On Aug 11, 2006, at 12:48 PM, James Edward Gray II wrote:

> On Aug 11, 2006, at 2:42 PM, Mat Schaffer wrote:
>
>>
>> On Aug 11, 2006, at 2:39 PM, James Edward Gray II wrote:
>>
>>> On Aug 11, 2006, at 7:58 AM, Ruby Quiz wrote:
>>>
>>>> 	- Try to fill the biggest board you can with this script, in a
>>>> 	  reasonable amount of time (let's say 48 hours minus scripting  
>>>> time)
>>>
>>> [snip: lots of fast filled boards]
>>
>> Do you by any chance have a background in artificial  
>> intelligence?  That's really impressive how quickly you cranked  
>> out a solution to a problem I can't yet see a decent strategy  
>> for.  Hats off to you, my friend.
>
> I bet you are going to be very surprised when I post my pathetic  
> little toy.  If it's good at anything it's cheating...  ;)

I've done a similar problem before, years ago called a Knight's Tour.  
You move a knight around a chess board and try to visit each square  
once. I've seen someone do it live, blindfolded which was pretty  
cool. Here's a brief English description of one technique I used.

MINOR SPOILER WARNING

One simple technique that helped get decent results was to keep track  
of how many still-open squares are connected to each square. Then  
when choosing what square to visit next, choose the square with the  
lowest number of open, connected squares. Corners are only connected  
to two others squares at the start (for the knight's tour). This  
makes sure if you move adjacent to a corner, you'll move to the  
corner next, which helps avoid being struck (it's not strictly always  
necessary because you could end in a corner). I can't remember if it  
ever had incomplete runs using just the open square scoring technique  
or not.

-- Elliot Temple
http://www.curi.us/blog/