Surprisingly, structured fuzzing (mutating inputs to a grammar's generator) has proven less effective than optimized traditional random fuzzing, so far.<p><a href="https://google.github.io/fuzzbench/" rel="nofollow">https://google.github.io/fuzzbench/</a> shows that grammar fuzzers (like AFLSmart) find less coverage per minute than standard fuzzers like aflplusplus, even for the cases where the grammar fuzzer has an extremely comprehensive grammar for that format.