Could you elaborate? Their website seems to indicate that they are a very polyglot shop, not someone pushing a JVM agenda:
"On the back-end, we use Java, Ruby, Python, .NET, PHP and others based on what makes sense balancing server performance, scalability, hosting costs, development efficiency, and your internal development team's capabilities."
"We have included our in-house Java web framework, Gemini, in our tests. ... "
Then you see results for some languages that are completely out of whack with most other such benchmarks (they themselves mention the weirdness of Sinatra vs. Rails, for example).
Then you see on a couple platforms that more performant mainstream options have been excluded, for no good reason.
Then if you look at the repo, there are deployment choices and code mistakes in some of the other languages which go well beyond elementary incompetence...
Feel free to issue a pull request to help our testing. We are not trying to push Java-based frameworks over any others and we believe we are being fair across the board. That being said, if there are "code mistakes... which go well beyond elementary incompetence", then we would love to correct and retest these.
Maybe you should've had the controllers execute raw SQL for a better comparison. I see you executing a regular query in Rails whereas your Java servlet is using prepared statements.
"On the back-end, we use Java, Ruby, Python, .NET, PHP and others based on what makes sense balancing server performance, scalability, hosting costs, development efficiency, and your internal development team's capabilities."