I would sure like to see a cite or two that backs up this belief.
There’s no possible way to determine that number, the reliability of polygraphs as lie detectors in situations where there’s a penalty for failure. To do that would mean that you’d have to know, independently of the polygraph, what was true and what wasn’t, and if you already know that, then the results of the polygraph can’t be high-stakes.
And why is the human operator so essential, and how do you know how good an operator is? How do you train them? If you can train a human to interpret a polygraph, then you can train a machine to interpret it, too (or contrapositively, if you can’t train a machine to do it, then neither can you train a human to do it). Polygraph data is a heck of a lot simpler than CCTV data, and even CCTV data, we’ve successfully made machines that can interpret it.
A successful polygraph operator needs to know how to do one thing to be successful: figure out what the people who hire her/him want to hear.
The technique used by the one person I know who fooled more than one polygraph for employment purposes was 2 martinis about an hour before the test. I hold his CV as evidence of the unreliability of the test.