Do Managers Make a Difference in One-Run Games?

The argument for one-run games being a possible indicator of a manager’s skill and effectiveness in game-on-the-line circumstances is that these are the games where his decisions would have the most obvious impact, as suggested by several exciting games of the 2014 post-season mentioned in my previous article.  The prevailing view among sabermetric analysts, however, is that it is misleading to evaluate any manager’s performance based on such indicators as his record in one-run games, which are decided as much by luck as a manager’s game-management skills. Inspired by Branch Rickey’s famous dictum, “Luck is the residue of design,” this Insight seeks to explore that issue.

As noted in my previous article, Nats’ manager Matt Williams’ decision to remove Jordan Zimmermann in a 1-0 game with the tying run on base but needing only one more out to even the division series blew up in his face; O’s manager Buck Showalter’s decision to put the potential winning run on base with the tying run on second, with only one out and his team just two outs away from advancing to the ALCS, proved a masterstroke; and Royals’ manager Ned Yost’s decision to use a starting pitcher instead of his go-to seventh-inning guy to protect a one-run lead in the sixth in the wild card game nearly cost KC the game and their ultimate road to the World Series.

Games decided by one-run are not simply low-scoring affairs, such as often associated with the Dead Ball Era, but also games decided by scores like 9-8, 8-7, and 7-6. Most are legitimately on the line in the late innings, even if some one-run games begin as blowouts and become close only because of frenetic comebacks that fall just short.

The percentage of major league games decided by one run was typically between 30 and 35 percent during the Dead Ball Era; mostly between 25 and 30 percent from the 1920s through the 1950s when the power game was prevalent; and consistently back over 30 percent in the 1960s—including 35 percent in 1968, which was so much the “Year of the Pitcher” that the permissible height of the mound was dramatically lowered beginning the next year.

There was a marked divergence between the American and National Leagues in the first two decades of the DH rule beginning in 1973, where one-run games accounted for about one-third of the outcomes in the NL but for only between 25 and 30 percent in the AL. The major league average was about 27 percent in the first decade of the 2000s, with the NL only slightly higher than the AL. In the last five years, about 30 percent of major league games have been decided by one-run, including 29 percent in the American League and 32 percent in the National League last year. (All of the above data is available on the baseball-reference website for every season played since 1900.)

As countless teams have experienced and millions of fans have endured, however, the outcome of games decided by only one run could very much depend on a lucky break or the ability (or inability) of players to execute—events that say nothing about the actual merits of a manager’s decisions. It is a matter of inches, after all (or so they say), between fair and foul balls, safe and out on the bases or at the plate, home runs or long fly ball outs to the wall that decide the outcome of games.

Precisely because so much is up to chance is a strong argument for why a manager’s record in one-run games as an indicator of his game-management skills and performance is misleading.

After all, the Orioles’ record in one-run games over the last three years under Buck Showalter, one of the most respected managers in baseball today, is marked by extremes. The 2014 O’s had a 32-23 record in games decided by one run — a winning percentage only 11 points below Baltimore’s 96-66 record for the season. But the previous year, Showalter’s record in one-run games was an extraordinarily ugly 20-31, below .400, for a team that in 2013 had a .525 winning percentage, while in 2012 the Orioles were at the other end of  extraordinary with an amazing 29-9 mark in one-run affairs. It would surely be ludicrous to suggest Showalter is only a good manager in tight games in the even-numbered years.

Grounded in a long history of play-by-play data, sabermetric analysis typically discounts teams’ records in such close games as an indicator of their manager’s decisionmaking prowess. For example, the probabilities of a team scoring a run indicates that managerial decisions to play for one run using strategies such as sacrifice bunts to advance runners at the expense of an out, or stolen bases—which, if unsuccessful, cost an out and eliminate a runner—often actually reduces the odds of scoring.

Furthermore, most game strategies are obvious and pro forma, such as whether to sacrifice bunt in the late innings in a tie game, to walk a dangerous hitter with the tying run in scoring position and first base open, or even to bring in a left-handed specialist to pitch to a dangerous left-handed batter at a critical spot in the game. Whether to go “by the book” or not is the manager’s choice and the outcome is often indeed dependent on his players’ ability to execute or a matter of luck.

Still, bearing in mind Branch Rickey’s Bartlett’s quotation about “luck being the residue of design”—which is itself a variation of Roman philosopher Seneca’s dictum that “luck is what happens when preparation meets opportunity” (well-read in the classics was Mr. Rickey)—managers making decisions to play for one run do so because they believe just one run is what is needed at that moment to ultimately win the game, particularly if it is the late innings, and therefore is worth the risk and the sacrifice of a potentially bigger inning. Even if their strategy to play for one run actually reduces the odds of scoring according to the probabilities of run-expectancy from their decision, their bias is towards not blowing the opportunity to score just one run. And it is the same concerning decisions about pitching changes, defensive alignments, and how and even whether to pitch to particular batters.

Their players may fail to execute. Fate may intervene in a bad bounce, a great defensive play, or a lousy call (although replay challenges are a remedy to that problem). But managers make the decisions they do at the moments they do for the very purpose, they believe, of maximizing their team’s chance to win the game. We should be careful never to assume that the manager did not have very good reasons for making the decisions he did—even those that backfired and have fans and pundits screaming, What was he thinking!

A manager’s decisions are made in the context of not only the game situation, but also his years (and years) of learned experience in the game, as well as the philosophical approach and willingness to take risks that he brings to the game as part of his personal and, sometimes, his team’s organizational history. In this regard, it would seem their decisions are less about luck (although still dependent on execution) than about the manager’s judgment. If so, a manager’s record in one-run games may indeed be telling as a valid indicator in the outcome—win or lose—of close and sometimes critical games.

There will be further consideration of this issue in the future.