. Against lefties, Carter hit five home runs, drove in 17, while slugging .494 in 83 at bats. Another platoon employed by the A’s is at designated hitter with Johnny Gomes and Seth Smith. Smith homered 12 times while slugging .454 and having an OPS of .805 against right handers in 313 at bats. While Gomes hit 11 homers to go along with a .413 OBP and .974 OPS against left handers in 164 at bats. The A’s have also prominently used other platoons throughout the season, such as at the catcher spot with Derek Norris and George Kottaras.

While the A’s use of platooning did spark their offense, this success did not stand out in traditional statistical categories. With a batting average of .238 the A’s finished 28th in the league, leaving only two last-place teams, the Houston Astros and the Seattle Mariners, with lower marks. Even On Base Percentage, a statistic prominently featured in the original Moneyball, cannot account for their 2012 success, as their .310 OBP ranks 24th in the league.

However, what Oakland lacked in getting on base and batting average they made up for in the power game. The A’s finished 7th in the league in home runs and while their slugging percentage ranked 15th in the league, that ranking is considerably higher than expected given their lowly batting average. In the sabermetric community, Oakland’s hitters excelled, as well. The A’s ranked 10th in the league in Weighted Runs Created (wRC+), an improved version of Bill James’s original Runs Created statistic, which “attempts to quantify a player’s total value and measure it in runs.” Overall, despite a ridiculously low batting average and lack of star power, the A’s were able to produce the 14th most runs in Major League Baseball, which, with their surprisingly impressive pitching staff, proved to win a lot of ballgames.

Similar to the offense, the A’s pitching staff proved that big names are not a necessity for success. Unlike the hitters, however, more traditional metrics can be used to quantify the A’s pitching staff feats. In 2012 the A’s staff posted a 3.48 ERA, which was good for 6th best in the league to go along with 1.24 WHIP. The one anomaly was strikeouts, in which the A’s finished 26th in the league. This lack of strikeouts, however, was not truly an anomaly, but rather an indicator of a general pitching strategy employed by the A's.

The A’s finished 9th in the league in fewest walks allowed going along with having the 10th lowest walk percentage. The A’s also had the 7th highest percentage of fastballs thrown. These statistics show that A’s pitchers used the strategy of attacking the zone with fastballs and forcing contact in order to limit opposing batters. This strategy does not require having pitchers with dominant strikeout arsenals, a skill set generally more expensive to obtain.  While pitching to contact by attacking hitters with fastballs worked in 2012 for the A’s, advanced sabermetrics suggest that the success of the A’s pitchers might be somewhat attributable to luck. The A’s had the third lowest BABIP (Batting average on balls in play) in the league, the 7th highest xFIP (Expected fielding independent pitching), and the 8th highest SIERA (Skill interactive ERA). Having a low BABIP is a signal that many balls hit by opposing batters went directly to Oakland fielders. Going forward, this might not always be the case and more balls could fall in for hits. xFIP is a statistic that attempts to judge pitchers on entirely what they can control, using a combination of strikeouts, walks, hit by pitches, and “how many home runs they should have allowed” (using home run to fly ball ratios and multiplying it by fly ball rate), to try and determine how effective a pitcher is without the effect of his fielders. The higher the xFIP a team or pitcher has, the worse they were and the A’s xFIP of 4.2 is categorized as “below average” by Fangraphs.

The last sabermetric in which the A’s pitching staff failed to excel was SIERA. SIERA differs from xFIP because it places more emphasis on balls in play, using groundball and flyball rates, as well as walks and strikeouts to attempt to determine a pitchers skill. The A’s finished the year with a SIERA of 4.03, which was the 8th highest in all of baseball. So while the A’s did have one of the most effective staffs in the league according to traditional measures like ERA, advanced sabermetrics predict that the A’s success might soon run out or is unlikely to reoccur. In total, however, the A’s only allowed the 6th lowest amount of runs in the league in 2012, which along with their power and timely hitting led to winning 94 games.

The next question one must ask when dissecting the success of the 2012 Oakland A’s is, how did Billy Beane assemble this team? What changes were made from 2011 to 2012 that led to this significant increase in wins?

The only headlining acquisition was the signing of international free agent Yoenis Cespedes for a 4 year, 36 million dollar deal. Cespedes might have been an international superstar, but many in baseball circles thought that it would take some time for him to adjust to Major League pitching. Yet, the A’s Director of Baseball Operations Farhan Zaidi, felt he could be a key contributor and convinced Beane to make the investment. The 2012 season is proof that Zaini made the right judgment, as Cespedes hit .292 with 23 home runs and 82 RBI’s.

