About programming competitions and selections

I found this very interesting post l about online competitions for computer programmers.
Taking part to the 2015 CPAN Pull Request Challenge, and agreeding totally to the post author about how useless these competitions can be, I want to enforce the concept.
Nowdays softwares are very complex beasts, and it is much more important to have a look at how to solve a problem in a way that can be maintanable, well documented, self explainatory, and so on, rather than having a bunch of code that performs the right computation with a strange and not well known alghoritm. Moreover, with the ubiquity of alghoritm libraries, I don't see the whole point in proving your knowledge of alghoritms anymore.

I remember at least three job interviews I made a few years ago when I was asked to solve a problem implementing a mathematical alghoritm. And I failed.
But I also remember at least two of the interviewers telling me they don't know git and FreeBSD, or don't knowing the difference between a log-shipping replication and a streaming one.
I tend to prefer to know a little about a lot of things, so to be able to choose the right tool at the right moment (and improve my skills on demand), rather than knowing a single tool/paradigm very well.

So I don't see the aim of having an online competition on alghoritms, or even asking anymore alghoritms (except if your business is based on those). Rather, I strongly believe that being able to prove you collaborated in FLOSS projects, have pushed changes and commits to real code makes you resume stronger.