At one time plans were for a rover to collect and analyze samples from various areas then send only the most promising back to Earth. The best of both worlds. That may however be too expensive a route. The cheaper more gradual alternative is to send landers to several places then a sample return mission to the most promising. All you have to do is look at the barrage of tests thrown at ALH84001 to know that no space probe is going to be able to do that much and that the uncertainty after all those tests means a high probability of unanswered questions within limited robotic probe.

Regards,
Mark