From: Larry Kaufman comcast net> Date: 29 jul 2004 Subject: Todai Shogi This is a multi-part message in MIME format. ------=_NextPart_000_0017_01C475A5.68D798F0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Your free subscription is supported by today's sponsor: ------------------------------------------------------------------- Get a Great Credit Card for You Today=20 You can find a credit card to fit your credit needs.=20=20 All types of credit cards -- 0% APRs, Rewards, & Bad Credit. http://click.topica.com/= caacpgda2i6YsbnuqMaa/411Web ------------------------------------------------------------------- I recently received a copy of "Todai Shogi" and thought I'd review it = and also ask a question or two about it for anyone who might know. It is sa= id to be either the strongest or among the strongest shogi programs (any fe= edback on this point would be welcome).=20 Anyway, I found it relatively easy to use with my very limited knowled= ge of Japanese, once a Japanese friend set it up for me. So far as I know, = it is the first program on the market with extensive knowledge of handicap = joseki, which makes it especially interesting for me. In very fast play (i.e. 10 seconds per move), it is much too strong fo= r me in even games (i've heard that the top programs are competitive with s= ome pros at this speed). However, in games played at a serious time limit (= say 30' plus 30" per move or at the NHK time limit), I believe I've won mor= e than I've lost, and some of my losses were due to moving too quickly in w= inning positions (which I would not do in actual tournament play). But what= is remarkable is that in the early stages of the game, the program seems r= ather weak, and I often obtain a big edge, but in the endgame it is absolut= ely amazing! In many games (both against me and when set to play itself), I= have seen it find and rattle off 20-25 move Tsume sequences in just a few = seconds, often involving the sacrifice of up to half a dozen pieces!! Surel= y its Tsume ability is far beyond any human being. So while its late openin= g and early middlegame play is probably only around amateur 2 Dan level, it= s endgame is about Pro 15 Dan level (if there were such a thing!). The disc= repancy is so huge as to be incredible. So overall I'd rate it somewhere be= tween amateur 4 and 5 Dan, which also seems to fit with its results in both= even and handicap games against other players in our club. But I can't but= help wonder how long it will be before the middlegame play can be improved= to around my level, at which point its incredible endgame skill should mak= e it competitve with some pros I would imagine.=20 Actually I may be understating its current strength, because no matter= what time limit I select, the program plays far faster than necessary for = that time limit. I'm wondering if anyone else can explain this phenomenon o= r can tell me if I might possibly be doing something wrong. It seems as if = the program is unaware of the speed of the computer on which it runs (2.8 G= HZ). I have the feeling that if the program slowed down to use its time ful= ly, it might already be too strong for me, but that's just speculation. It was also interesting for me to observe the program playing itself i= n quick games at various handicaps. At bishop, rook, and rook & lance handi= caps the receiving side won about 85% of the games, roughly as might be the= case with amateur 4 Dans taking turns giving each other handicaps, but at = lance the side without the lance actually won by 6-4! This fits with the wi= despread opinion that lance handicap is not so suitable for amateurs; they = don't know how to exploit it. Although in the case of the computer, it alwa= ys played one of a dozen or so different standard joseki lines, so presumab= ly the side with the lance was always better after the opening. Anyway it a= gain reminded me of the need in shogi for a handicap between lance and bish= op. I've also tried playing some handicap games against the program, both g= iving and receiving, but the receiving side almost always wins. Other feedback on this and other programs would be interesting. Larry Kaufman, amateur 5 Dan Your free subscription is supported by today's sponsor: ------------------------------------------------------------------- Save up to 67% on Omaha Steaks + Get 6 FREE Burgers and a=20 FREE Cutlery Set + Cutting Board! http://click.topica.com/= caacpgja2i6YsbnuqMaf/OmahaSteaks ------------------------------------------------------------------- --^---------------------------------------------------------------- This email was sent to: = shogi-l shogi net EASY UNSUBSCRIBE click here: http://topica.com/u/?a2i6Ys.= bnuqMa.= c2hvZ2kt Or send an email to: shogi-unsubscribe topica com For Topica's complete suite of email marketing solutions visit: http://www.topica.com/?p=3DTEXFOOTER --^---------------------------------------------------------------- ------=_NextPart_000_0017_01C475A5.68D798F0 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable =
