That academics who do not want to succeed in doing something tend to be grossly unsuccessful in doing something is weak evidence that it cannot be done.

These academics would be richly rewarded, in and out of academia, for finding human lie detectors and even more so for finding techniques to train people into such things. This is true for all the obvious reasons, and for the more subtle reason that saying '99.75% of people suck and the ones who don't think this are self-deluded' is a negative result and academia punishes negative results.

(Also, bizarre ad hominem with no real world backing. How on earth are you getting upvotes?)

Some business places have a lot of small high value stuff, easily stolen, and a lot of employees with unmonitored access to that stuff. Somehow they succeed in selecting (as close to 100% as makes no difference) employees who do not steal.

'Shrinkage' is and remains a problem in retail; the solutions to this have nothing to do with human lie detectors. The solutions involve filtering heavily for people who have demonstrated that they haven't stolen in the past, summary termination upon theft, technological counter-measures, and elaborate social sanctions. If human lie detectors existed in such quantities or humans were so analyzable, why does do the diamond dealers of NYC resort to such desperate means as dealing as much as possible with their co-ethnics who have decades of reputation and social connections standing hostage for their business dealings?

(Non sequitur; how on earth is this getting upvoted?)

The evidence that people cannot detect lying resembles the evidence that the scientific method is undefined and impossible.

No evidence cited, and what is this juvenile relativism doing here?

The existence and practice of certain business places shows that some people are very good at predicting other people's behavior, even when those people would prefer that they fail to predict that behavior.

I like how this looks like an argument, yet completely fails to include any information that matters at all. 'existence and practice', 'certain business places', 'some people' - all of these are empty of semantic content.

And even assuming you filled in these statements with something meaningful, so what? The point of the OP was not that predictions cannot be made about humans, the point is that the predictions are not made by a hypothetical 'character'. Predictions made by situation are quite powerful, and I would expect that many businesses exploit this quite a bit in all sorts of ways, like placement of goods in grocery stores.

(Non sequitur again; good grief.)