To Run or Not to Run…?

Mike Trout has stolen 5 bags through the first 48 games of the 2014 season. Compare this to the first 48 games of 2013 where he swiped 11 bags and to his rookie season where he swiped 19 bases. The Millville Meteor is definitely not flying on the basepaths this year but why? Most of us believe this is a vast underuse of his talent. Is it? Should he be leading off as so many people seem to continually suggest so he can steal more bases?

First let’s look at Trout’s splits leading off and batting second in his last two seasons. In his rookie year he was exclusively the leadoff hitter. Mike put up a .326/.399/.564 triple slash line and a 166 wRC+. wRC+ is a measure of how good of an offensive player Trout was that year compared to league average of 100. In 2013 he played 18 games leading off and 89 games in the second spot in the order. Leading off his line was .325/.398/.519 with a 158 wRC+ which is very similar to his rookie year. When Trout hit second his numbers were .329/.420/.595 with a 181 wRC+.

Simply put, Trout had better numbers hitting second over his last two seasons. Not only are his numbers better but his power is more useful with the leadoff hitter on base in front of him. Lastly, a team’s best hitter should bat second to maximize his run production, minimize the number of outs in a lineup, and maximize his number of chances at the plate. I have argued this in another article on lineup construction on this site, please read that if you’d like to learn more about this. Better yet, read “The Book” or google ‘lineup optimization’ for a ton of articles on why this is mathematically true.

The next question is; if Trout bats in front of Albert, or even Josh, should he be stealing second or not? I recently read a very interesting article from Baseball Prospectus on Fox Sports about this exact issue but talking about Joey Votto and Billy Hamilton. It was specifically talking about lineup protection for Votto provided by Hamilton’s ability on the basepaths. Could Trout provide “lineup protection” for Pujols by being a greater threat on the basepaths?

First we need to understand the run expectancy chart. A run expectancy chart looks at each position a runner can be on the bases and each out state and determines for the past year how many runs a team could’ve expected to score. Each year has slightly different run expectancy for each base position and out situation but each is very similar. For those who wonder why it changes at all, run expectancy changes slightly based on the run scoring environment of each year and Baseball Prospectus keeps each year’s run expectancies on their site.

In 2013 with 0 out and a runner on first the average team is expected to score 0.8262 runs that inning. With 0 outs and a runner on second the same team is expected to score 1.0499 runs. That means if the runner successfully steals second he has added approximately 0.2237 runs to the inning, a little less than a quarter of a run.

What happens if the runner is caught stealing? With 1 out and nobody on base, assuming the runner is thrown out, the average run expectancy drops to 0.2489. That is a decrease of 0.5773. It would take almost 3 successful steals, .6711 runs, to make up for the runs lost by one caught stealing. The exact percentage varies slightly with outs and whether you’re stealing second or third but the generally accepted success rate to make the stolen base beneficial is 75%, or 3 of every 4 attempts. Any success rate less than this and you’re costing your team runs. If you want to see all the different percentages you can read it here on FanGraphs with cool charts.

Stolen base success rates are dependent on a few factors besides the skill of the base stealer. First, the situation has to present itself. Trout needs to be on first or second with the base in front of him open. Trout has had 76 opportunities so far this year. The second limiting factor is the pitcher’s time to the plate. If a pitcher has a good slide step and a quick time to the plate it decreases the success rate of the base stealer. The final limiting factor is the catcher’s arm and pop time. If a catcher is a particularly good catch and throw guy, the best is probably Yadier Molina, then the success rate also drops.

I don’t claim to have any particular knowledge of how many of those 76 opportunities were against good catchers and fast pitchers. What we can look at is Trout’s stolen base (SB) attempt rate. This is found by dividing his total SB attempts (this is SB+CS) by the number of opportunities he’s had. In 2012 he attempted a SB 21.95% of the time with a 91% success rate. In 2013 he attempted to steal 12.27% of the time with an 83% success rate. So far, in 2014, Mike has attempted to steal 6.58% of the time with a 100% success rate. Trout stole almost half as often in 2013 with an 8% drop in effectiveness. He was still above the 75% threshold but it was much closer than his rookie year.

Something else we need to consider to determine if Trout should steal more is when are stolen bases most effective? The three times they are most effective are: late in a game when one run is more valuable than multiple runs; when the batter is a double play threat; and when the hitter is not likely to score the runner from first (i.e. not likely to hit for extra bases.)

3 of Trout’s 5 stolen bases this year have come early in games which is not the ideal time. He hits in front of Pujols who we’ve seen hit into 8 double plays this year. This is far above Albert’s career norm and ties him with 6 other players for the sixth worst mark in the league. Albert’s GIDP tendencies do put some pressure on Trout to run more but Pujols also has been hitting with a lot of power this year. Pujols is tied for 11th in the league in ISO (isolated power) with 14 homeruns. Albert is likely to score Trout from first. For example, Trout has scored 15 of 32 times in his career when on first and the batter hits a double. If Pujols hits a double with Trout on first he will score about half the time. That is significant.

With Trout batting in front of a power hitter (or two) it is less valuable for him to steal. Pujols is a double play threat and maybe he should run a bit more, especially late in games when runs are at a premium. But he should only run when he can be certain he will be successful. None of this leads us to a final answer to the question of whether he should steal more or not. There is still one more variable to check; does it affect the guy at the plate?

Lindbergh, in his article on Fox Sports about speedsters helping the guy at the plate, shows that a base stealer at first base does force the pitcher to throw about 4.3% more fastballs than he normally would. Since hitters prefer hitting fastballs and generally hit better against fastballs this should mean those guys batting while an “aggressive runner” is on base should hit better. Lindbergh found that not to be true. He found that on average those hitting behind disruptive runners hit 25 points lower in slugging percentage and their batting average decreases about 8 points.

“So why is a batter better off with a slow runner or a fast runner who rarely steals on first than he is with someone who’s a constant threat to steal? Stolen bases are distracting. When a runner goes, the movement can catch a batter’s eye at the moment when he’s trying to focus on the pitcher’s release point. And worse, the hitter might feel that he has to take a pitch because the batter got a good jump, or swing at one to rescue the runner from being out by a mile.”

Lindbergh argues that it is even more disruptive to the hitter than the defense. Stolen bases are a weapon but they are a weapon that needs to be used appropriately. In the bottom half of the lineup with slash and dash guys hitting. The stolen base is most valuable later in the lineup. Also late in the game in high leverage situations the stolen base can be very valuable if success is highly likely. But with a power hitter at the plate it generally isn’t as valuable.

This is not to say that Trout shouldn’t steal bases. Mike should steal when he feels he is nearly certain he can swipe it. The odds should be 95%+ in his favor (slow pitcher and average to below average catcher) and when he can do it early in the count. He should also steal bases late in the game when the game is close. It should be used sparingly and for effect.

As a base runner this year Trout is scoring 31% of time even with Albert batting a pedestrian .268. Trout will score and he will drive in runs, even if he doesn’t swipe a ton of bases. His base running skills are still being put to use scoring from first on a double and going first to third on a single.

This is a hard pill for many of us to swallow but Scioscia isn’t completely mismanaging Trout on the basepaths. Trout is not necessarily missing out on opportunities to score a bunch more runs by not swiping bags. He may actually be scoring more runs that wouldn’t have been scored if he got thrown out at second. This may not be popular, but according to the numbers this is best for the team.

Fire away, but read the other articles linked to first. Please understand the data before blasting the article. Much appreciated.