Speaking hypothetically, suppose all of the electronics on a JDAMS guided bomb failed, say because of a bad battery. At that point, it reverts to being a dumb bomb. The same would apply for a laser guided bomb. After all, both laser guidance and JDAMS packages are designed to be added to ordinary bombs.

So, what would that do to accuracy? Well, it's reasonable to say that the aircraft would've dropped it expecting the guidance package to do its job, so it might not be aimed as accurately as a dumb bomb has to be. Under those circumstances, the CEP (circular error probability) could be very large. It's also possible that the guidance package might affect the bombs aerodynamic properties, making the trajectory even more unpredictable.

In WWII, there were a lot of factors that reduced bombing accuracy. These included such things as relatively inaccurate bombsights, the planes being tossed about by antiaircraft fire, relatively poor aerodynamics on the bombs themselves, and unpredictable motions of the bombs as they were released from the planes. Even a small lateral motion at the instance of release could throw the bomb off the aim point by hundreds of feet as it fell from 25,000 or 30,000 feet.

I went to squadron officer school in 1991 with an F-111 pilot (he's now commander of a test squadron at Edward AFB). He told me that a good F-111 crew could put a bomb within 100 yards or less of the target every time, and a really good crew could do a lot better than that. Virtually all of the factors I listed above had been improved that much. He was a stan-eval (standardization-evaluation) check pilot. He said that when some pilot tried to make a 100 yard miss sound good, he'd rag them, "Yeah, that'd be OK if you were aiming at the Rose Bowl!"