It's not only that. The problem with most theories of eugenics is that they draw from experience with agricultural breeding of domesticated species. Humans are not domesticated; we're a wild species with massive genetic diversity compared to, say, purebred Arabian horses.
This means that with us sexual reproduction still does what it is supposed to do: generate genetic diversity in offspring. Look at large families. You get some who are tall and some who are short; some who have Grandpa Joe's nose and others that have Grandpa John's jaw, others who get both or neither. Even with litter of pedigreed puppies you'll get one total loser and if you're lucky one champion; and pedigreed dog litters are much more alike than any set of human siblings. And that's just physical traits; in terms of interests, talents, and success there is massive variability among siblings, although there is some correlation, in part due to economic circumstances, upbringing and education.
Nature works this way because variability is good for the species, and that variability comes from combinations of genes being shuffled. Add to that the massive behavioral plasticity of our gigantic brains, and the idea that you can sample some of, say, Steve Jobs DNA for successful CEO markers is ludicrous. If you'd raised Jobs in a different family and sent him to a different set of schools, and didn't get him luck out by ending up close friends with Woz, then while he may well have been quite successful in some other way, he wouldn't have been the Steve Jobs we knew.
Of course, willingness to go along with the DNA test is a good test for one phenotypical trait: the willingness to put up with pseudo-scientific baloney.