As Jerricho pointed out, there is an azimuth error inherent in TCAS (30 degrees iirc).
That's where the danger lies in using TCAS to 'identify' target aircraft.
The resolution of the clock code used for giving traffic info for visual acquisition is also 30 degrees. If you're really worried about the crew failing to tell the difference between 2 contacts that cannot be distinguished on TCAS, would you rely on the crew to get the right aircraft visually?