Posted by: Mike Cornelius | October 14, 2018

Testing The Limits Of Moneyball

Much has been written over the last few seasons, including in this space, about how the Great Game is changing. The concept of relying on advanced metrics rather than the subjective judgments of scouts and coaches was popularized by “Moneyball,” the 2003 Michael Lewis book that highlighted how the low-budget Oakland A’s made it to the playoffs in the previous season. The book became a critically acclaimed movie in 2011, giving A’s general manager Billy Beane the luxury of forever being able to remind anyone listening that he was played by Brad Pitt. But in recent seasons faith in sabermetrics has accelerated among all thirty major league teams, with detailed statistics on every aspect of the game far beyond the days of “Moneyball.”

Every front office now has an entire department devoted to statistical analysis, and scoreboards across the land let fans know the exit velocity of every ball put into play. The data now available on every hitter’s proclivities has led to a dramatic increase in defensive shifts, which in turn has spawned a new generation of players with uppercut swings trying to lift the ball over the picket fence of defenders on one side of the infield and into the seats. Home runs have spiked, while traditionalists have lamented the loss of the stolen base, the hit-and-run, the sacrifice bunt, all key elements of baseball offense in another time.

But as has been highlighted this postseason, it is not just defensive positioning and the response of players at the plate that has changed. So too, fundamentally so for some teams, has the role of the starting pitcher. Increased reliance on a team’s bullpen, and with it the emergence of specialty relievers who are called upon in specific situations, has been a growing trend for decades. Starters, who once were handed the ball and expected as often as not to finish the game, have seen their innings steadily shrink. In 1975 Catfish Hunter threw thirty complete games for the Yankees, just three years after Steve Carlton led the National League with the same number for the Phillies. That was the decade that the major leagues last saw a thousand complete games in a season. By the end of the next decade the number had dipped below five hundred. This year eleven starters, six in the American League and five in the National, tied for the lead in complete games by going the route just two times. Those twenty-two complete games were just over half of the forty-two thrown in the majors in 2018, a record low.

A major reason for the willingness of managers to go with a quick hook on their starting pitcher has been clear statistical evidence that in the relationship between the man on the mound and his opponents in the batter’s box, familiarity breeds contempt. The success rate for batters, be they elite hitters with gaudy averages or journeyman struggling to hit .220, spikes once they step in to face a pitcher for the third time in a game. In an ideal world from the pitcher’s perspective, that might not happen until the 7th inning. But under the right circumstances it could come as early as the first batter of the 4th inning, even in a scoreless game.

That statistical evidence has made the third time through the order, whenever it takes place, the time that most managers now think about calling the bullpen, even when their starter is doing well. This season it also led to a few teams, most frequently the Tampa Bay Rays, to go with a so-called “opener,” a relief pitcher who starts the game with the purpose of pitching just one or at most two innings. The idea is to force batters to face several different pitchers during the game, eliminating the increased success rate from seeing the same hurler multiple times. Bullpen games, when a team might use six or seven pitchers, used to be a sign that the starter had imploded early. Now for some teams they are both intentional and not infrequent.

The Milwaukee Brewers captured the NL Central crown by closing the season with seven straight wins to tie Chicago and added an eighth with their victory over the Cubs in the tiebreaker game. The Brewers then won their ninth, tenth and eleventh in a row as they swept the Rockies in the NLDS. During that streak Milwaukee’s starting pitchers averaged less than four innings pitched per game, and manager Craig Counsell used on average just a fraction under six pitcher per game. The second game in the streak, a 6-4 win over the Cardinals, was a bullpen game. In his only start of the entire season Dan Jennings retired the only batter he faced before giving way to a parade of fellow relievers that eventually numbered nine. Members of the Brewers starting rotation notched just two of the eleven wins, in part because they only reached five innings pitched, the minimum required for a starter to be credited with a win, in four games.

Milwaukee’s approach is a product of statistics that suggest the team’s bullpen is very strong, while its starting rotation is not. Not surprisingly, Counsell stuck with the tactic through the first two games of the NLCS against the Los Angeles Dodgers. In Game One Friday night, starter Gio Gonzalez lasted just two innings before giving way to the first of six relievers who were called upon the record the final twenty-one outs. On Saturday Wade Miley went five and two-thirds, a remarkable outing by Milwaukee’s recent standards. Five relievers followed Miley to the mound.

But the Brewers are heading west, where the next three games will be at Chavez Ravine, all square against their NLCS foe. That fact, and what took place in the two contests at Miller Park, should be a reminder to fans, pundits, and members of every team’s analytics department, that the Great Game is not played on a spreadsheet, and that there is still value in paying attention to what’s taking place on the field.

Gonzalez left Game One trailing 1-0 after surrendering a home run to Manny Machado. But he allowed just one other base runner and obviously had not run up a big pitch count. The Brewers came back against Clayton Kershaw, but what looked like a safe 5-1 lead nearly evaporated in the late innings as L.A. pecked away against the vaunted Milwaukee bullpen. The final was 5-4, with Milwaukee relievers giving up four earned runs on seven hits for a less than lights out ERA of 5.14.

When Miley was lifted Saturday, he had thrown just seventy-four pitches and was blanking the Dodgers on two hits, both singles. But the second of those was off the bat of Chris Taylor, the final batter he faced. Taylor was the Dodgers leadoff hitter, meaning Miley was starting his third time through the L.A. order. So Counsell marched in lockstep with the statistical evidence and went to his bullpen. Corbin Burnes got Justin Turner to fly out to end the inning but failed to retire a man in the 7th. The Dodgers plated two in that inning and two more in the 8th on Turner’s drive into the left field seats. The final score was 4-3 L.A., with Milwaukee relievers giving up all the runs while posting an ugly 10.80 ERA.

Any single game is of course an exercise in small sample size. Perhaps over the course of the series Milwaukee’s approach will win out. Perhaps Counsell has so little faith in his starters that he genuinely believes he must continue down his chosen path. But before the Great Game turns managing over to robots, perhaps he might consider going with his gut.

Advertisements

Responses

  1. This is a very insightful look at what’s going on in the majors. I guess we are entering an age of specialists on the mound.
    Ω


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Categories

%d bloggers like this: