[computer-go] Numbering of MC-players
Hi, I have been playing with Lukas Lew's go library, and created my own version of a pure MC player. I see there are a few on cgos, with a number appended to the name, like Mogo_xxx-3k. I take that to be the number of playouts the program plays. But my program plays (say) 1000 playouts from each possible move on the board. Should I call it xxx_1k, or xxx_81k, when playing on a 9x9 board? - Heikki P.S. The program is called Halgo-1. I will need to test a bit more before I announce any results. But I will do so. -- Heikki Levanto "In Murphy We Turst" heikki (at) lsd (dot) dk ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] Numbering of MC-players
On Mon, 2007-02-19 at 11:16 +0100, Heikki Levanto wrote: > Hi, > > I have been playing with Lukas Lew's go library, and created my own > version of a pure MC player. I see there are a few on cgos, with a > number appended to the name, like Mogo_xxx-3k. I take that to be the > number of playouts the program plays. But my program plays (say) 1000 > playouts from each possible move on the board. Should I call it xxx_1k, > or xxx_81k, when playing on a 9x9 board? If a program says, "xyz-10k", it means it plays exactly 10K play-outs and they are divided between all the moves. - Don > - Heikki > > P.S. The program is called Halgo-1. I will need to test a bit more > before I announce any results. But I will do so. > > > ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] Numbering of MC-players
On Mon, Feb 19, 2007 at 09:12:39AM -0500, Don Dailey wrote: > If a program says, "xyz-10k", it means it plays exactly 10K play-outs > and they are divided between all the moves. Thanks. I'll change my program to behave that way, and give it a number. -H -- Heikki Levanto "In Murphy We Turst" heikki (at) lsd (dot) dk ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
[computer-go] CGOS
I made a few minor fixes/changes to CGOS, nothing major. 1. There is a condition where a program can disconnect but CGOS doesn't always get this right and you can never log back in until the server is restarted.I cannot easily test this, because I have been unable to actually reproduce the bug, but I now trap that error and try to gracefully recover from it. I believe it has been fixed. We will see 2. I silently add 1/2 second to the time control for each move. This is to help programs on slow internet connections and other timming ambiguities. The amount of time your program takes for a given move is never allowed to be reckoned as negative due to this change - so the least amount of time a move can take, even with the 1/2 bonus, is 0.0 seconds. This is not designed to deal with network lag in a "fair" way, if your program is running on a fast reliable network it will still have an advantage, however this should improve the situation where a program on a slow network continues to lose time on the clock even when it's playing instantly. The idea is to try to guarantee that your programs internal clock is a lower bound on the time actually used from CGOS point of view. I would be interested in any reports on time usage vs what CGOS reports. I know that some programs log the discrepancy. This may also benefit programs with aggressive time control algorithms. 3. The SGF files are formatted slightly different. 3.1 The rules are "Chinese" instead of "TT--" to make it more compatble with game readers. 3.2 Instead of placing 1 move per line in the SGF file, I place several moves per line.This makes it appear more compact if you are looking at the SGF file manually (and it eliminates several line-feeds, so it's actually slightly more compact in reality.) - Don ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
[computer-go] a superfast random number generator
I love these comix: http://xkcd.com/c221.html Terry McIntyre UNIX for hire software development / systems administration / security [EMAIL PROTECTED] Get your own web address. Have a HUGE year through Yahoo! Small Business. http://smallbusiness.yahoo.com/domains/?p=BESTDEAL___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
[computer-go] Re: CGOS
On Mon, 2007-02-19 at 11:25 -0500, Don Dailey wrote: > > The idea is to try to guarantee that your programs internal > clock is a lower bound on the time actually used from CGOS > point of view. Actually, I said that backwards. The idea is that if your program thinks it used 3.64 seconds, the server believes it used NO MORE than 3.64 seconds. - Don ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
[computer-go] Big board
Here is a completed game of Go between two random players... on a very large board. For ascetics, the eyes have been filled after both players passed. http://fantius.com/RandomGo1600x1200.png Sorry, no SGF available :) ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
[computer-go] More than one Surrounding Block
I have a question for the group: How is a miai connection strategy created for a color enclosed region when their are multiple enclosing blocks involved with one or more interior defender blocks? Martin Muller's paper "Recognizing Secure Territories in Computer Go By Using Static Rules and Search" doesn't go in sufficient detail for my understanding. At first it seems obvious, especially given the example; however, given a different example where there are more than one possible miai strategy, does it matter which one is used before generating another miai strategy for protecting interior defender blocks? I hope the answer is no. But if so, is there an optimal way to generate the miai connection strategy? I'd appreciate any references to other papers on the subject. - Phil___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
