Unlike some of our stats masters on the site, I haven't made a study of this, but stealing third seems to be a much higher percentage strategy in PB than in real baseball. I have spot checked success in stealing second, and it seems to run pretty true to form that the varibles of the caught stealing ratios of pitcher and catcher and the steals percentage of the baserunner run somewhere close to where they should be.
Bunt singles in the game appear random. Lots of players bunt for singles who wouldn't do so in real baseball; in a recent series Albert Pujols did that to me. Probably bunting for a hit occurs rather randomly for pitchers as well. I think it's been explained on this board before that the algorithm controls whether a hitter gets on base, but the calls are random. The same seems to be true, for example, in hitters' having opposite field power.
Tomj is right that a delayed steal of home used to be a high percentage play in PureBaseball, much more so than in real life, and that that seems to have been fixed.

Osceola Dave