Three of us sat down to shortlist candidates on Friday. Now I know what to expect it was a pretty straightforward 90 minutes and no surprises. I still think I've uncovered a fundamental failure to understand some basic math when I get into discussions about the scoring system we used last time we had to rate someone. [1] Anyway, I had an insight an over the weekend. The whole process of looking at the applications, shortlisting candidates and interviewing them, isn't to get the right person for the job. It's to give the panel time to adjust to the idea of having to appoint the person who happens to come out 'best' by a restricted set of criteria. I.e. let's rate people against a job spec and try and quantify their performance in interview, but take no account whatsoever (or very little) of personality, who they have to work with, or 'rightness' for the job. You can see why such woolly measures get ignored and I suppose I can give some kind of intellectual assent to the system being the best that there is. But it still seems to me, well, rather flawed. Does this ever get any easier? [1] If you have several 'woolly' criteria to rate between 1 and 10 but *tend* to uses numbers between, say, 4 and 9, and you have one criteria that's measurable and uses the whole scale from 1 to 10 then it seems to me obvious that an outlier on the measurable criteria is going to skew the points of someone who is conservative with marking. Apparently that's not obvious to others even *after* I've pointed it out and shown the mathematics.