Author: Vincent Diepeveen
Date: 07:45:14 11/23/99
There is a number of ways to play against a computer.
Let's list the different methods (i'm giving them a cool name here
feel free to modify the name if you refer to it)
- grandmaster preparement
- Mr. exchange
- running pawn formations
- closed positions
- following a single plan
Grandmaster preparement
To start with one that's only possible for strong players:
play mainline of your favourite opening and wait till the
computer does a bookmove that isn't good, or simply play
your preparement and get a much better position and then win
the game.
Obviously the grandmaster way is possible if you don't play openings
that are very dubious. there are however less obvious openings that
are not possible to play too: some people play openings which gives
the opponent a small advantage or an equal position which happens to
be a position where the computer has the right plan, follows this
plan and therefore wins chanceless.
Grandmasters play a lot of different openings. Not usual they have
very professionally prepared at least 2 openings, to have choice
against the different opponents.
As openingsbooks are based upon the few things grandmasters release
from their preparement, and computers not too well in opening,
it's obvious that computer books are always
behind the preparement of grandmasters, as they are dependant upon
the few published findings of them and basically their games.
It's for me very obvious that in future grandmasters will remain
a big problem for computers as they can get a very good
openings position and they already proved when getting their GM title
that they can finish it without blundering pieces.
Mr Exchange
Something derived from the past is the idea to just exchange everything
against the computer preferably into an endgame or a position with
very little pieces where tactics, though still important, are easier to
predict for humans.
Not seldom i follow this concept myselve with changing success. Last
years computers got a lot stronger in endgame, not because they search
deeper in it, but basically because most programmers have improved
the endgame of the different programs. A big risk human is having
here is that the computer sometimes while exchanging finds a shot,
small or big, that directly turns the position after exchanging into
a lost one for the human.
Starting endgame is still a good idea against computer,
the progress of the programs in these endgames is very obvious though.
Running pawn formations
This is a method against what i call 'piece square table programs',
though i use this pst program word quickly when a program doesn't have
much knowledge.
Basically
programs that have very little pawn knowledge in them and therefore misjudge
running majorities, minorities or pawn runs for mating purposes. Also
placing pawns on the right color is included in this section.
The bottom line is that the outcome of the pawn run decides the game.
Sometimes sacrafices need to be done for this, which is very dangerous.
Some programs respond very well on pawn formations. This can be used only
when you know what program you're playing. good examples are blitz games of
mine against programs where i setup opening with 1.e4 2.c4 3.Nc3 4.nf3 5.h3
6.g3 7.d3 8.bg2 9.0-0 10.Kh2 and then slowly i occupy f5 with the knight
hope for programs to respond wrong with h6 (bg5 they fear!) and then crush
them with f4-f5 and g4-g5. Basically the pawnstorm getsunderestimated here.
Another obvious example of a pawn storm is kasparov-deep blue the last game
of the previous match where kasparov wins by a majority queen side, even
ignoring a winning tactical move bxh7 somewhere.
Closed positions
Very bad is the computer in closed positions. Even weak players can get
bigtime won or at least drawn positions using the concept of closed positions.
I saw some messages here about closed centres. Closed centres are not
the same like closed positions. a position with a closed centre but open
c file and half open f file (like french tarrasch variation) is definitely
not closed. Many caro-kann lines and exchange Slav lines are also with
closed center but definitely not a closed position.
However sometimes a position with open file(s) and half open files can
be seen as closed, good examples are kings indian defense E99 and such.
Usual computers have the biggest problem with this, as all normal rules
might be not valid.
the person playing this against a computer however must not have too much
adrenaline or be wise. Usual the idea is to play for the draw as closing
theposition with black is sometimes hard.
