Probe set algorithms: is there a rational best bet?

Probe set algorithms: is there a rational best bet?