Our results highlight how difficult it is to find an unambiguous association between intelligence and chess skill. When we tested the whole sample of children, some of whom had just recently started to play chess, we found a moderately positive correlation between intelligence and chess skill thus confirming some previous studies ... But when we examined the role of intelligence among highly skilled young chess players we found not only the same absence of the association between intelligence and chess skill that is usually reported among adult chess players ... but also that smarter children had actually achieved a lower level of chess skill.
It's not exactly the same as saying the more stupid you are, the better you are at chess - I hope.
Anyhow. The article is chocabloc with such interesting, often counter-intuitive stuff ("visuo-spatial ability ... was arguably the worst predictor of chess skill among all other abilities we used in this study ... we believe that the common view of the great importance of visuo-spatial ability is a myth" and "Holding (1985) speculated, on the grounds of the observation that many remarkable chess players were journalists and that there was no evidence that visuo-spatial ability is connected with successful chess playing, that verbal ability is more important for chess than visuo-spatial ability" particularly intrigued me) and well worth a read.
Note. The PDF of this article available via the above link is actually a close-to-final draft. The original is available via Science Direct, but will cost you. The copyright publishers insist on does not apply to drafts of academic articles - hence many academics get around this problem by providing a draft on-line, often only with one or two differences from the finished product: a missing comma or two, that kind of thing.