For a land landing, I believe the landing pad *is* the safest crash location.
If the reentry burn is normal but there's a problem in the landing burn, it'll likely crash somewhere on or west of the pad (but not far west, still in the general vicinity of the pad). But if the problem is a failure in the reentry burn, it could be on a much wider range of possible impact trajectories - generally significantly west of the pad. The dV from the reentry burn shifts the ballistic trajectory's impact point to the east in the process of reducing velocity, so if you have lower dV than intended, you have a further west impact.
There's an even worse possible failure case (excessive boostback burn), but it's not nearly as likely. Normally one's failure cases cause too little dV, not too much.
For sea landings, it could make sense to try to avoid drone ship damage, but I'm guessing that they don't want to bother putting in the engineering effort, at least not yet. There are so many ways that a rocket could fail to land.
I hope your assessment is correct, that it's just been lower on their priority list. But when it comes to land landings, proper detection of abort scenarios and handling of them simply can't be allowed to be lower priority. And the fact that they don't detect it for sea landings says they don't for land landings either; if they had the code for land landings, why would they deactivate it at sea?