[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Gnushogi-devel] [Patch 0/1] Switch to getopt cmdline parsing

From: H.G. Muller
Subject: Re: [Gnushogi-devel] [Patch 0/1] Switch to getopt cmdline parsing
Date: Thu, 8 Feb 2018 10:18:13 +0100
User-agent: Mozilla/5.0 (Windows NT 6.3; WOW64; rv:52.0) Gecko/20100101 Thunderbird/52.5.2

I have been involved in the development of GNU Shogi a few years ago, in particular for making it XBoard compatible. (Originally it was using some mutilated dialect of XBoard protocol to communicate with xShogi, a now totally obsolete fork of XBoard.)

I was never motivated to do something about the AI, however. The problem is that GNU Shogi's AI must be excessively buggy, because it truly plays like crap. Even general-purpose Chess-variant engines like Sjaak II, when configured with Shogi rules and not completely idiotic piece values, will totally crush it, because virtually
ever move it plays is a blunder of astronomic proportions.

I have already written two Shogi engines of my own, however (one of them even able to play several Shog variants, including the 11x11 Wa Shogi) which are again far stronger than general-purpose engines, although not early as strong as Shogi
engines based on data mining, such as Bonanza (which is also open source).
And there seems to be no reason at all to patch an engine that is ~3000 Elo weaker then state of the art, and ~1500 Elo weaker than what a trivial (say 200-line) code
written from scratch could do.

Op 2/8/2018 om 2:47 AM schreef Chen Chen:
Hi Justin,

The patch seems quite good. The hand-written argument parsing is indeed

And yes, the development of this repo has been stalled for years.


reply via email to

[Prev in Thread] Current Thread [Next in Thread]