My family did a test at a Christmas party just a few days ago. We did a blindfolded tasting of four cheap wines, two red and two white. The goal was solely to distinguish between red and white wines. On average, each person made one incorrect determination (n=8, avg=1.1, mode=1).

It wasn't very scientific - it wasn't double-blind, each person knew the number of reds and whites, and it's pretty clear the order in which the wines were presented made a difference. In any case, I think our findings were reasonably consistent with the article.

It was also a fun time.