- It's mathematically flawed bet:
Case of sending 2 ships on the single mission:
+--------+--------+-----------------+
| Ship A | Ship B | Science outcome |
+--------+--------+-----------------+
| OK | OK | 100% |
+--------+--------+-----------------+
| KO | OK | 100% |
+--------+--------+-----------------+
| OK | KO | 100% |
+--------+--------+-----------------+
| KO | KO | 0% |
+--------+--------+-----------------+
Case of sending 2 ships on 2 missions
+--------+--------+-----------------+
| Ship A | Ship B | Science outcome |
+--------+--------+-----------------+
| OK | OK | 200% |
+--------+--------+-----------------+
| KO | OK | 100% |
+--------+--------+-----------------+
| OK | KO | 100% |
+--------+--------+-----------------+
| KO | KO | 0% |
+--------+--------+-----------------+
I'd rather take my chances and get 200% given the same initial investment.
- There is already redundancy:
Lots of redundancy is built-in inside the probe, to ensure it will not fail. Multiple sets of thrusters, reaction wheels, CPU, ... Basically we are already sending 2 ships on one package.
- Same causes causes same effects.
If there is a design flaw in your twin spacecrafts, it's likely that both your spacecraft are doomed: you couldn't learn from your mistakes.