You would have to test an incredibly large number of people to rule out a statistical or mathematical variance.
No you wouldn't. ...At least you wouldn't if the product was significantly better (as it ought to be for the price).
40 people tested three times would give you 120 results. If 100 results indicated that the Kamui chalk performed better (when the user didn't know which chalk they were using), then a 5:1 preference would surely be a strong statistical indicator.
Simply keep the cues, tables, temperature etc the same, and change chalk every 30 mins for three or four hours as I say, and you'd end up with some hard results which would indicate that Kamui was either better, worse, or approximately the same as other chalk which costs a fraction of the price.
Of course this is a scary idea to anyone who has paid so much more for a product unless he or she is deeply sure that it's worth the extra money, otherwise the 'Emperor's new clothes' effect might turn out to apply.
Now I'm not saying that Kamui chalk isn't worth the extra money. I'm just saying I'd like to see the double-blind test before even trying a product which is so much more expensive than most of the competition.