----- Forwarded message from Andrew Sayman <[EMAIL PROTECTED]> ----- Date: Thu, 21 Dec 2006 02:34:36 -0500 From: Andrew Sayman <[EMAIL PROTECTED]> To: Chris Boyle <[EMAIL PROTECTED]> CC: crack-attack-devel <crack-attack-devel@nongnu.org> Subject: Re: Crack-Attack "not easy enough"? OpenPGP: id=E3E55E0F
Chris Boyle wrote: > Any recommendations for this user? :-) > > http://bugs.debian.org/403176 > I have some suggestions, but they require modifying the code as none of this stuff is configurable otherwise. Right now the AI are actually differentiated by how fast they respond to player actions, and what height the invisible AI stack must be at for them to lose. You can modify both of these settings pretty easily in the code and recompile. src/ComputerPlayerAI.cxx:154 is where EasyAI::baseSteps() is defined. The multiplier 15 is a general timing variable that determines how long the AI waits to respond. This is measured gamesteps, which isn't useful to explain. Raising the number makes the AI wait longer to send a return volley of blocks. src/ComputerPlayerAI.cxx:161 is where EasyAI::lossHeight() is defined. This is the height you need to get the invisible stack to before the AI does a custom shatter. Lowering this number means you have to send smaller combos to win. Considering how parametric this is right now, it probably should have command line options, but we still officially intended this as an interim AI. Real Life(tm) has a way of rearing it's ugly head. Andrew Sayman ----- End forwarded message ----- -- Chris Boyle http://cmb.is-a-geek.org/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
