Games Make the Man

There are a lot of things in the world. Trillions of cells in your body, for example. But the number of ways to arrange these things swamps the actual number by an unfathomable margin. Suppose there are an even trillion cells in your body. If you scrambled the locations of these at random by swapping each cell with some other chosen in a cosmic lottery, what are the odds that after two such iterations, you'd end up back with all cells in their original position? So you scramble, scramble again, and hope for the best. You can think of it as the first operation randomizing everything, and the second giving a (very small) chance of putting it right. So what are the odds?

One in a trillion factorial. That's two times three times four...up to a trillion. The number of subatomic particles in the known universe doesn't come close to that number.

My point for this post is that connections and relationships are important. This sometimes has unexpected effects. I have two examples, both of which involve game theory. Both come from the 20 January 2007 issue of New Scientist.

1. The editorial (sub. required) describes some medical experiments showing that a chemical called dichloroacetate might be a potent anti-cancer drug. It's currently being used to treat other diseases, and basically enhances mitochondrial function. Most cancer cells don't use mitochondria to provide their energy (it requires a blood supply, which may or may not exist), so they rely on glycolysis. The drug switches the cancer cells' mitochondria back on. They apparently wake up, take a look around, and self-destruct the cell.

The twist is that because this drug has been around for a long time, it's unpatentable. Therefore no drug company is going to spend millions on a drug trial. They might well spend years and millions developing a similar drug. But not this one.

2. Blue-eyed men prefer to mate with blue-eyed women.
[M]en with blue eyes are drawn towards blue-eyed women, and prefer to choose them as their partner because this can provide reassurance that the woman's babies are theirs too.
In other words, because the blue-eye gene is recessive, a blue-eyed guy will know something's up if the baby has brown eyes. This is not true for brown-eyed mates.

There's actually another article in the same issue that fits the theme here. It's on people who lie in Internet dating sites, and is a fairly obvious instance of cheating. In this case, people lie, but not too much. They make themselves a little taller (men) or thinner or younger (women). But not so much that the face-to-face meeting will explode the myth.


My Wish for Sticky Conferences

Here's my wish for 2007, vis the Web 2.0 universe we inherited from 2006. I would like, pretty please, a web site that allows data and metadata to be attached to conferences. An example will illustrate what I mean.

I just got back from the annual American Association of Colleges & Universities conference. As usual I made a slew of contacts, have sheafs of handouts and my own notes, the conference program annotated with the sessions I attended, etc. The sessions were good, and the best ones were those that involved extensive discussion during or afterwards. Roundtables are the best, actually. Unfortunately, those conversations tend to get cut off just when they get interesting. Wouldn't it be great if these moments could be tagged so that you could pick them up later? I imagine a site that you log in to, identify the conference, session particulars, etc., and then are allowed to add comments or documents under that rubric.

Some conferences attempt to do something like the above, but it would be better to have a central clearinghouse with a standard interface and a dedicated team making improvements. There would be gobs of advertising money to be made, I'd imagine, since the entropy of the groups addressed is low (people with common interest are interested in some of the same products and services).



Somebody was telling me about a new kind of noise generator used for office spaces. Instead of creating a 'whoosh' of white noise at low levels, it makes vowel sounds. This is supposed to confuse your brain when you hear a faint conversation. The overlapping vowel noises mask enough of the actual voices so that you can't hear--and don't even notice--the voices. Cool idea.

By coincidence I was at a party the other night where a similar topic came up. People were saying that when they were by themselves in a house at night they liked to have a cat around. Failing that, the TV will do. This way if there's a bump in the night, they can blame it on the animal, and not worry about it.

So if you put these two ideas together, you get my noisemaker idea. This box would have several settings. On the Subdivision setting, it would create small sounds like minor house creaks, refrigerators clicking, and other normal house sounds, with an occasional rattle or loud bump thrown in. This way, one could sleep soundly knowing that all noises can be blamed on the machine. On the Farm setting, one could imagine more slamming doors, rattling shutters, and such. The Urban setting would feature sirens and gunshots. Now where's that patent application?