The untranspose . untranspose == id tests only run for singleton lists. The chance of generating a list of equally long lists is small, very small. You should be using a custom generator.

The equal, non null tests do not run at all. The chance of generating a non-empty list of equal elements is negligible. You should be using a custom generator.

The last test is a complete non-sequitur. As a solver, I have no idea what it means when it runs, and I can't look into the test until I've passed it. Given unconsCol _ = Nothing, it just crashes ( head: empty list ). You're assuming things about the output which may not be true.

Please try to write maximally readable code for random testing. Anybody doing maintenance ( this may be you in six months' time! ) will thank you. Resist the urge to use pointfree lambdacases - just write out the stuff, and don't go plucking arguments out of thin air.

The describe wrapper means all your it headers are invisible, because the describe is collapsed by default. Clean, but possibly not what you intended. I wouldn't mind seeing a bit more green on success. :]