The Boston Red Sox are on pace to set a modern scoring record as they are scoring near six runs a game. But could they be better? Lineup optimization is a dynamic strategy that is nowhere close to being perfected. It’s potentially something that can never be perfected. If every player was a robot that performed the exact same regardless of any human variance, it might be possible. A “perfect” lineup could be put in place for each opposing pitcher, including making proper substitutions when a relief pitcher came into the game to give a chance for the best potential outcome.
The previously mentioned Red Sox have scored as many runs as they have due to David Ortiz improbably having the best year in his illustrious career. Ortiz is pacing the league in many offensive statistics, standard and advanced, and has hit cleanup in the majority of his performances in John Farrell’s lineup. Xander Bogaerts is also having an MVP-type season, and he has hit third in the lineup nearly every day. Veteran and team-leader Dustin Pedroia has had a very good year, not quite to Bogaerts’s levels, and is most often put in the second spot in the lineup.
The new trend in lineup construction has argued that the “best” hitter in a team’s lineup ideally should hit second. The second spot in the lineup is the most “important” place – calculated by run-scoring chances and frequency of plate appearances. With those arguments available to the Red Sox, (who have the sabermetric godfather Bill James in their front office) should Bogaerts and Pedroia flip spots in the lineup? Taking it to another level, should Ortiz be hitting second?
There may not be a right answer. It is possible when moving Pedroia out of his comfortable second spot, the entire machine may malfunction. Jackie Bradley Jr. throws in another wrinkle, as he is also having an unpredicted great year leading the Red Sox to have the highest run production (wRC+) out of their ninth spot in the entire league. On the other hand, the Red Sox have one of the lowest run production values from the fifth spot in the lineup. The fifth spot is arguably one of the top three spots in importance, and there are players that the Red Sox could be using to better optimize that spot. It’s clearly not hurting them to a great degree in actual run production, but the theorized run production predicts it would. But again the question is could the Red Sox be scoring even more? Are other teams showing similar examples of optimization and could certain team’s lineups cost themselves enough wins to lose a playoff spot? Below are examples of a team that is outproducing its metrics, one under-producing, and a third which is bucking all theoretical patterns.
The correlation between the league’s wRC+ value and runs per game is very strong – .853. Knowing the definition of wRC+ would lead one to believe this to be true, as more base runners and more total bases would theoretically lead to more runs. There are some outliers however, and that’s where the deep dive begins. The Colorado Rockies have the twentieth most productive offense in the league according to wRC+, yet have scored the fifth most runs. Two theories (and possibly more) could explain this: The Rockies are more “clutch” than any other team. The Rockies are indeed fifth in batting average with runners in scoring position, and they are fourth in slugging percentage with RISP. That doesn’t seem enough to jump them fifteen spots between the two categories however. The other theory is that the Rockies lineup is optimized. If a team has their production come through in the most important situations, that could bridge the gap between wRC+ and actual runs scored. If their best hitters (ex. Nolan Arenado) are always coming to base with runners on, and he is always succeeding in those situations, that could cover for the rest of his under-producing teammates. Arenado’s OPS with RISP is indeed sixth in the league with a fantastic 1.218 OPS while also garnering the fourteenth most at-bats in those situations. Teammate Carlos Gonzalez, no slouch himself with a .875 OPS with RISP, has the same amount of at-bats. These chances could be just that – a chance, a coincidence, created due to just small sample size. But the numbers are at least showing for now, the Rockies have the right guy up at the right time.
This isn’t of course the situation every time. In fact, the first two months of the season showed that Gerardo Parra, he of the 65 wRC+ in 2016, was actually coming to bat with runners on base more often. Of Parra’s first 222 plate appearances in the 2016 season, there was on average, 0.635 runners on base compared to Arenado’s 0.613 average. Parra began the season further up in the lineup, usually fifth, and for the first 33 games the Rockies averaged 5.09 runs per game, a very respectable total. As of game 34, Parra moved down in the lineup (to sixth or seventh) and in the next 31 games, the Rockies were up to 5.16 runs per game. In fact, when the Rockies later moved DJ Lemahieu to the second spot in the lineup, the 20 games after saw the Rockies putting up 5.7 runs per game, which would place them second in the league, only behind the Red Sox. There may be other moves that the Rockies can still continue to look at. Their production from the leadoff spot remains below league average, for example. However, the fact that the Rockies currently are getting their best production from the second spot, and next best from their fourth spot, could correlate to their lineup being the most optimal in the league.
