I was at the market with friends today and right at the check-out counter, I noticed small boxes of Harry Potter-inspired Every Flavor Beans (made by Jelly Belly, so they're jelly beans). On the box was a loud advertisement that these Every Flavor beans come with two extra flavors. Somehow, despite the fact that they had every flavor before, they now have two more than before. I think I got a few odd looks at the market when I was gleefully pointing this out, but I typically get some odd looks wherever I go, so I ignored them.
One of these two new flavors was bacon, so apparently Jelly Belly has figured out the solution for a puzzle that one of my cousins and I had previously considered: Namely, what combination of flavors can one use to get a meat-flavored jelly bean? (Naturally, we had nothing better to talk about the night before my cousin's sister's wedding. Actually, it might have been the night of the daytime wedding. It was one of those two nights.) I wonder whether the solution is unique? Only existence has been established.
Have you tried the meat-flavored jelly bean, bacon? I was curious, so I bought the Every Flavor Beans on sale at Walmart. Gross! Licking cement tastes better.
ReplyDeleteI didn't actually buy it, but I was tempted to get one to send to my cousin.
ReplyDeleteI don't have the same experience licking cement that you apparently do, so I can't make any comparisons here. The closest I can do is to mention the song "Lick the Pavement" by Garbage.
I can't say that I've licked cement, unless I have suppressed memories of doing so during my childhood. I haven't heard any music from Garbage in a couple years.
ReplyDeleteIf you ask me, the easiest way to make a meat flavored jellybean is to make it out of meat! How come no one comes and asks me when they hit these "hard" questions?
ReplyDeleteThere are some configurations where the additional two flavors would be observable... Consider, for example, that the probability distribution isn't flat. In particular, consider that p(n) (the probability of drawing flavor # n) = 1 / 2^(n+1) (for n = 0,1,2...). Inserting the new flavors at any finitely-numbered index would be observable with a reasonable probability in finitely many samples.
During the 2002 discussion, it didn't occur to me to consult expert Techers. We kept it in-house. Besides, how would I look in front of my cousins if I actually had to ask for help? (Instead, I just said I didn't know and gave up, so I looked much better in front of them...)
ReplyDeleteThe thing here is that it isn't just infinitely many flavors or a full measure of flavors. Both of those cases could certainly lead to an insertion of finitely many flavors (which would have finite measure).
In terms of making the jellybean out of meat, the problem we had was slightly more difficult, so I should phrase it more precisely: Given some Jelly Belly assortment of jellybeans (which is what we had available to test at the time), devise a constructive algorithm to combine those flavors to obtain a meat flavor when eating them. (Hell, this is starting to sound like a quantum computing problem. Once we measure [aka, eat] them, we can't test them again. I wonder if we can use entanglement here?)
I'm personally waiting for the "Bombay sapphire martini on the rocks with olives" flavored jelly bean.
ReplyDeleteAh, I getcha. I suppose then it's a question of whether the Jelly Belly basis vectors sufficiently span the space. Given the subjectivity of human taste, a solution in the least-squares sense would probably be good enough.
ReplyDeleteAs my shortcut suggestion I was just imagining bean-shaped chunks of beef jerky.
Hmmm... the idea of resolution of human tastes is a good point as far as effective dimensionality is concerned. (And Randy was complaining that the lunch conversation was too nerdy...)
ReplyDeleteTo be precisely, Randy was amused and, in fact, thought that it wasn't too nerdy. But I digress. :P
ReplyDeleteprecise, even.
ReplyDelete