The Major League Baseball regular season begins in fewer than three weeks, when the Los Angeles Dodgers play the Arizona Diamondbacks in Sydney, Australia. Opening night is Sunday, March 30th (featuring the Dodgers again, it is as if they come from a big market or something) and everyone will jump into the mix on opening day on the 31st.
I myself am currently in Arizona, spending some time with family while watching Cactus League spring training games. On my flights from Chicago to Phoenix (easily the most cheerful flights of the 50-plus times I have flown) I perused some chapters of Scorecasting, which I have read a few times. It both reminded and inspired me to reflect a little bit on baseball, the original moneyball.
With box scores going back over 100 years, and Bill James’ first baseball abstract approaching its 40th anniversary, baseball tactics have adapted to (or been reinforced by) the power of (useful) statistics more than any other sport. A walk may not be as exciting as a hit, but by and large all teams now value it accurately as a viable way of generating runs. Michael Lewis wrote about the power of on-base percentage (and other less conventional measures) in Moneyball and the secret has been out for some time. Says Andrew Friedman, general manager of the Tampa Bay Rays:
The game is incredibly efficient right now relative to where it was ten years ago. Our greatest fear is it becomes perfectly efficient.
What are the remaining inefficiencies in the game? What could teams change to improve their chances of winning, however slightly?
A team’s ninth-best batter should bat sooner than ninth (especially when ninth-best by a wide margin).
A team’s leadoff hitter receives the most plate appearances, and a team’s ninth hitter receives the least. Certainly the best hitters should come first. But the top of the order is guaranteed to begin an inning only once a game; if the rest of the time the best hitters are preceded by an easy out, they are less likely to produce runs. This effect is more pronounced when the difference between a team’s eighth-best and ninth-best hitters is significant (say, 0.050-0.100 in batting average), hence the usefulness of batting the pitcher eighth in the national league.
How useful? When the gap is significant (as it is in the national league), putting a slightly better (or “less-bad”) hitter in the ninth spot is worth about two runs a season. It sure is not much, but it is something. Teams do occasionally miss the playoffs by one or two games; one or two runs at the right time could be the difference.
Pitchers should rotate more often, for less time.
Tobias Moskowitz and Jon Wertheim discuss this in Scorecasting; in 1993, Tony La Russa actually tried this for a few games before relenting to his starting pitchers’ complaints. Tried what, exactly? La Russa tried cycling his pitchers every 30-50 pitches or so, maybe two to four innings. There could be many benefits to this: pitchers work less intensively and so more are available at any given time to face a particular batter or situation; hitters do not get to see a pitcher more than once, maybe twice1; (formerly starting) pitchers will never be as tired in a game as they routinely are in the current system; etc, etc, etc.
The players themselves, however, and especially starting pitchers who become ineligible to officially record wins, seem to hate the idea, as La Russa discovered in his short-lived experiment. It has never been tried again. Perhaps in this age of league-wide moneyball, with pitching salaries tied more to strikeouts and walks than wins and losses, such a rotation would be more feasible. Given that starting pitchers still chase the ‘W’s and complete games, though, it seems pretty unlikely.
Closing pitchers are often more useful in middle relief.
Even in the current system, a closer–typically a team’s top relief pitcher–generates a similar problem for efficiency. When a manager saves his closer for the ninth, he may be costing his team runs in the preceding innings. Consider: with a one-run lead, would you want your best reliever to come in to pitch the 9th, starting with zero outs and nobody on, or in the 7th, with two on and nobody out? The chances of your opponent scoring (to tie or take the lead) are much higher in the latter situation, yet teams usually employ a “middle reliever” in such a situation. Saving the best for last is no good if you give up runs in the process.
Closers live to acquire saves just as starters do for wins. But with the evolution of the “hold” as a statistic and the continual search for low-hanging strategic fruit, perhaps the coming years will see a shift towards a more optimal relief pitching strategy.
All three of these strategies are a departure from the convention, even the modern, data-driven convention of today. And, even though they may work, it is unlikely to see them employed widely any season soon. Turns out the childhood adage still holds: winning really isn’t everything, even in professional sports.
- In Scorecasting, the authors mention that batters hit and get on base .030 less often in their first at-bat against a pitcher relative to all other at-bats against the same pitcher in the same game. ↩