Note that some papers do come close to meeting what we view is the right standard. Table 1 shows that T-Fuzz does pretty well, for example.