I can't see how there can ever be a bug in 2). It's double dummy analysis is perfect, and there is really very little logic involved with running a simulation.
Bugs such as Gib randomly throwing a vital high card really therefore have to be due to flaws in 1) - it must be making some inaccurate assumptions based on what has happened so far that causes it to think nothing it does matters.
Fastforward to West on lead at trick 6.
From a daylong tourney, so basic Gib. This means less simulations than advanced, but still some reasonable number.
How could a simulation possibly result in a club lead being equal or better than any other lead? (Clearly not better, which means every simulation had it equal.. but surely that couldn't be.)