MySQL vs. Neo4j performance on a Large-Scale Graph Traversal

In conclusion, given a traversal of an artificial graph with natural statistics, the graph database Neo4j is more optimal than the relational database MySQL.

I get what they mean, but "more optimal"? Please...

Maximal, minimal, optimal, and the corresponding verbs, require proof. Why not use "better", "improved", "shorter", "reduced", "faster" or "sped up" instead?

I sometimes use Neo4j but I don't know much about its implementation. Assuming that there was a lot of disk I/O involved in this benchmark, does anyone know how graph data is stored on disk, indexed, etc.? This is a very different problem than in-memory graph operations.

Well, luckily I would choose neither of these for traversing a graph, neo4j being bound to java, and MySQL lacking recursive CTEs.