Stolen Bases and Batting Order by Scott Spratt December 14, 2017 I’ve already done some research on batting order’s impact on RBI and runs scored, and if a fantasy player is indifferent between the two—the way he likely would be in drafting a new team in a traditional roto format—then there isn’t a ton at stake. Over the first five spots in the lineup, a typical fantasy-relevant batter can be expected to gain or lose at most 6 RBI plus runs in a full season. Intuition tells me that is not the case with stolen bases, and that isn’t just because managers like certain archetypes of hitters to bat in specific positions in the lineup. Runners really only have a chance to steal bases when certain conditions are met. To make things easy, I’ll call it an opportunity whenever a runner reaches first base with fewer than two outs and second base unoccupied. Those opportunities are not evenly distributed among the spots in the lineup because every team starts every game with 0 outs and at the top of their lineup. You can see what that does in a table of the average number of stolen base opportunities per season from the last few years. Stolen Base Opportunities / 162 Games Bat Order AL NL 1 115.3 122.4 2 104.2 111.5 3 76.4 80.6 4 82.2 78.8 5 83.0 84.2 6 77.4 79.3 7 72.3 75.4 8 75.2 73.4 9 70.6 57.2 Each subsequent spot in the order should see fewer plate appearances over the course of the year, but stolen base opportunities do not decline uniformly. Instead, the No. 5 hitter has more opportunities than both the No. 3 and No. 4 hitters, especially in the NL. That is because so many first innings end after the first four hitters, which then eliminates the chance the No. 5 hitter could reach base with 2 outs or with a runner on second base in the second inning. Managers don’t treat the No. 5 hole in the lineup as a secondary leadoff spot, and, probably, they shouldn’t. A team’s goal is to maximize their runs scored, and sometimes, the No. 5 hitter comes up with multiple runners on base in the first inning. A manager needs to balance all of those concerns. But it would certainly help fantasy owners if managers batted their stolen base threats fifth instead of third or fourth, or at least it seems it would based on the actual trends. The problem with those trends is they are tainted by the real quality differences in batters at the various lineup spots. Look no further than the low 57.2 stolen base opportunities for nine-hole hitters in the NL. Turns out, pitchers aren’t great at reaching first base. To really determine how much batting order impacts a player’s ability to steal bases, we need to again dive into some hypotheticals. In my previous research, I established an expected number of plate appearances for each lineup position. Then, the model batter I have been using has established chances to hit a single (18.0 percent of his plate appearances) and walk (8.0 percent of his plate appearances), and I’m going to add a small 0.5 percent chance he reaches first base on a hit by pitch. If we assume that those events are randomly distributed, then the question that remains is how often those events result in a base-out state with fewer than 2 outs and second base unoccupied. I calculated those rates from the last few years split by batting order. These rates will allow me to calculate an expected total of stolen base opportunities for the model hitter—or for any hitter—while also accounting for the likely quality of batters around him based on his place in the order. How Often 1B/BB/HBP Become Stolen Base Opportunities AL NL Bat Order 1B BB HBP 1B BB HBP 1 60.7% 52.5% 54.6% 62.5% 54.7% 50.5% 2 57.8% 45.0% 44.0% 57.9% 47.9% 48.8% 3 44.3% 31.6% 28.8% 46.2% 31.9% 29.8% 4 50.5% 38.2% 39.1% 47.8% 36.9% 35.4% 5 52.0% 42.9% 45.8% 52.2% 43.5% 41.6% 6 50.2% 40.2% 36.4% 50.0% 41.5% 43.5% 7 48.7% 40.4% 36.8% 49.2% 40.3% 40.8% 8 51.2% 41.4% 42.5% 50.0% 40.8% 41.3% 9 50.6% 41.3% 41.3% 51.9% 44.7% 37.0% The peak in these rates around the No. 5 hole matches the trend we saw in the previous table and provides further evidence of the narrative I assigned to it. You can also see that the weirdness around the No. 9 hole in the NL has mostly vanished. The massive dip in stolen base opportunities there are clearly the result of pitchers’ inability to hit. Now that I’ve established the model batter’s chances of reaching first base, the chances that reaching first base will create a stolen base opportunity, and the expected number of plate appearances of each spot in the order, I simply need to multiply them together for the model batter in each spot in the lineup. I’ve split the results into separate tables, the first for the AL and the second for the NL, but the differences between the leagues turn out to be pretty small. Expected Stolen Bases by Lineup Spot, Model Batter, AL Bat Order PA SB Opps SB – Fst SB – Md 1 691 106.4 53.2 31.9 2 691 98.3 49.1 29.5 3 670 71.3 35.7 21.4 4 652 80.5 40.2 24.1 5 631 82.2 41.1 24.6 6 615 76.5 38.2 22.9 7 601 73.2 36.6 22.0 8 584 74.4 37.2 22.3 9 563 71.0 35.5 21.3 The leadoff spot is the unsurprising victor for most stolen base chances, but the No. 2 hole is much closer to it than to any other spot in the order. That means that, as was the case in the RBI and runs scored research, it’s not as big a deal as I expected for the critical member of your fantasy team to bat second. If he’s a Dee Gordon or Jose Altuve type, then he may lose about 4 steals over the course of a full season. If he’s more like Mike Trout or Lorenzo Cain, then it will only cost him 2 stolen bases. However, if he falls further in the order, then he stands to lose 10 or more stolen bases if he’s fast. Expected Stolen Bases by Lineup Spot, Model Batter, NL Bat Order PA SB Opps SB – Fst SB – Md 1 691 109.7 54.9 32.9 2 691 100.2 50.1 30.1 3 670 73.8 36.9 22.1 4 652 76.5 38.2 22.9 5 631 82.6 41.3 24.8 6 615 77.1 38.6 23.1 7 601 73.8 36.9 22.1 8 584 72.8 36.4 21.8 9 563 73.8 36.9 22.1 In the NL, the tail in the 7-8-9 spots is flatter, but that doesn’t mean much in practice. Pretty much every NL team bats their pitcher ninth, which means that aspect of this research remains purely hypothetical.