Fundamentally, the attacks that StackGuard prevents are not very interesting. They are serious security faults that result from minor programming errors. Once discovered, fixing each error is easy. The significant contribution that StackGuard makes is not only that it patches a broad collection of existing faults, but rather that it patches a broad collection of future faults that have yet to be discovered. That StackGuard defeats the attacks against Samba and wwwcount discovered after StackGuard was produced is testament to this effect.
Using StackGuard does not eliminate the need to fix buffer overflow vulnerabilities, but by converting root vulnerabilities into mild degradation-of-service attacks, it does eliminate the urgency to fix them. This gives software developers the breathing room to fix buffer overflows when it is convenient (i.e. when the next release is ready) rather than having to rush to create and distribute a patch. More importantly, StackGuard eases security administration by relieving the system administrators of the need to apply these patches as soon as they are released, often several times a month.