The Tampa Bay Rays are on the other side of the spectrum. Despite the seventh highest wRC+ in the league, they only rank 17th in runs per game. If one were to compare the Rays to the Rockies, they could come away with either the Rays aren’t “clutch”, or they could say that the Rays lineup isn’t optimal. The Rays do have the seventh lowest batting average with RISP in the league; that certainly tells a portion of the narrative. Evan Longoria has been the Rays best hitter this season and he finds himself in the third spot in the lineup every game. Out of the everyday batters in the Rays lineup, Longoria has only the fifth highest ratio of having runners on base during his at-bats. To worsen matters, the disappointing Corey Dickerson (81 wRC+) has come up to bat most often with runners on base. The Rays have taken notice of this. After game 35, the Rays were averaging 3.66 runs per game. It seems like this was the point where Kevin Cash and others believed it was time to make a change. They moved Dickerson down in the order, and in the 27 games after, they were averaging 4.82 runs per game – over a run per game increase. Cash has been selective with his batting orders, for example hitting Brad Miller in the second spot only against right-handed pitchers, but the lineup construction could be causing the team’s lack in production.
The St. Louis Cardinals are breaking all of the trends that are listed above. The Cardinals currently get their best production out of their leadoff spot. They get their second best production from their eighth spot. Their fourth spot is barely out-producing the ninth spot, generally populated by the pitcher. Even with these oddities, the Cardinals have the second highest runs per game as of this writing – how? Brandon Moss is currently hitting 12 for 33 with five home runs with RISP while Stephen Piscotty is 28 for 55 with RISP. Matt Adams has had an average of 0.728 runners on base, one of the top totals in the league, and has scored 20 percent of those runners, a fantastic total that paces the league. There are many factors that would lead one to believe that the Cardinals are greatly fortunate to the small sample size. Cluster luck has also been much in the Cardinals’ favor. The Cardinals are having the right people up to the plate when they need it. Much like the Red Sox, the first few months have seen the Cardinals lineup be very productive. Unlike the Red Sox who put their best hitters in the conventional 2-4 spots, the Cardinals have seen their production more spread out through the lineup. Both teams can put out lineups that have the majority be productive hitters, which may be the most important coincidence. As the sample size grows, there may be more normalizing to show which figures rise and which fade away.
A few other league-wide notes:
- The batting order position’s wRC+ value that has the highest correlation to runs per game is surprisingly the eighth position. A possible theory why is that teams who have production throughout their lineup will score more runs.
- The next highest correlation goes to the fourth spot, followed by the second spot, making sabermatricians happy everywhere.
- The third spot has a much lower correlation to actual production, yet teams still continue to bat their best hitters third. Most noticeably, no team is getting awful production from their third spot, while some teams still do in their second and fourth spots. The team with the lowest production in the third spot is the Cincinnati Reds, where Joey Votto has had a disappointing season (note: he’s been much more productive since a recent move to the second spot).
- Eight teams get their highest production from the third spot (none being in the top nine of R/G), six from the second (three of the top eight R/G), while only three get it from their fourth spot.
- The Chicago White Sox and the Oakland Athletics get their lowest production from their second spot, while the Washington Nationals and the Kansas City Royals get their lowest production from their leadoff man.
There are many ways to build a lineup and have a successful offense, but teams analytic departments will continue to try to create an edge. The numbers shown below display the 2016 wRC+ values (through 6/14) for each spot in every team’s lineup. Hover above (or click if on mobile) to see the actual values. The colors show where the strength of each lineup lies – the more green, the better; more red, the worse.