I'm very surprised to see the definition offered by Robinson; if you can fly to a safe place to land it can't be that serious a problem!
SASless raises some interesting points but I think it is also quite easy to be sucked into the thought process that it's too dangerous to land so carry on anyway. I think the key is how sure you are that the problem actually exists that you are being warned of. If you have a fire warning that won't go out, then I agree that the scenarios SASless paints may well temper your actions if you have any doubt about it being real. On the other hand, if you have a double or triple indication that your MGB is about to fail, then it might be better to risk a ditching, with a good chance of some people not surviving, rather than have it seize at 200 ft as you try and make the platform 10 nm away, and all die.
Your peception of the severity of a problem will also be a function of your own experience and knowledge, which may not (i.e. almost certainly will not) match that of the manufacturer's. You may have never seen a Nicad battery explode from thermal runaway, or seen how long it takes from the torque spikes and kicks from a failing gearbox to the point where the blades stop turning. If you had, maybe you would consider the situation differently.
If you do decide not to land immediately, you had better have a damm good reason why for your widow, and those of your passengers, to understand!
It may all sound a bit heavy and grave, but sadly that's the nature of the consequence of mechanical failures in a helicopter.