Harry Potter (no, not that Harry Potter!) writes "I hope you will devote a column to the subject of using databases. Do I need one to play correspondence chess? Are they useful? Do they save study time? If a database is worth having, how do I use it? And which one is the best?"
I did a column on data base use in the June 1995 Chess Life. It is a fascinating subject, though, so here is another.
Do you need one to play correspondence chess?
No. Many postal players prefer to think for themselves. Many other players prefer to use previously played successful moves. In general, the stronger your opposition, the more likely you are to want to use a database.
Are databases useful?
Yes. The database works like books on chess openings. What may be missing is the evaluation, giving instead the result of the game. This may be misleading with a limited number of games, but with a large number of examples a conclusion can be drawn about the evaluation of that opening.
Do databases save time?
Probably not. The proper use of databases will increase the time spent on deciding on a move: more choices, more variations, more games, more time.
How do I use a database?
The following game may shed some light on the use of a database for research.
GAME OF THE MONTH
The U.S. vs. Cuba match has had some tough sledding. Most of the games have dragged on for quite some time. I was fortunate in my game. My opponent, a government official, had access to a computer and the game was played by e-mail.
RUY LOPEZ [C85]
W: Alex Dunne (2359)
B: Rafael HernandezU.S.-Cuba match, 1999
1. e4 e5 2. Nf3 Nc6 3. Bb5 a6 4. Ba4 Nf6 5. 0-0 Be7 6. Bxc6
My database of master games played in 2000 shows six games played with this line, White scoring +2 =1 -3. But in 1996-1999, things were better for White, with results of +18 =20 -12.
6. ... dxc6 7. Nc3
The choices in the 1996-1999 master database were -
7. Nc3 scored +2 =3 -0
7. Qe2 scored +4 =3 -4
7. Qe1 scored + 10 =13 -6
7. d3 scored +12 =15 -13
7. Re1 scored +1 =0 -2
Percentagewise, 7. Nc3 stands out. Next, the games with this line are played out. IM Walter Muir, writing before databases, gave the advice: "The best and only way to learn chess is to play over a great many games annotated and played by masters, to get ideas for one's own play. In addition, one will get a feel for the middlegame complications and endgame techniques required for winning."
Here I decided to choose 7. Nc3, a change from my usual 7. Re1. Because of the better percentage? No, just for a change of pace.
7. ... Qd6
There are no games with this in recent events. I will have to turn to larger, older databases. I could combine all my databases and have one 400,000 game database, but I prefer to divide them up for more specialized use. After checking all my databases, I find no example of this move. Are my databases washed up? No, I can still find many similar positions in the Ruy Lopez with 1. e4 e5 2. Nf3 Nc6 3. Bb5 a6 4. Bxc6 dxc6 5. 0-0 Qd6, so I can look at some of those games to get ideas of middlegame play.
At this point I am out of my database use, and it is time to act on my own. Some programs allow the player to search for positions with similar pawn structures and similar distribution of pieces (White: two knights plus bishop vs. Black: two bishops plus knight) but I don't have that capability.
Now that my databases are exhausted, is it all over? Well, there are databases of 1,000,000 or more games. Would they do better? Perhaps, but as always there comes a time when you are on your own. This is such a time, probably for both players.
When choosing a database, there is a give and take between quantity, quality, and annotations. For opening strength, probably the best database would be one of master-level postal players such as Mega-Corr, with 270,000+ games.
8. ... Nd7 9. Be3 0-0 10. Qe2 exd4
The first slide down the slope: White's pawn structure is superior with the 4-to-3 kingside majority.
11. Bxd4 Qg6 12. Rad1 Bc5?
The second descent toward 1-0: Black loses his one hope for the future, the two bishops.
13. Qe3! Bxd4 14. Rxd4 Nb6 15. Rfd1 Bg4 16. Ne5 Qe6 17. Nxg4 Qxg4 18. b3 h6 19. h3 Qe6 20. Qg3 Qe7
With this advance, White announces the endgame doesn't interest him as much as the middlegame.
