Stealing 3rd with 2 outs is collosally dumb. Stealing it with less than that makes sense if you can make it. Well sure, if you can make it. Stealing a base ALWAYS makes sense if you can make it. The problem is, there's no way for you to know if you can make it. The risk:reward ratio for steals in general is already so high that you have to have a roughly 75% success rate just to break even in terms of productivity (the number varies depending on the source, but it's usually around 75, isn't it?). Well, that ratio is even higher as far as stealing third. The potential benefit of stealing third is you move a runner in scoring position into... more scoring position. The only real boon is the possibility of sac flies and advancing groundouts, which is minimal. You don't help stay out of a double play, and you don't dramatically increase the likelihood of that player scoring. The benefit of stealing third is therefore much smaller than it is for stealing second. The cost of creating an out is also higher. Not only do you create an out, but you take a runner in scoring position off the basepaths. So it stands to reason that, since the risk:reward ratio is even higher than normal for stolen bases, that it would require an even higher success rate to make it worthwhile, probably at least 85-90% (I've not seen a study done on this, so I'm approximating). But, to make matters even worse, stealing third is harder than stealing second. There are very few baserunners who could reasonably be expected to maintain even a 75% success rate stealing third, let alone 85-90%. It's simply never really worth it to try. I suppose I could be convinced that the scenario David mentioned would be a possibility, but there'd have to be some pretty extreme circumstances (excellent runner and basestealer, awful catcher at throwing out base stealers, terrible hitters coming up to the plate, etc.). Even then I'm not sure I'd buy it.