It helps to be fast to steal bases, but stolen bases certainly aren’t all about speed. So much of the steal is about reading the pitcher and knowing when to run. As Vince Coleman told Anna McDonald of ESPN last year: “You don’t have to be fast to be a great baserunner…You just have to be smart, alert, aggressive and [able to] anticipate.”

As Coleman noted in that piece, it was the tendencies of individual pitchers that gave him the information he needed about when was the best time to run. In an interview with Blake Murphy of Vice Sports, Eric Thames credited his sudden surge in steals in the KBO to first base coach Jeon Jun-Ho, also known as the “Stolen Base King of Korea”, who taught Thames many of the key tells to watch out for from the pitcher.

It therefore seems logical that being more familiar with the pitcher should help, so consequently it seems reasonable to assume that the longer a player is in the league, the more successful they should be at stealing bases, at least up to the point where their physical tools start to decline and offset the benefits of experience.

As a result, I’ve always expected improvement from speedsters who have a strong minor league track record and good scouting reports, but a poor stolen base percentage in the first year or two of their career. While I can certainly think of anecdotal examples of young, fast players getting caught more often in the early years of their career before achieving a much higher success rate later on – Billy Hamilton, for example – I have no idea if that actually holds true. Time to find out.

The overview of the data doesn’t suggest that there’s any reason to expect an improvement in stolen base success as young players adjust to the league. Using data from the Baseball Reference Play Index, since 1998 the average success rate for players with at least 5 stolen base attempts per year in any of their first five seasons is as follows:

Year 1: 70.41%
Year 2: 68.68%
Year 3: 68.45%
Year 4: 70.16%
Year 5: 69.33%

No trend at all, and given that this is only the initial five seasons and only includes players with a minimum of five attempts, it should eliminate the effects of aging or players who switch teams. Less than 5% of these seasons were from players in their age-30 season or older; 68% were players 26 or younger, the peak of the stolen base aging curve according to FanGraphs.

If you total all of the attempts and steals and then calculate the average, giving greater weight to the high-volume basestealers, there is a slight improvement in year 4 and 5, around 2%. However, if only high-volume stealers are included by limiting the sample to players who have attempted 100 or more steals over our five-year sample, there’s essentially no change from year 1 to year 5.

The overall success rate of 75.95% is unsurprisingly much higher than the original sample, but with the year 1 success rate at 76.28% and year 5 at 76.24%, any benefit that basestealers might have from experience seems to be negated by other factors. These likely include the fact that the players are slowing down, and that this process works both ways: just as the runners are getting familiar with the pitchers and catchers, the battery members become more familiar with the runners.

The first year success rate is actually second-best in the sample, behind year 4, suggesting that youth and unfamiliarity are perhaps just as beneficial to budding base thieves as experience with the league is. The fact that the success rate hasn’t really changed by year 5 despite the aging curve does suggest that the learning process is offsetting the physical decline, but there’s also a survivor bias here; the players who do slow down or find themselves unable to sustain a respectable success rate against major league quality batteries have disappeared from the group before year 5.

There’s an awful lot of other noise here too, of course. Players might be playing hurt or simply fatigued. Managers or team philosophies can change, especially based on team fortunes and composition, affecting both the number of opportunities and possibly even the level of patience with a young baserunner who gets caught more often than would be considered acceptable. Players might simply find themselves on in tough spots, against a catcher with a terrific arm or a pitcher with a deadly pickoff move. Some runners come up and are so effective from the start – Jose Reyes, for example – that there isn’t that much room to improve. It wasn’t until his age-30 season that Reyes posted a success rate below 75%, but when he started out at over 80%, there’s only so far to trend upwards.

If there’s too much noise to really see anything from the league-wide numbers, we can at least identify those players who did seem to exhibit some real improvement over their early performance. Splitting the sample in two, I compared the success rate from the first two years with the next three. Every player with a double-digit improvement is below.

