"While both issues could probably be avoided by investing more effort into the round preparation, I think they actually highlight one of the main good aspects of competitive algorithmic programming: it's extremely objective. Every solution is judged against the same testcases, under the same time and memory limits, and completely automatically." — in my opinion that is exactly what makes it subjective. Objectively good solutions definitely should pass all possible testcases. I would definitely want to live in a competitive programming world where problemsetters don't have to worry about how stupid solutions contestants can create and what to do in order not to let them pass. Specific choice of testcases is what makes everything subjective. However I think that I've read somewhere that people on FBHC were given significantly different testcases, some of sets of testcases weren't including some special cases which other testcases covered and that is of course even worse.