21. ... Rad8 22. Rg4 Rxd1+ 23. Nxd1 g6 24. Ne3 Kh7 25. c4 Rd8 26. e6!
This shot leaves Black's king without proper cover as 26. ... Qxe6 27. Qxc7 wins.
26. ... Rf8 27. Rh4 fxe6 28. Ng4 h5 29. Ne5 Qg7 30. Nxg6!
White effectively ends Black's resistance as Black's king will find no shelter.
30. ... Qxg6 31. Qxc7+ Rf7 32. Qxb6 Qb1+ 33. Kh2 Qxa2 34. Qd4 Qe2 35. f3 Rf5 36. Re4, Black resigns.
Isay Golyak (Rochester, NY) fashions a slow-motion attack, sniping off pawns all over the board.
CARO-KANN DEFENSE [B12]
W: Isay Golyak (2220)
B: Steven Fairbairn (2048)
1997 Golden Knights
1. e4 c6 2. d4 d5 3. e5 Bf5 4. g4 Bg6 5. h4 h5 6. e6 fxe6 7. Bd3 Bxd3 8. Qxd3 hxg4 9. Qg6+ Kd7 10. Qxg4 Nf6 11. Qe2 Kc8 12. Nf3 Nbd7 13. Qxe6 Kc7 14. Bf4+ Kb6 15. c4 dxc4 16. Nc3 a6 17. d5 Ka7 18. dxc6 bxc6 19. Qxc6 Qc8 20. Be3+ Kb8 21. Qe6 Qb7 22. Ke2 Rh5 23. Rhd1 g6 24. b4 Rf5 25. Nd4 Rh5 26. Bf4+ Kc8 27. Qxc4+, Black resigns.
The master's touch is evident here. Cross simply puts his pieces on their best squares and every exchange brings a winning endgame closer.
FRENCH DEFENSE [C11]
W: Ted Cross (2248)
B: Richard Hanlon (1897)
1999 Golden Knights
1. e4 e6 2. d4 d5 3. Nc3 Nf6 4. e5 Nfd7 5. f4 c5 6. Nf3 Qb6 7. Be3 a6 8. Na4 Qa5+ 9. c3 cxd4 10. b4 Qc7 11. Qxd4 a5 12. b5 b6 13. c4 Nc5 14. Nc3 dxc4 15. Bxc4 Bb7 16. 0-0 g6 17. Ng5 Bg7 18. Nce4 Nxe4 19. Nxe4 Bxe4 20. Qxe4 Ra7 21. Rac1 Qb7 22. Qxb7 Rxb7 23. Be2 Nd7 24. Bf3 Rb8 25. Rc7 Bxe5 26. fxe5 Nxe5 27. Bc6+ Nxc6 28. bxc6 f5 29. Rd1, Black resigns.
Thematic tournaments allow the contestants a chance to experience life on both sides of the street.
KING'S INDIAN DEFENSE [E97]
W: Ernst Melchoir (1908)
B: Robert Irons (1604)
KID Thematic 1991
1. d4 Nf6 2. c4 g6 3. Nc3 Bg7 4. e4 d6 5. Nf3 0-0 6. Be2 e5 7. 0-0 Nc6 8. d5 Ne7 9. Nd2 a5 10. a3 Nd7 11. Rb1 f5 12. b4 Kh8 13. Qc2 Nf6 14. f3 axb4 15. axb4 f4 16. c5 g5 17. cxd6 cxd6 18. Nc4 g4 19. Nb5 g3 20. hxg3 fxg3 21. Nbxd6 Ne8 22. Nxb7 Bxb7 23. Bg5 Nf6 24. Bh4 Ng6 25. Bxg3 Nh5 26. Bf2 Qg5 27. Be3 Ngf4 28. Bxf4 exf4 29. Rfd1 Rg8 30. Bf1 Qh4 31. Nd6 Raf8 32. Nxb7 Ng3 33. Bc4 Bd4+ 34. Rxd4 Qh1+ 35. Kf2 Nxe4+ 36. Ke2 Qxg2+, White resigns.