Toward the end of an interesting see-saw battle against Samurai on 6/1/01 that had some very interesting tactics in the midgame, my move against Samurai (move ...
I've been emailing back and forth for a few months with the organizers of the Mind Sports Olympiad in London. They're interested in having Octi as an event,...
Hi,<br><br>I've made some good progress with the learning algorithm and the net seems to play pretty good now. I played Aaron's bot set to 5 ply, and the net...
Jason, your robot has improved remarkably. It's cut-and-run style enables it to play competitively against low-ply searches and (my) overconfident children....
Hi Don,<br><br>There's still some issues with training. Too much training appears to degrade performance, so I need to either stop training at the right time ...
Hi,<br><br>I've uploaded the latest and greatest version of my OFK program. It has a more usable user interface and generates a thunk log for viewing the move ...
Your program is improving by leaps and bounds. It's setting up a much more intelligent offense. <br><br>Am I right to assume that it's still looking just 1 ...
Yesterday, Chuck and I were looking back at the Mitchell book on Machine Learning. In chapter 1, he briefly and without much comment suggests that one can tune...
Don,<br><br>It really depends on the structure of f(x). It has been shown that if f(x) is a look up table for all possible states, and you visit each state an ...
Don,<br><br>What Mitchell describes is essentially the learning algorithm used in TD-Gammon. (It is TD(lambda) with lambda=0.) <br><br>TD(lambda) used with a ...
I'm trying to understand better the relationship between the lamda(0) tuning approach and the depth at which the opponent searches. If my opponent, perhaps ...
Hi Don,<br><br>First you need to define your policy for play. I'll assume youy AI is greedy, always taking the highest ranked action. Secondly I'll assume your...
Thanks to AaronA, a puzzle viewer program is in the works, similar to the current format of the on-line log viewer (which you should check out, if you haven't...
It looks like the Mind Sports Olympiad will indeed happen this August in London. I'm planning to be there to run an Octi tournament during the first few days,...
It's hard to believe that Octi first shipped to stores two years ago tomorrow. A lot has happened in just two years, and I feel deeply indebted to a lot of ...
Stephen Tavener has just about finished the code to put Octi up at gamerz.net. We'll be doing some beta testing over the next few days, and I'll let you know...
We've playtested a bit of the new octi program on pbmserv, and so far, so good.<br><br>The issue arises: what version of the game should be the default. My ...
Hi...<br><br>I'm interested in developing some Octi "robotic" players... but I'd rather not reinvent the wheel if I don't have to.<br><br>Has anyone developed...
Hi,<br><br>I am not aware of Octi development that's been packaged in DLL or ActiveX form. The current Octi GUI(s) are written in Java. The UI that people ...
"for a tournament-level game player, well, you just don't want to use basic AI routines..."<br><br>I guess I don't have that those aspirations just yet, being...
I know there are some interfaces out there, and Chuck Chiou and others may be working on something similar to what you have in mind.<br>Perhaps you can team ...
I am working on the GUI and client/server interfaces. The code is in C++ and I'm developing on UNIX platforms... The code is highly portable so I plan to...
Having problems with message search?
Fill out this form to ensure your group is one of the first to be migrated to the new message search system.
