One thing we’re especially unsure about is whether our double-blind commit model still holds up if one party delays reveals strategically. We modeled this, but I’d love critique from cryptography folks here.