Tagged: chocolate

Fun-sized statistics that melt in your mouth

I’m told that there are treatments available for those of us who are compelled like to count and sort our candy by color before eating it. I can stop any time I want, of course, but today’s fun-sized pack of M&Ms gave me pause.

This Unit Not Labeled for Individual Retail Sale

No yellow. There’s a yellow guy on the package, but he’s a peanut M&M. Do the regular chocolate ones still come in yellow? Yes, they do, but not in this pack. What are the odds?

Well, assuming that the little packages are filled from an effectively bottomless vat of M&Ms representing the official color distribution (24% blue, 14% brown, 16% green, 20% orange, 13% red, and 14% yellow) and that every fun-sized pack has 17 candies like this one, the odds of no yellow should be (1-p(yellow))^17, or 0.84^17, which is about 5.16%.

In a much more rigorous and costly investigation, the estimable Josh Madison bought 48 packages (larger than mine, each had an average of 55 candies) of M&Ms and ran the numbers. No word on the number of pepcid tablets he needed, but not one of this 48 bags was bereft of any single color. Madison did, however, find that the actual distribution in his sample was not so close to the published ratio, with a lot less blue and more of all the others, especially green.

On the off chance that you’re still with me, I bet you’re wondering, “ok, professor chocoholic OCD, is there something I can use here?” Well, I think there are two ideas worth remembering if you’re in the business of counting or estimating or forecasting things:

Size Matters. Sample size, that is. If you based your view of the M&M world on my single 17-candy packet, you’d have a pretty messed up view of reality. If you used one of Josh’s 55-candy packs, you’d be a lot better off, but even with 48 such packs, you’d have only partially cracked the code.

Seemingly Rare Events aren’t always as rare as you think. Intuitively, a small pack of M&Ms missing a color feels like a rare thing – how often have you seen it? But the math puts it at 5%, one pack in 20. Huge businesses are built on the preferences of market segments smaller than 5% of the population. The United States contains about 4.5% of the world’s people.  And sometimes rare events are even rarer than you think. What are the odds of picking up a fun-sized pack that’s all blue (the most common color by the official stats)? Do you think that would be 100 times rarer, maybe 10,000 times less common than no yellow? How about 0.0000000029% or about 1 all-blue pack in 34 billion, compared to 1 pack in 5 for yellow-deficiency? If all 400 million M&Ms made each day were put into 17-candy packs, they’d make only 8.5 billion packs a year. That would be a rare event.

The odds of a rare event happening are 1.0 after it happens. No matter how crazy it might seem, events with long odds can happen, and once they happen, they have happened. You could choose to believe something’s not right in your calculations or in the world, and you’d probably be smart to check. But once the all-blue pack is in your hands, there it is. Just remember that the odds of getting another one are just as small as they were before.

So think hard next time you put M&Ms into some kid’s trick or treat bag. Who knows what might happen?

Halloween Kit Kat? Give me a break

It’s the week after Halloween, and that means tons of leftover candy. Especially since it’s the stuff you couldn’t palm off on the young extortionists in costume, I usually don’t pay much attention. But this week, somebody left a bunch of Kit Kats in the office, and some of them had orange labels. And we all know that means peanut butter, right?

kit1

Wrong. These orange-wrapped Kit Kats are in fact “Halloween Kit Kat” and they are orange in color but not in flavor.  They’re not peanut butter, they’re not pumpkin pie, they’re not orange Creamsicle, they’re not even candied yam or cantaloupe flavored, they are white chocolate that’s been dyed orange.  It’s a bit like a wafers wrapped in a crayon.

kkit2

This is so disappointing because I know they could have done so much better. In Japan, Kit Kats come in all sorts of seasonal varieties, many of them fruity. Even right here in the USA, there are candies in seasonal flavors like pumpkin spice, so why would Hershey phone in such a weak Halloween Kit Kat?

We gave the world the bacon chocolate bar, can we not put some real pumpkin flavor in a candy bar? Don’t America’s children deserve more squash in their goodie bags?

This chocolate pudding could be A+

I’ve called out food trucks before for obfuscating the name and content of common dishes, but when Mei Mei Street Kitchen put Sanguinaccio Dolce on the board, they helpfully, if bluntly, glossed it with “Taza chocolate, John Crow farm pigs blood.” $2, what could possibly go wrong? If there’s going to be blood in my dessert, I’d rather it be local.

Mei Mei Street Kitchen Menu

Sanguinaccio Dolce (don’t you just love saying that?) is a traditional carnival dish of the Basilicata region of Italy, the arch of the foot of the boot, if you will, or maybe a spat, since it has coastline on both the Tyrrhenian and Ionian seas. It’s more or less chocolate pudding with some fresh pig’s blood in it, sometimes served with biscuits.

Mei Mei provided no biscuits, but did include a nice dollop of cream and some sesame(?) seeds. The characteristic flavor and texture of Taza chocolate was evident in the apparently creamy pudding. There was no obvious or intrusive blood or pork flavor, not even the saltiness that I was expecting. It’s just a subtle twist to the chocolate, a bitterness that I doubt you’d even be able to identify as blood if some evil person gave it to you without disclosure.

Sanguinaccio Dolce

Ingredients aside, $2 for such a small serving of pudding may seem a little steep, but you do get to say that you ate it. Plus, how much chocolate pudding (and blood) belongs in a balanced diet? If you need to load up, go get yourself a venti chocolate cookie frappuccino with two strips of bacon. [N.B., at the time I wrote this I was as yet unaware of the Dunkin Donuts bacon egg and cheese on a donut breakfast “sandwich”.]

I applaud the Mei Mei team for putting something different out there and also for making an effort to use the whole animal. Their menu is ever-changing and seasonal, so get your sanguinaccio dolce while you can. Maybe as summer heats up they’ll add ice cream and call it Sundae Bloody Sundae.

Also, if you haven’t heard, Mei Mei is opening a brick and mortar restaurant, and you can support them on kickstarter.

All candy should come with technical cross-section diagrams

While snagging a fresh Mozart Kugel from the snack table at the office I noticed this informative diagram inside the box. Behold the majesty of two different kinds of marzipan on one chocolate ball.  What really drove the Salieri Kugel to madness was how easy the Mozart Kugel made it look.

Inside the Mozart Kugel

Now that’s my kind of infographic. It’s too bad you typically only get this sort of diagram with German or Japanese candy. To my mind, it should be as required as the nutrition information or the candy guide for the perplexed. Via Steve Almond’s CandyFreak, you can also test your ability to identify candy bars by their cross sections, and there’s a whole load of cross-sectional chocolate fun at Edible Cartography. It should go without saying that I really like that name.