Gnu Shogi Database (GSDB)

Browse hundreds of historical professional and amateur shogi games. Select a collection, pick a game, and replay it on the interactive board.

Kifu is a shogi game record. Every move is written down so the game can be replayed and studied later. Serious students of shogi build up - and work through - large libraries of kifu to understand openings, middlegame plans and yose (end-game) technique.

The Gnu Shogi Database (GSDB) is a collection of Japanese chess games and sample openings. Hundreds of shogi kifu (game records) have been documented. The orignal GSDB comes with an applet that allows viewing and manipulating of the collections. Unfortunately, the web browsing industry has discontinued support for Java applets. That makes browsing the collections trickier, without a little help.

For those techies interested in browsing the original GSDB 2.7.4 source code and license, you can download it here.

Just a note on shogi rankings: The lowest ranking is 15-kyu. The rankings go down to 1-kyu as the player improves. After 1-kyu, rankings begin counting up with 1-dan all the way up to 9-dan. Professional shogi players begin around 5-kyu and go up to 9-dan (the champion professional). Ameteurs of the same rank as professionals are typically less skilled than their professional counterparts.

Meijin – Kifu #22

[Name "Reijer Grimbergen"] [Email ""] [Country "Japan"] [Sente "Tanigawa Koji"] [Gote "Habu Yoshiharu"] [Black_grade "Chall"] [White_grade "Meijin"] [Result "1-0"] [Comment "Comments in the game by Reijer Grimbergen & Shukan Shogi"] [Source "Reijer Grimbergen , translated from Shukan Shogi"] [Event "55th Meijin Title Match"] [Date "19970410"] [Round "1"] [Venue ""] [Proam "Professional"] {~ Furigoma decided that Tanigawa could start the match with black. A couple of~ years ago there was talk that Habu could even mesmerize the pawns thrown to~ decide black and white, but lately he seemed to have lost that ability :-).~ ~} P7g-7f P3c-3d P2g-2f P4c-4d S3i-4h S3a-4b P5g-5f P5c-5d G4i-5h S7a-6b {~ It was expected that Habu would play ranging rook, but Habu decided to play~ Yagura. It was a year and three months ago that Habu and Tanigawa last played~ the Yagura opening against each other.~ ~} S7i-6h G6a-5b P3g-3f G5b-4c G6i-7h G4a-3b K5i-6i K5a-4a P4g-4f P8c-8d P2f-2e S4b-3c S4h-4g P8d-8e S6h-7g B2b-3a N2i-3g P7c-7d P9g-9f N8a-7c P6g-6f {~ The sealed move. 31.S6f also seems possible, but after P6d P5e P6e the silver~ has to retreat to 5g which is not something one would like to do.~ ~} P6c-6d G5h-6g S6b-6c B8h-7i B3a-4b B7i-6h K4a-3a K6i-7i K3a-2b P4f-4e {~ Tanigawa: 'If I allow white to attack first, I have no faith in the position,~ so! I! pushed on'. Now the fight starts.~ ~} P4dx4e N3gx4e S3c-4d S4g-4f P6d-6e {~ The classic counter pawn push. The whole board will now be the battleground. ~} P6fx6e N7cx6e S7g-6f P8e-8f P8gx8f {~ Interestingly enough, the same position already occured in a game between the~ same two players. In November 1988 Habu had black against (at that time) Meijin~ Tanigawa.~ ~} B4bx8f { ? ~ Almost ten years ago, Tanigawa played 52.P*8e, which is much better. After~ Px8f P2d Bx2d P3e Rx8e P*8h white had a good position and went on to win the~ game from there.~ ~} B6hx8f R8bx8f P2e-2d P2cx2d P'2c {~ A severe attack. If Kx2c then P*2e Px2e P*2d Kx2d B*4a and the attack only~ becomes stronger. Taking with the gold is also not very attractive. In the~ analysing room some people already wondered if the game was over. However,~ even at this level the game is never that simple...~ ~} G3bx2c S6fx6e P'6f G6g-7g R8f-8e N'1e {~ This knight drop is always a problem if the pawn is not pushed to 1d. ! However,~ also good seems 63.P*2e Px2e Rx2e P*2d Sx5d (Sx5d N3c+ and Rx8e). ~ ~} S4dx4e B'4a B'3b {! ~ A great fighting move and the only defense that keeps Habu's position together~ (at least for the time being).~ ~} B4ax6c+ N'6g G7gx6g P6fx6g+ G7hx6g R8ex6e S'4a {?? ~ A horrible mistake. Tanigawa: 'I thought I was threatening an easy mate, but~ after having dropped the silver I realised that it wasn't mate at all'.~ ~} R6ex6c {?? ~ Habu is victim to the same illusion. Both players had plenty of time here~ (Tanigawa 25 minutes and Habu more than 90 minutes). Actually Habu took 15~ minutes to play this move. His shallow analysis: Sx3b+ Kx3b B*4a K3a N*2c~ and a quick mate. However, after N*2c K2b G*3b K1b Nx1a+ Kx1a G2b Kx2b Rx2d~ K3a there is no mate. There are actually quite a few variations that seem to~ lead to mate in this position, but none does. For example, Sx3b+ Kx3b Nx2c+~ Kx2c N*1e K3c B*2b K3b and no mate. Conclusion: if Habu would have played~ 74.Rx6g+ instead of !Rx!6c, he would have won the game!~ ~} S4ax3b+ K2bx3b N1ex2c+ K3b-4a P'6d R6cx6d P'6e R6dx6e P'6f R6e-8e B'6c S'5b B6cx7d+ R8e-8g+ {~ Habu: 'G*6i might have been better.. '. However, even if 88.G*6i then K7h~ B*8g K8h is no mate and black wins.'~ ~} G6g-7g S'8h { ~ Nothing helps. For example 90.G*6i Kx6i +Rx8i G*7i B*4g G*5h and black wins.~ ~} R2hx8h G'6i K7ix6i +R8gx8h S'3b K4a-5a +B7dx5b K5ax5b S3bx4c {~ Resigns~ ~ Not an easy mate. Habu took three minutes before resigning to confirm the variations:~ a) 100.Kx4c N*5e Px5e G*4d Kx4d Sx5e and easy mate~ b) 100.K6a S5b= (S5b+ is also good) Kx5b G*5c Kx5c N*6e and mate~ c) 100.K6b N*7d K6c G*7c Kx7c S*6b and mate~ ~ A hard blow to Habu who not only lost but also had his endgame confidence~ shaken. Let's see how he recovers in the second game on April 26th and 27th.~ ~ }