Matchbox Computing

April 1, 2008 6:50 AM

"Once you've constructed your Hexapawn opponent, it's time to start playing against it... If you play this game repeatedly ... you'll quickly notice that your matchbox opponent plays better and better until it is unbeatable!" Martin Gardner created a game called Hexapawn, and also devised an artificially intelligent opponent you can build yourself out of matchboxes and colored beads. Bonus link: An interview with Martin Gardner.
posted by Fuzzy Skinner (17 comments total)

Good post - I heard about something like this years ago but had lost the reference to it. Along the same lines (and very topical for the date) is A.K. Dewdney's Apraphulian Wonder.
posted by jamespake at 7:10 AM on April 1, 2008

This is awesome, as are the CS Unplugged videos inside the link.
posted by DU at 7:14 AM on April 1, 2008

I remembered reading about this in my school library cough cough years ago. I would tell people about it when similar things came up in conversation, but never tracked it down again until now.
posted by Fuzzy Skinner at 7:21 AM on April 1, 2008

24 matchboxes and a bunch of multi-colored jellybeans? That seems cumbersome and physical.

Wouldn't it be easier if we just drew a directed acyclic graph of possible moves, and struck out losing moves? Or represented it as tuples (board, resulting board) in a database table, and deleted losing moves?

What do you mean they're all equivalent representations?
posted by orthogonality at 7:21 AM on April 1, 2008 [1 favorite]

Thanks for this post.

As a nerdy kid interested in math, I loved Gardner's writing, though in the small town where I grew up I had to be driven about 50 miles to get the next book he'd done or even a copy of Scientific American for which he wrote a column. But though occasionally some of the math was over my head in fifth and sixth grade, the clarity and beauty with which he spoke of math and the ways in which he practically applied difficult concepts really spoke to me.

Thus began my quest to learn everything about math that I could. Once again, it was a small town and their weren't many pre-Internet opportunities there, and by 8th grade I'd digested most of what the schools would be able to teach there through high school, but still there was a lot of fun involved in this strange new world that was opening up to me. Topology was especially fascinating.

But the hexapawn article Gardner wrote was always my favorite. And finally as a freshman, I did a science fair project with a friend that took the hexapawn matchbox theory and applied it to computer programming for games. It was of course not that different from what many others had done much better by that point, but we had fun with the project, and after more plays than I like to remember, we had a computer program that played a pretty kick-ass game of hexapawn or tic-tac-toe, among other possibilities we had looked at.

Then came the actual science fair itself. I believe we were the only entry in the computer category at our school. We set up our program, and then as per our backward school's rules, participants were forced to leave the room while judges completed the judging and awarding of prizes. When we came back, we were very disappointed that we hadn't received a ribbon indicating a top prize. Or any prize. But worse yet, we noticed that the computer itself was reset and not running properly. Later, we were to find out that the judges--being barely familiar with computers--had somehow been able to both stop our program from running and delete the program itself from the floppy disk. They had never even seen the program running properly or asked anyone to fix it so that they could see it in action. They simply just assumed it was our fault and moved on to the next table.

I have a feeling that it is this event that provides the foundation for why I ended up not pursuing math professionally and now teach Greek and Latin instead.
posted by zeugitai_guy at 7:28 AM on April 1, 2008 [6 favorites]

zeugitai_guy, thanks for that story. Gardner's math was, and remains, over my head. But he had a way of explaining things more clearly than anyone else. I first became aware of Martin Gardner because he was an amateur magician, and his ideas were often referenced in magic books, particularly by his friend James Randi. I saw his name on math and puzzle books in the school library, and read them cover to cover. Even though my brain just skimmed over the things I didn't get, what I did glean really made an impression on me, and opened up a whole new way of thinking.
posted by Fuzzy Skinner at 7:36 AM on April 1, 2008

There was a science fiction story way back with this exact premise, only it was used by a guy in a spaceship who needed to fool a malevolent computer intelligence into believing he (the guy, not the computer) was still mentally functional despite some mind zapping ray,and for some reason the malevolent computer had challenged him to play a sort of checkers, so he trained his primate helper (chimp?) to do the thing with the beads, and managed to stall the malevolent computer long enough for reinforcements to arrive.
posted by signal at 7:52 AM on April 1, 2008

Is 5 in binary flipping me the bird?

(honestly, though, this is really really cool)
posted by Navelgazer at 8:10 AM on April 1, 2008

There's something I saw as a young child, which I am almost certain was in a Gardner book/article, that hinted at a turn of the century book devoted to mental exercises that would enable you to properly imagine, visualize, and mentally manipulate tesseracts. Never could find the reference again. I often wonder if it was a prank or simply something I imagined.

But coming from Gardner, I almost thought it was possible.
posted by adipocere at 8:18 AM on April 1, 2008

This is cool. I also liked reading Gardner when I was a kid, again also for the writing as much as for some of the math. I had a pile of Penguin reprints of his columns. I remember building a matchbox computer to play tic tac toe; it was a lot of fun. My mum couldn't figure out what I was up to ...
posted by carter at 8:35 AM on April 1, 2008

There was a science fiction story way back with this exact premise

"Without a Thought" by Fred Saberhagen, the first of his Berserker stories.
posted by SPrintF at 8:56 AM on April 1, 2008

Well that's awesome, I was never taught binary that way--conversion seems much easier now.
posted by Citizen Premier at 11:44 AM on April 1, 2008

American kids might want to check into their local laws before trying this, since purchasing that many matches could very well get them flagged as a meth cook.
posted by First Post at 12:01 PM on April 1, 2008

I did this for a science fair with a couple of friends. No one got it, specially the judges, except us. So sad.
posted by Dr. Curare at 1:10 PM on April 1, 2008

If you play this game repeatedly ... you'll quickly notice that your matchbox opponent plays better and better until it is unbeatable!"

This is actually true of any drinking game.
posted by anthill at 5:35 PM on April 1, 2008 [1 favorite]

In any discussion about the difference (if any) between simulated vs real intelligence, the classic Chinese Room question, it's worth putting in a prop for Peter Watts' Blindsight.
posted by aeschenkarnos at 7:52 PM on April 1, 2008

Adipocere: the book you're thinking of is A New Era of Thought, which was written by Charles Howard Hinton in 1888. Hinton's book was indeed mentioned by Gardner in his article on Hypercubes, which is reprinted in The Colossal Book of Mathematics.
posted by steadystate at 3:39 PM on April 5, 2008