While the Cespedes deal made headlines, it was the more low key moves made by Beane that really boosted the A’s all season long. Similar to the 2002 season depicted in Moneyball, Beane was able to find value in players where other teams did not. In a deal with the Red Sox, Beane sent reliever Andrew Bailey and Ryan Sweeney to the Red Sox for inexperienced outfielder Josh Reddick and two minor leaguers. As stated earlier, Reddick exceeded everyone’s expectations by hitting 32 homers and slugging .463. Also, in a multiplayer deal with the Diamondbacks, Beane traded starter Trevor Cahill and veteran reliever Craig Breslow for rookie Jarrod Parker, reliever Ryan Cook, and Colin Cowgill. In 2012, Parker was arguably the best rookie pitcher in the American League posting 3.47 ERA in 181.1 innings, while Ryan Cook was an All-Star reliever in the set-up role with a 2.09 ERA and 42 holds.

The third trade which significantly increased the strength of the team was the deal with the Nationals, which sent Gio Gonzalez and a minor leaguer for rookie Tommy Milone, Derek Norris, Brad Peacock, and a minor leaguer. While Gonzalez won 20 games, he would have soon been a contract the A’s could not afford. The newly acquired Milone, a soft-tossing righty with an average fastball velocity of 87.7 went on to also be one of the best rookie pitchers in the American League, posting a 3.74 ERA while winning 13 games. Norris also stepped in, playing a crucial role at the catcher position as half of the platoon with George Kottaras.  All of these trades represent examples of Beane dumping established players for relatively unproven talent, yet in all three cases the unproven talent was able to significantly contribute the A's success in 2012; this is a great tribute to the effectiveness of the A's scouting department.. These three trades, along with a trade with the Rockies for Seth Smith and the signings of journeymen Jonny Gomes and Brandon Moss, are all examples of Beane’s ability to trade and acquire underutilized cheap talent.

Though the 2002 season will probably be most remembered when people look at the Billy Beane’s career as General Manager, it is the 2012 season that is by far his finest so far. Unlike 2002, in 2012, Beane did not have the luxury of marching out three all-star quality starters (Barry Zito, Mark Mulder, and Tim Hudson). Instead, in 2012, Beane not only had to piece together a lineup full of mostly unknown veterans and rookies, but he also had to construct a pitching staff of underrated parts as well.

Beane’s ability to assemble talent as well as manager Bob Melvin’s coaching staff’s ability to maximize it, through the implementation of platoons and pitching strategy, led to the 2012 Oakland A’s far exceeding anyone’s expectations and the revival of the belief in Moneyball in its original city.

Reflections on an Unexplainable Game

As pure and natural as kids on a sandlot

As the calendar moves to the latter stages of February, and the early days of March, the people's attention once again will turn to back to the Diamond. Yes, it’s that time of the year, where hope springs eternal, anything seems possible, and everything feels right with the world. Baseball is back.

For a game that very recently has been predicated on the technical through analysis of statistics, finding meaning in numbers, and identifying trends through various metrics, the reality is a simple one: sometimes there’s just no scientific explanation for the game we love.

Now don’t get me wrong, I, myself, am one of these individuals, searching for answers, predictions, and ways to advance the game. I love thorough knowledge. When you take a step back, though, like many other things in life, there may just be no scientific answer.

How do you quantify the crisp crack of the bat, as players practice in the warm sun, under the clear, blue sky? How do you analyze the camaraderie, the chatter brandied about by grown men? Can you really analyze what that new glove feels like when the ball meets it or that sweet “pop”?

What I'm trying to say is, there’s something magical about this game, something that deep down draws us all in. It’s that feeling of youthfulness, of seeing the green, fresh-cut grass, of listening to the sounds of the players, of seeing the white ball go from pitcher to bat to glove. It’s this thing that can’t be quantified, the human element of Baseball, and it's the thing that keeps us coming back, hungry for more, as a rite of passage every Spring.

While our society continues to speed up, and our game continues to become more methodologically-based, the unidentifiable variable will continue to exist. Players will be analyzed by the brightest minds who utilize the likes of all-encompassing statistics such as WAR, UZR, and PECOTA, but in the end, it's the game, a game for young men, that will remain. The human element, memories that our national pastime invoke, emotions that it conjures up, and feelings that it draws out of us, while un-quantifiable, is what will draw us all back again and again.

It's like an old friend calling up after a long, cold Winter, just to say, "Remember me?" You answer the phone, and pick up right where you both left off, nothing has changed. Welcome back, Baseball. We can't even begin to tell you how much we've missed you.