Like Finagle said; space exploration is difficult.
When things go wrong on a spacecraft or probe, it’s so very easy for them to go completely wrong, due largely to the harsh environment that the equipment has to endure
It’s possible to engineer probes to be more durable (bigger, stronger, components) or reliable (more redundant systems) but typically this means that weight-for-weight, they are less functional (if they work at all, of course) - the more payload weight you squander on making the thing solid, the less you have left to cram the actual science stuff into.
So a break-even situation must be reached; it might be decided, for example, that launching three relatively cheap and flimsy probes, each with only 75% probability of success (as far as it is predictable anyway) is preferrable to launching a single, expensive, chunky one with 95% probaility of success, but only the capacity to do half the experimental science that each of the three cheaper ones could do.
It’s different when there is a human crew, I’m sure.