To Steal or Not to Steal?

Determining if teams should risk base runners to improve Win Expectancy  

Liquid Biomechanics LLC.

By Operations Interns: Brandon Smith and Colin Campbell 

Project Supervisor: CEO Roy Krishnan 

May 4th, 2021

October 17th, 2004, Mariano Rivera trots in from the mound looking to close a four to three run lead and a three to none series lead to punch the Yankees ticket to the World Series. The Yankees are one win away from the World Series. One of the most iconic plays in the history of baseball is about to transpire: Dave Roberts stealing second base in the bottom of the ninth inning, after he had entered the game to pinch run for Kevin Millar. 

The Red Sox would later tie the game and force extra innings, before eventually walking off the Yankees in the twelfth inning. Roberts and the Red Sox ultimately rallied around his stolen base and the momentum of the series flipped, as the Red Sox finally overcame the curse of the Bambino and brought the Commissioner's Trophy back to Boston. 

Stolen bases, once a common occurrence, have gradually made its way out of the major leagues. Teams have slowly changed their small ball philosophies and have focused on elevating batted balls, in return increasing overall offensive production. Today’s teams have shown that they are not willing to risk losing base runners and take the bat out of players' hands that can potentially yield a positive return. 

As more data has been released during the Statcast Era (2015 to Present), it is clearly evident that teams have become less likely to risk an out by stealing. This could be due to a variety of factors. For example, pitchers velocity has significantly increased which as a result, has generated higher whiff rates for hitters. The increase in velocity has made it more difficult for offenses to string together rallies, as a result teams have sold out for power, as opposed to manufacturing runs. Furthermore, teams have placed a significant value on outs, and are not willing to risk advanicng 90 ft. This can be seen with the small ball example as teams are also less likely to bunt in addition to stealing. Catchers have also become more athletic as teams have moved away from hiding the less athletic players on the field behind the play.     

To further examine the stolen base trend in baseball, we collected data from 2000-2019 and examined how stolen bases correlated with runs and wins. Stolen bases to runs had a -0.04 correlation coefficient and stolen bases to wins had a -0.01 correlation coefficient. Based on our data set, there was no significant relationship which swayed towards stealing or not stealing. From this, we can conclude that stolen bases and the impact that they have on a game should be examined at a micro level and the in game situation, rather than observing stolen bases on a macro level.  

Although there is no positive correlation coefficient between runs and wins there are numerous positive outcomes that come with stolen bases on a micro level.  Besides the obvious example of moving a runner up 90ft, stolen bases apply pressure on defenses. Currently, teams shift more and this often leads to infielders playing out of position, or cause infielders to cover bases in which they are not typically used to playing around (i.e. A third baseman playing behind second base and covering second). The shift also forces catchers to throw to a moving fielder which also creates chaos on the defensive side of the game. Stolen bases also adds additional pressure on pitchers to ensure their time to the plate is quick enough for the catcher to make a play. This could lead to missed locations by pitchers or balks for left handed pitchers.  Furthermore, a successful stolen base has an impact on run and win expectancy. For example, a successful stolen base from first to second in a tie game in the bottom of the sixth increases a team's win probability from 62.6% to 66.8%. This change in win probability has a significant impact on the opposing team, as it may make a manager use a certain reliever earlier in the game rather than saving for later innings.  The advancement of ninety feet increases the leverage for a pitcher as well as the defense because both positions will now be forced to execute in a higher leverage situation. 

On the flip side, there are numerous flaws when it comes to stolen bases. As previously mentioned, the velocity of pitchers makes it difficult for teams to string together rallies. That being said, a base runner reaching first base is highly coveted by a team and a caught stealing significantly dampened a team's win and run probability. Building off of the bottom of the sixth inning example mentioned above, a caught stealing would decrease a team's chances of winning from 62.6% to 54.2%. Furthermore, more times than not a stolen base attempt forces a hitter to take a pitch, which in some instances is a called strike. This alters the count for the hitter and as the count draws closer to two strikes, the wOBA for an average hitter significantly decreases.  Lastly, player salaries are significantly greater in today's game than in the early 2000’s. The risk that comes with stealing a base and a franchise player sliding head first into a base increases the chances of an injury. 

So in looking at individual situations, win probably must individually examined to make independent decisions.

Examining Win Probability: 

4/1/2021 Toronto Blue Jays vs. New York Yankees 

Situation: Bottom 9, score tied 2-2, runner on first 

First situation: Mike Tauchman of the New York Yankees pinch runs for Gary Sanchez following a one out walk. The Yankees were approaching the bottom of their order and were looking to win the game by putting a runner in scoring position. 

The pitcher for the Blue Jays is Jordan Romano. His time to home plate on average is 1.62 seconds. The catcher is Danny Jansen whose poptime average is 2.05 seconds. Mike Tauchman 90ft sprint speed is 3.3 seconds. By stealing a base in this situation, the Yankees win probability increases from 71.1% to 81.1%. If he is caught stealing, the win expectancy drops from 71.1% to 58%. Based on the subpar time to home and catcher pop time in combination with the fast runner, this situation suggests that the Yankees take the chance at the stolen base. As a result, Tauchman ultimately steals second base and is safe. 

In addition to stealing second base, Tauchman successfully stole third. This was an extreme gamble with a high risk low reward at stake. Tauchman's steal of third only increased the win probability from 81.1% to 82.7%. Had he been thrown out, the win probability would have dropped from 81.1% to 53.5%. For this situation to make sense for a team, the stolen base needs to be almost guaranteed in order for this situation to make sense given the marginal reward and high risk.             
Conclusion: 

Based on our findings, it is evident that there is no relationship on a macro level that stolen bases significantly contribute to a teams success in terms of runs and wins. More so, stolen bases should be examined on a micro level and in game situations when the reward of stealing is greater than the risk when relating back win probability. Furthermore, stolen bases should be attempted when the runners sprint speed is greater than the pitchers time to home and catchers pop time.