...is that even though there's the aforementioned high bar for taking second in most cases, the numbers say that there are at least as many situations where managers (even the, ahem, smallball guys) don't take the base nearly as often as they should try.  Sending a runner home from first on a double (ie. the windmill vs. the stop sign) has a break-even point far lower than I think most fans (and third base coaches) would guess, something like 40%.  Which is to say, if you think your guy has even a 50/50 shot at making it ahead of the throw, or beating the tag, or whatever, the numbers say send him!

But since we generally think of baserunning as stealing second (which, to be fair, is almost certainly the most numerous occasion) and the bar is set higher than a lot of managers hit -- 73.5%, as Craig pretty much nails -- the overall feeling that people get from baserunning studies is that the numbers say you shouldn't.

But not only is that not true for the reasons given above, it's also almost never true about tactical studies.  If there's one thing I've picked up from reading MGL's work at The Book blog, it's that everything should be regressed -- but if there's a second thing I've picked up, it's that game theory abounds in the in-game decision making that is being a manager.  Yes, the bar might be 73.5%, but if you're only sending guys who are successful say 75% of the time, and only in situations where guys are successful on average of say 78% of the time, you're setting yourself up for being picked off or pitched out on.  You can't always do the optimal thing, because it's only optimal in the context of every attempt.  If you only steal when the numbers say you should, and never when they say you shouldn't, you're going to come out with sub-optimal results.

Luckily I don't think Fredi is at that point, most likely because I don't think he knows enough.  Sometimes that's for the best.

If it's Hanley or Maybin, send 'em.

If it's anybody else (except maybe Hermida), then don't.