Not a real answer, as it doesn't answer you question -- but why not change the question instead? Surely you can augment it enough that it fits the same basic principle, but it shows the person is thinking instead of reciting?

Have you reached out to the recruiter to point out that they seem to be doing more harm than good? Beyond that, I'm not sure that I see the issue. If a developer doesn't trust themself to write FizzBuzz in an interview and decides to try to memorize a non-obvious version and can't reproduce the non-obvious version they've memorized, would you really want to hire them?

Have you confirmed that your Fizz-Buzz question correctly weeds out good developers from bad developers in your organization? Only good devs can pass? Only bad devs fail? "The issue is some people that memorized the one-liner did ok during the job interview, but the one-liner doesn't compile/run, so I fail them." makes me wonder.

My code never compiles from the first time, even for the simplest syntax mistakes. I am still a good programmer. Then, a few unit testing for the logic, and ship away.

What is the problem that you are trying to solve? meta.stackexchange.com/questions/66377/what-is-the-xy-proble‌​m

It seems that your screening system is highly effective at quickly identifying people you don't want to hire. What's the problem?

Are you sure the candidates are even actually trying to write a 'clever' solution? It can be hard to tell the difference between that and being unable to break the problem down into parts and getting stuck on trying to do it all in one step. The latter is fairly common among inexperienced developers.

VTC as unclear what you are asking. I am unable to understand the problem you are trying to solve here.

Fizz-Buzz should be used as the kind of problem you should ask in an interview to show minimum proficiency. You should not actually ask Fizz-Buzz itself.

Ever time I get asked the FizzBuzz question in an interview. My first thoughts are "oh boy, here we go again". A candidate who fails FizzBuzz leaves a bad impression, but the interviewer who asks FizzBuzz leaves a worse impression.

@SandraK - mine generally doesn't either. If someone writes Fizz-Buzz with a loop and some if statements, I don't type it in because it's very obvious if the logic flow is correct. One-liners are usually pretty dense, and I'm asking what almost everyone would consider a super easy question so you can demonstrate minimum programming skill.

@cgTag - Several people have flat out failed Fizz-Buzz. As in 20 minutes of writing and have a loop that won't work. This includes "senior" developers.

If a dev still manages to fail Fizz-Buzz after prepping for it, I feel like the question did its job.

Oh, if you'd ask me FizzBuzz, I will take 20 minutes. Or more. If I get to the interview, and you still need to ask me FizzBuzz because you still don't know whether I can even program, you haven't done your homework. And I will take my time and create some monstrous program. (I've probably by then concluded I don't want to work for you)

I don't care about the effectiveness of FizzBuzz. I'm saying it's 2018 and this is an old interview question. There is irony in the OP's suggestion that candidates are googling the answer when clearly the OP googled the question. It's been around for over 10+ years now. Why not follow up with some jQuery questions and test their Windows 95 knowledge. That's why I said "oh boy, here we go again". Been there. Done that and moved on.

For what it’s worth, I retired recently after more than thirty years in software engineering (four jobs, six or more interviews) and this is the first I ever heard of “fizz buzz”

Even an expert programmer may write code on paper that fails to compile/run. Are you giving people a paper coding test and then failing them based on it not compiling/running without looking at the code to assess their understanding and competence? If so, you are failing to do your job of using the test to evaluate their competence. If you insist on running code, you have to give them a compiler. Otherwise, you have to look at the code and think about it.