While I don't have a deep knowledge of peer-reviewed literature on transfer of skills from playing tournament chess, I am an experienced tournament chess player as well as a math educator. Based on my knowledge of what tournament chess players actually do, I find it very unlikely that tournament chess would provide any significant boost to mathematical skills that could not be gained by other means. Sure, the mental discipline and working memory development can't hurt, but those things are hardly specific to chess.
The simple reason is that playing a game of chess, even at the highest levels, doesn't involve thinking processes anything like those of graph theory and the Knight's tour, or even mathematics in general. A low-level tournament player will have thought processes centered around tactical calculations a few moves into the future, e.g., "If I capture on e4, then he captures with the bishop, then I recapture with the queen, ..." A mid-level player will add positional thinking such as "If I make him recapture on c3, his pawn structure will be weak." The strong players will be adept at choosing the right broad plan, such as "I'll spend a few moves locking up the queenside, then I'll start a pawn storm on the kingside before he can get in position to defend."
I don't see any way that a task analysis of any of these thought processes gets even close to chess math problems such as the Knight's tour, or to any problems of higher math in general. The thought processes involved in tournament chess are very domain-specific. So-called chess math problems such as the Knight's tour may appear to be superficially related to chess, but being entirely divorced from real chess considerations such as pawn structure, capturing pieces, etc., there is little reason for even a grandmaster to have any special insight or advantage in such problems.