(which is the research underlying many articles and blog posts on this subject) is that Ericsson found results that no other researcher predicted before trying the experiments. In one case, Ericsson developed a training program for undergraduates to increase their "digit span" from the typical seven digits to more than 100 digits. That's a striking result, because "digit span" used to be regarded as a reliable proxy for IQ. (There are still IQ tests that have a digit span item as part of the item content.) Scientists who obtain surprising, replicated results show that sometimes our intuition about how the world works needs correction from factual observations.
Have you tried the experiment? The really striking thing about K. Anders Ericsson's research
http://www.psy.fsu.edu/faculty/ericsson/ericsson.hp.html
(which is the research underlying many articles and blog posts on this subject) is that Ericsson found results that no other researcher predicted before trying the experiments. In one case, Ericsson developed a training program for undergraduates to increase their "digit span" from the typical seven digits to more than 100 digits. That's a striking result, because "digit span" used to be regarded as a reliable proxy for IQ. (There are still IQ tests that have a digit span item as part of the item content.) Scientists who obtain surprising, replicated results show that sometimes our intuition about how the world works needs correction from factual observations.