Your free subscription is supported by today's sponsor:
-------------------------------------------------------------------
Get a Great Credit Card for You Today=20
You can find a credit card to fit your credit needs.=20=20
All types of credit cards -- 0% APRs, Rewards, & Bad Credit.
http://click.topica.com/=
caacpgda2i6YsbnuqMab/411Web
-------------------------------------------------------------------
     I recently receiv= ed a copy=20 of "Todai Shogi" and thought I'd review it and also ask a question or two a= bout=20 it for anyone who might know. It is said to be either the strongest or amon= g the=20 strongest shogi programs (any feedback on this point would be welcome).=20
     Anyway, I found i= t=20 relatively easy to use with my very limited knowledge of Japanese, once a= =20 Japanese friend set it up for me. So far as I know, it is the first program= on=20 the market with extensive knowledge of handicap joseki, which makes it=20 especially interesting for me.
     In very fast play= (i.e. 10=20 seconds per move), it is much too strong for me in even games (i've heard t= hat=20 the top programs are competitive with some pros at this speed). However, in= =20 games played at a serious time limit (say 30' plus 30" per move or at the N= HK=20 time limit), I believe I've won more than I've lost, and some of my losses = were=20 due to moving too quickly in winning positions (which I would not do in act= ual=20 tournament play). But what is remarkable is that in the early stages of the= =20 game, the program seems rather weak, and I often obtain a big edge, but in = the=20 endgame it is absolutely amazing! In many games (both against me and when s= et to=20 play itself), I have seen it find and rattle off 20-25 move Tsume sequences= in=20 just a few seconds, often involving the sacrifice of up to half a dozen pie= ces!!=20 Surely its Tsume ability is far beyond any human being. So while its late= =20 opening and early middlegame play is probably only around amateur 2 Dan lev= el,=20 its endgame is about Pro 15 Dan level (if there were such a thing!). T= he=20 discrepancy is so huge as to be incredible. So overall I'd rate it somewher= e=20 between amateur 4 and 5 Dan, which also seems to fit with its results in bo= th=20 even and handicap games against other players in our club. But I can't but = help=20 wonder how long it will be before the middlegame play can be improved to ar= ound=20 my level, at which point its incredible endgame skill should make it compet= itve=20 with some pros I would imagine.
     Actually I may be= =20 understating its current strength, because no matter what time limit I sele= ct,=20 the program plays far faster than necessary for that time limit. I'm wonder= ing=20 if anyone else can explain this phenomenon or can tell me if I might possib= ly be=20 doing something wrong. It seems as if the program is unaware of the speed o= f the=20 computer on which it runs (2.8 GHZ). I have the feeling that if the program= =20 slowed down to use its time fully, it might already be too strong for me, b= ut=20 that's just speculation.
     It was also inter= esting=20 for me to observe the program playing itself in quick games at various=20 handicaps. At bishop, rook, and rook & lance handicaps the receiving si= de=20 won about 85% of the games, roughly as might be the case with amateur 4 Dan= s=20 taking turns giving each other handicaps, but at lance the side without the= =20 lance actually won by 6-4! This fits with the widespread opinion that lance= =20 handicap is not so suitable for amateurs; they don't know how to exploit it= .=20 Although in the case of the computer, it always played one of a dozen or so= =20 different standard joseki lines, so presumably the side with the lance was= =20 always better after the opening. Anyway it again reminded me of the need in= =20 shogi for a handicap between lance and bishop. I've also tried playing some= =20 handicap games against the program, both giving and receiving, but the rece= iving=20 side almost always wins.
     Other feedback on= this and=20 other programs would be interesting.
 
     Larry Kaufman, am= ateur 5=20 Dan
=
Your free subscription is supported by today's sponsor:
-------------------------------------------------------------------
Save up to 67% on Omaha Steaks + Get 6 FREE Burgers and a=20
FREE Cutlery Set + Cutting Board!
http://click.topica.com/=
caacpgja2i6YsbnuqMag/OmahaSteaks
-------------------------------------------------------------------
--^^-------------------------------------------------=
--------------
This email was sent to: =
shogi-l  shogi  net

EASY UNSUBSCRIBE click here: http://topica.com/u/?a2i6Ys.=
bnuqMa.=
c2hvZ2kt
Or send an email to: shogi-unsubscribe  topica  com

For Topica's complete suite of email marketing solutions visit:
http://www.topica.com/?p=
=3DTEXFOOTER
--^^---------------------------------------------------------------
------=_NextPart_000_0017_01C475A5.68D798F0--