> but I am sure their test incorporated a great deal of detailed knowledge of algorithms used in the computation Actually no. In this case it exploits the know

> but I am sure their test incorporated a great deal of detailed knowledge of algorithms used in the computation Actually no. In this case it exploits the knowledge that "low frequency inputs tend to explore more program space", It was found using numbers with a low transition probability between zero and one. This is an approach also used for digital circuit verification too. It's true that I also know the code is full of hard to hit branches for carries, which does partially motivate that selection. But importantly, "We weren't even testing for that." the OpenSSL code in general has an entirely different algorithm than the libsecp256k1 code; and I hadn't recently looked at any OpenSSL innards.