FWIW, it's trivial to get benchmarks into their comparisons, provided they aren't microbenchmarks, at least in my experience. Saying they pick benchmarks that a

FWIW, it's trivial to get benchmarks into their comparisons, provided they aren't microbenchmarks, at least in my experience. Saying they pick benchmarks that are favourable is untrue: the majority were added because of PyPy performing badly on them, and they've improved as a result of being included thus making them mostly benchmarks PyPy does well at.