Let me politely disagree with you. I conduct interviews regularly and often the candidate's perception about of own performance doesn't match reality.
Agreed. In my own case, when I'm interviewing people I frequently ask at least one or two questions that I really don't expect anybody to know the answer to. It's more of a "here's your chance to really extra impress me", but not knowing the answer is in no way an indication that you aren't qualified. One of my favorites of this type is "Can you explain the difference between deadlock and livelock?" Anyway, when asking stuff like that I make it a point to always try to remember and emphasize to the candidate "there are no auto-fail questions here, and not knowing the answer to this does not mean we won't want you", etc. I've interviewed (as a candidate) enough times to understand how nerve wracking it can be, and I try my best to help the candidates not feel intimidated or whatever.
And yet, I think sometimes people stumble on a single question, and suddenly get more nervous thinking that they've "failed the interview". Interviewing is not an easy process, for either side of the table.
> I frequently ask at least one or two questions that I really don't expect anybody to know the answer to.
Just to expand on this: It seems really optimal, in terms of extracting information, if the candidate can answer about half the questions (maximal entropy, if you will).
If the questions are too easy and the candidate nails them all, you're wasting time. If the questions are too hard and the candidate fails them all, you're wasting time. You want to quickly get to the limit of the candidate's knowledge, and spend some time there.
This is particularly true if you aim to cover several areas of interest quickly.
Being told by a candidate that they don't know the answer to something is a really useful thing to hear.
There are times when you want someone to show initiative and come up with a solution on the spot and others when you really just want someone to clearly say, I don't know.
It's always good to have at least one question where someone can demonstrate that intellectual honesty.
Agreed. In my own case, when I'm interviewing people I frequently ask at least one or two questions that I really don't expect anybody to know the answer to. It's more of a "here's your chance to really extra impress me", but not knowing the answer is in no way an indication that you aren't qualified. One of my favorites of this type is "Can you explain the difference between deadlock and livelock?" Anyway, when asking stuff like that I make it a point to always try to remember and emphasize to the candidate "there are no auto-fail questions here, and not knowing the answer to this does not mean we won't want you", etc. I've interviewed (as a candidate) enough times to understand how nerve wracking it can be, and I try my best to help the candidates not feel intimidated or whatever.
And yet, I think sometimes people stumble on a single question, and suddenly get more nervous thinking that they've "failed the interview". Interviewing is not an easy process, for either side of the table.