Ichiro & Stolen Bases

Inspired by Larry Stone's awesome entry and some of the subsequent comments, I thought I'd revisit an old subject - that of Ichiro's stolen bases and clutch situations. We've all heard the popular criticism. Ichiro runs, but he doesn't run when it's important, preferring instead to pad his numbers when the game isn't close. Or something. It's one of those weird things people throw out there as a justification for disliking the guy when they can't find a better and more significant reason since his numbers themselves are terrific.

Guess what? As we've mentioned before, it isn't true.

Baseball-Reference.com provides some handy leverage splits, with each player's numbers getting separated into High Leverage (1.5+ LI), Medium Leverage (0.7-1.5 LI), and Low Leverage (below 0.7 LI) situations. Here are the results of a comparison between Ichiro and the 2001-2008 AL average:

Ichiro's distribution is very similar, and his success rates blow the league averages out of the water. So not only is he running when it's important - he's running successfully. The difference between an 80% success rate and a 72% success rate in high leverage situations is huge, especially when you run as often as Ichiro does. He's pretty good. It's worth pointing out that the average LI of Ichiro's SB attempts last year was 1.43, and the sum WPA of those attempts was 0.595. While some people may choose to criticize Ichiro for his behavior on the basepaths, the fact of the matter is that his stolen bases provide a great deal of value.

Could Ichiro run more often? Sure he could. He could try to steal every single time he's on base with an empty base in front of him if he wanted to. But there's a game theory component that comes into play here that prevents him from doing so. Speed isn't the only factor in determining whether or not a steal attempt is successful; there's also the element of surprise, and the more Ichiro runs, the more the pitcher and catcher will be looking for him to run. And the more they look for him to run, the less successful he'll be. It's the same reason Ryan Feierabend doesn't try to pick off every single runner at first base. Lose the surprise and you'll lose the success.

I don't see any reason to ask Ichiro to change his basestealing habits, because everything he does has worked exceedingly well, and you don't mess with success. Run whenever the hell you feel like it, little dude. By this point I think we can trust you to know what you're doing.

Other stuff:

• It's interesting to see how average SB success rates change with the leverage. Since 1998, the Major League average success rate is 72% in high leverage situations, 68% in medium leverage situations, and 72% in low leverage situations. My suspicion is that, in high leverage situations, runners are more cautious, pitchers are more focused on the hitter, and catchers may be less likely to attempt or hurry a throw, leading to greater success. In low leverage situations, I imagine the same is true of pitchers and catchers. I wonder, then, if 68% might best represent the league's "true" basestealing ability, in that both the runners and defenders are acting normally.
  
  
• A graph:
  
  
  Stolen bases are down. Overall, there were more than 4000 attempts each season between 1998-2002, but since then, the Majors have yet to reach that mark. The trend is more evident in the AL, where we have a seven-year streak of ~1800 attempts per season. Between 1998-2001, there were an average of 2196 attempts per season. Between 2002-2008, that figure dropped to 1800, a decrease of 28 attempts per team. Along with the drop in attempts we have seen a corresponding increase in success, from 69% in 1998-2001 to 71% in 2002-2008. 
  
  In 2001, 26 AL players stole 20+ bases, 12 stole 30+ bases, 3 stole 40+ bases, and the leader had 57. In 2002, 15 AL players stole 20+ bases, 6 stole 30+ bases,1 stole 40+ bases, and the leader had 41.
  
  As far as the NL is concerned, there were an average of 2384 attempts between 1998-2002 and 1979 between 2003-2008, a drop of 25 attempts per team. The success rate, meanwhile, has jumped from 68% to 72%.
  
  Over the past half-decade or so, basestealers have definitely become more successful, if a little more cautious. I wonder if this is a cyclical thing, where the increased success leads to more attempts, which leads to more failures, which leads to fewer attempts, which leads to a greater rate of success, and so on. It'll also be interesting to see if the blossoming renewed interest in fast defensive players leads to a little more emphasis being placed on the stolen base.
  
  
• Average SB attempts, 1998-2008, AL team: 139
  Average SB attempts, 1998-2008, NL team: 135
  
  Why are stolen bases considered a part of "NL-style baseball" again?