I'd actually bookmarked that node for use when I rewrote the Jotto programs (soon, I must find time to do this soon).

As I find myself saying a lot, I really do know a lot more know than I did when I wrote that code. I'm glad it proved useful in motivating you to improve it, though.

I did some messing around with a "most constraining" heuristic. I did a bunch of tests, and then, after many days of running tests, I discovered that I had implemented the algorithm incorrectly. By that time, I didn't feel like going back to fix it. Rewriting the code might get me to try it again, and I think I can probably cut down on the time it takes to play a bunch of test games.

One thing that I did, though, was only "turn on" the heuristic once the number of possible words dropped below a certain threshhold value. It just takes way too much time to think about it, otherwise, and, really, it doesn't matter too much for the first couple of words, anyway.

I'd be interested to hear about what you come up with for improving the guessing algorithm.


Queue/Kevin

BTW, anyone up for an e-mail game of Jotto? I'm currently playing X-Jotto (secret word can be anywhere from 3-8 letters, score as normal), which I find much more interesting than regular Jotto, but I'm happy to play any variety.