Player Y1-2 SB% Y3-5 SB% % Change
Jon Jay 42.11% 70.00% 27.89%
Jay Bruce 43.75% 61.11% 17.36%
Eric Young 67.74% 83.65% 15.91%
Julio Lugo 62.96% 77.78% 14.81%
Yunel Escobar 46.67% 60.87% 14.20%
Billy Hamilton 74.19% 87.20% 13.00%
Gordon Beckham 52.38% 65.22% 12.84%
Jacque Jones 52.63% 64.58% 11.95%
Denard Span 70.69% 81.67% 10.98%
Adam Dunn 67.65% 78.26% 10.61%
Yasiel Puig 59.46% 69.57% 10.11%

Hey, there’s Billy Hamilton again. Hamilton came up, blew everyone away with his speed in a brief 13-for-14 cameo in 2013, then surprisingly got caught 23 times in 2014. He still stole 56, good for a 71% success rate, but it wasn’t until the third year that his success rate reached levels commensurate with his speed. Many of these players had so much room for improvement that all they had to go was go from terrible to simply poor; Hamilton went from great to outstanding.

I’m burying the lede, which is that Jon Jay’s pedestrian career 62% success rate masks a very strange history. Jay went 8-for-19 on the basepaths in his first two major league seasons, making it fairly remarkable that he was allowed to attempt 26 more steals in his third, 2012. Jay somehow managed to swipe 19 bags in those 26 attempts, good for a 73% success rate. 2012 was also Mike Matheny‘s first year in charge. As a catcher, Matheny threw out more than 35% of would-be basestealers, with a peak around 2000-2001 in which he nailed over half. Did Matheny impart some of his wisdom from seeing the game from behind the plate to Jay? Somehow, Jay went from being downright awful to perfectly respectable, but this five-year sample represents almost all of his work as a stolen base threat: he has attemped just five in total in his three years since.

There’s one other name that I can’t just skate past without a mention. Before he became the poster boy for the three true outcomes, Adam Dunn stole bases. Not Hamilton levels of bases, but enough to appear in our group of five-plus attempts for every year of this sample. Dunn might have been 6’6″ and well over 250 pounds by the time he hit the majors, but he was also an incredibly athletic star high school quarterback. In 2002, then-Reds assistant director of scouting Johnny Almaraz told Sports Illustrated that Dunn “ran four flat down the line.” Dunn continued to run in the minors, stealing 24 in 29 attempts for the Dayton Dragons in 2000.

Dunn’s start to his major league career was less auspicious, getting caught three times in his first seven attempts. Following that last caught stealing, Dunn went 53-for-68 from mid-April 2002 to the end of 2007, a 78% success rate. Despite a considerable slowdown in the following years, in which he rarely ran and was caught more often than not when he did, Dunn’s career stolen base success rate is 72%. He improved his success rate in each of his first four seasons in the majors, peaking at 86% and, after dropping back to succeed just two-thirds of the time in his fifth year, went a perfect 7-for-7 in 2006.

All things considered, there’s no real reason to expect any individual player to steal bases more successfully as they get further into their career. Like anything else in baseball, it’s natural to expect some development, but there’s so much noise here that it’s probably more useful to look at a scouting report and extrapolate from a league-average success rate than try to project anything statistically. For most players, the likelihood is that as their other skills improve – including their ability to read pitchers – their speed and acceleration is already declining. For young players who are athletic and fast, just maintain those characteristics for long enough while also learning to pick your spots and take advantage of pitchers, and perhaps you can be as good as Adam Dunn.

Next post:
Previous post:

One Response to “Stolen Bases, Experience, and the Surprising Success of Adam Dunn”

  1. John Shreve

    In Barry Bonds’ last eight years, leaving out 2005 when he played in only 14 games, his stolen base record was 69-11, or 8-1 per year.


Leave a Reply

Your email address will not be published. Required fields are marked *