Engr2012
I would not recommend doing that. Try to solve a mixture of easy, medium and hard questions and under timed conditions. These mini mocks are just to let you do timed practice, make you undertsand how the various sections will be worded officially and nothing more. Consolidated scores depend on may different factors in what position was that particular question, did you make a string of incorrect answers, what was the difficulty of that question etc in order to determine what score you should be getting.
I realize that the real algorithm is complicated but there must be
some sense to it. Getting every quant question correct will NOT be a Q30, obviously. And I'd bet that getting 36/37 Q's right is a lot closer to a Q50 than a Q40. We should take estimates with a grain of salt, but we can certainly do better than, "no clue."
FWIW, I think getting around 80% right on the paper tests results in a 700, maybe >95% is a 760+. With a balance of Easy-Medium-Hard I would think that would still be roughly true. Of course, you can make the quiz only Medium-Hard if you're already above average and wouldn't see the Easy stuff on a CAT.
A peson should be humble enough to accept someone else's response than being arrogant to scoff off the given suggestion. If you feel comfortable with your "sense" , go for it.