Should there be the facility to see how much percentage of test cases were successfully passed ?
I disagree with you on this point. People here are extremely clever and it might be possible to deduce information about test cases from that. Just handle a different test case in each run (using assert or “if” statements) and you might be able to extract the nature of the test cases (This is already done in challenge problems i think) like n is large, n is 1 , range of input is small,etc which can be used to devise hacks to solve the problem.This is not very helpful in learning as people will try to take shortcuts to get to the answer quickly.
yes you are correct on this point but sometimes the problem is bit too rough to understand and the explaination for those problems is also not so satisfactory, so this facility may help in to know whether the coder is going in right direction or not.