Do You Miss Baseball? Here’s How the Pentagon Might Have Predicted the Nationals’ World Series Upset
Red teaming can help spot flawed assumptions, root out dangerous biases—and explain how oddsmakers got the Nats-Astros matchup so wrong.
Last fall, the Houston Astros were prohibitive favorites to win the World Series. They were, in fact, the biggest frontrunners since 2007, when the Boston Red Sox went on to sweep the overmatched Colorado Rockies in four games.
But this time, the anointed superteam did not win the world championship. Instead, the Washington Nationals defeated the Astros in a memorable seven-game series, winning all four games in Houston, an unlikely feat that Major League Baseball fans may never witness again.
Hindsight is helpful in seeing the depth and strength of the 2019 Nationals team. But as Major League Baseball fans await the delay or possible cancellation of the 2020 season, it might be a good time to explore the question of why oddsmakers and pundits got it so wrong.
At the Pentagon, where I work, national security analysts spend a lot of time debating the relative strengths and weaknesses of America’s potential adversaries. Through a process known as “red teaming,” planners examine myriad topics related to military strength and use data to explore all kinds of plausible outcomes in the event the United States has to go to war. This red teaming is critical to spotting flawed assumptions, groupthink, and all kinds of biases that could get people killed in a real conflict.
Even a cursory round of red teaming before last year’s World Series would have revealed the two ballclubs to be similar in all but one key respect: their relief pitchers. The Astros had the second-best bullpen in baseball, the Nationals the second-worst. In the world of analytic “what-iffing” that is deeply ingrained in military exercises, that bullpen deficiency might not have mattered nearly as much as odds-makers and pundits imagined as they went about setting the series odds.
But before we run that exercise, let’s look at the two teams in depth and explore other ways in which observers might have been biased to assume large advantages for the Astros.
Houston won 107 games and boasted a roster filled with stars. MLB Network’s list of the top 100 players in baseball is filled with 2019 Astros: Alex Bregman, Jose Altuve, Justin Verlander, Gerrit Cole, Carlos Correa, Michael Brantley, Zack Greinke, Yordan Alvarez, and George Springer. Houston's was perhaps the most talented roster in baseball. But that same top-100 list is filled with almost as many 2019 Nationals: Anthony Rendon, Juan Soto, Max Scherzer, Stephen Strasburg, Patrick Corbin, and Trea Turner. In the terms familiar to national security scholars, the Astros and Nationals were by many measures “near-peer competitors.”
Red teamers would have likely deemed the depth of talent in position players and starting pitchers on both teams as roughly equal, or perhaps given a slight edge to Houston. And in examining the disparity in bullpens – during the regular season, Astros relief pitchers allowed the second-fewest runs in baseball while the Nationals ’pen surrendered a whopping two runs per game more than the Astros – they would have asked several simple questions. The first is, could the Nationals mitigate that vulnerability and, if so, what difference would it make? As it turns out, the answer to that question was hiding in plain sight. During the first three rounds of the playoffs, the Nationals deputized members of their elite starting pitching staff four times, all but eliminating their middle-inning bullpen vulnerabilities.
The second question that red teamers would have likely asked is related to the baseball equivalent of military “readiness.” In wargaming, the quality of a potential adversary is judged by many factors, including training and preparedness against a skilled enemy. And this is where things start to get interesting. A close examination of the teams’ regular-season schedules reveals a wide disparity. The Astros, who won 107 games, played 99 against teams with losing records and just 63 against teams who won more than they lost. The Nationals played 96 games against teams .500 or better and only 66 against teams with losing records. In short, one team’s win total was juiced by the baseball equivalent of facing third-world military forces; the other’s was suppressed by routinely competing against well-trained NATO allies.
The Astros won 73 percent of the games they played against losing teams, but just 55 percent against winning teams. The Nationals won 68 percent of their games against losing teams and went 48-48 against teams .500 or better. The Nationals won 93 games during the regular season, but based on the difference between their runs scored and runs allowed, known as their Pythagorean expectation, the Nats should have won 95 games, a more-than-respectable result in the National League East, a far stronger division than the Astros’ home in the American League West. The NL East had perhaps the best pitching in baseball last year, headlined by the likes of the Mets’ Jacob deGrom and Noah Syndergaard, the Phillies’ Aaron Nola, and the Braves’ Mike Soroka. The Nationals entered the World Series having repeatedly faced some of the best starting pitchers in baseball.
Red teamers would no doubt have looked closely at the skill level of the Astros’ opponents and wondered: What if the Nationals had played that same schedule? The answer is revealing. If the Nats had played the same number of games against losing teams as the Astros, based on their season-long performance, their win total would jump from 93 games to 98. And if the Astros had played 96 games against teams .500 or better, as had the Nats, their wins would have dropped from 107 to 100. The difference between the two would have been virtually imperceptible.
Analysts red-teaming the World Series would have noticed something else that could be classified under “readiness.” During the regular season, the Astros seldom — if ever — faced starting pitchers as good as the Nationals’ duo of Strasburg and Scherzer. The baseball analyst Eno Sarris ranks Scherzer as the major leagues’ fourth-best starting pitcher, and Strasburg as No. 6. Meanwhile, the best pitcher in the American League West who is not an Astro, according to Sarris’s rankings, is Frankie Montas of the Oakland Athletics, judged the 24th-best starting pitcher in MLB and the division’s lone top-30 starting pitcher outside Houston. Good pitching stopped good hitting: Washington’s aces effectively shut down the powerful Astro offense during the four games the Nationals won in Houston.
Several years ago, Vince Gennaro, the president of the Society for American Baseball Research, introduced the idea that one useful predictor of how batters might fare in postseason play is how they performed against top-flight pitchers during the regular season. During the 2019 regular season, the Astros faced starting pitchers ranked in the top 30 on Sarris’s list 19 times, scoring an average of just under four runs per 9 innings. The Nationals faced top 30 pitchers a total of 31 times and scored an average of 4.35 runs per game. Against top-flight pitchers during the season, the Nats offense was more productive, even though they faced elite pitchers far more often — including eight games in September as they battled for a playoff spot. The Astros faced no top-30 pitchers after Aug. 27.
There is one final question war gamers would have asked about the Nationals. How does a team with an historically bad bullpen, playing a challenging schedule in a tough division, win 93 games? The obvious answer is that no ordinary team does that, something the Astros and baseball found out last October.
The opinions in this article do not necessarily reflect those of the U.S. Air Force or Defense Department.