For example:Secret (2979) vs. Kingway (2709)
Rated blitz match, initial time: 5 minutes, increment: 3 seconds
Move Secret Kingway
---- ---------------- ----------------
1. e4 (0:00) e6 (0:01)
2. d4 (0:00) d5 (0:00)
3. Nc3 (0:00) Bb4 (0:01)
4. e5 (0:00) c5 (0:01)
5. a3 (0:00) Bxc3+ (0:04)
6. bxc3 (0:00) Qa5 (0:01)
7. Bd2 (0:00) Qa4 (0:01)
This line is said to be very bad for black.
My book (eco) says HUGE advantage for white,
however see how the international master
carefully closes the position and slowly
converts his lost position to a won one.
8. Nf3 (0:00) c4 (0:16)
9. Be2 (0:22) h6 (0:06)
10. Qc1 (0:18) Nc6 (0:05)
11. Qb2 (0:13) Nge7 (0:01)
12. O-O (0:21) b6 (0:01)
13. Bf4 (0:12) Bd7 (0:02)
14. Be3 (0:11) O-O-O (0:04)
15. Nh4 (0:11) Ng6 (0:02)
16. Nxg6 (0:09) fxg6 (0:00)
17. Bg4 (0:07) Rdf8 (0:09)
18. Rab1 (0:14) Kd8 (0:18)
19. f3 (0:11) Ke7 (0:04)
20. Bf2 (0:10) g5 (0:03)
21. Be3 (0:08) Be8 (0:01)
22. Qc1 (0:07) Bg6 (0:01)
23. Bxg5+ (0:06) Kd7 (0:39)
24. Bh4 (0:06) Qxc2 (0:08)
25. Qxc2 (0:00) Bxc2 (0:02)
26. Rb2 (0:04) Bd3 (0:42)
27. Re1 (0:00) g5 (0:10)
28. Bf2 (0:07) h5 (0:17)
29. Bh3 (0:00) Rhg8 (0:09)
30. g4 (0:02) Rxf3 (0:19)
31. Bg2 (0:00) hxg4 (0:47)
32. Bxf3 (0:07) gxf3 (0:01)
33. Be3 (0:06) g4 (0:05)
34. Kf2 (0:02) Rh8 (0:13)
35. Bf4 (0:09) Ne7 (0:05)
36. Rg1 (0:07) Bf5 (0:03)
37. Ke3 (0:08) Ng6 (0:09)
38. Bg5 (0:07) Rh5 (0:12)
39. Bf6 (0:07) Nf8 (0:42)
40. Rg3 (0:07) Nh7 (0:03)
41. h3 (0:07) Nxf6 (0:11)
42. exf6 (0:06) Rxh3 (0:25)
43. Rxh3 (0:00) gxh3 (0:02)
44. Kxf3 (0:02) Ke8 (0:15)
45. Kf4 (0:10) Kf7 (0:01)
46. Ke5 (0:10) Be4 (0:01)
47. Rh2 (0:04) Bg2 (0:01)
48. a4 (0:09) a6 (0:03)
49. Kf4 (0:02) Kxf6 (0:02)
50. Ke3 (0:03) Kf5 (0:01)
51. Kf2 (0:06) Kf4 (0:01)
52. Kg1 (0:06) Ke3 (0:01)
53. Rxg2 (0:07) hxg2 (0:02)
54. Kxg2 (0:03) Kd3 (0:01)
55. Kf3 (0:04) Kxc3 (0:01)
56. Kg4 (0:04) Kxd4 (0:02)
57. Kf3 (0:03) c3 (0:01)
58. Ke2 (0:05) e5 (0:02)
59. Kd1 (0:05) Kd3 (0:01)
60. Ke1 (0:05) e4 (0:02)
61. Kf2 (0:05) d4 (0:01)
62. Kg3 (0:05) c2 (0:01)
63. Kg4 (0:05) c1=Q (0:02)
64. a5 (0:05) b5 (0:02)
65. Kf5 (0:05) e3 (0:01)
66. Kg4 (0:05) e2 (0:02)
67. Kf5 (0:02) e1=Q (0:03)
68. Kg4 (0:00) Qcd1+ (0:04)
69. Kf4 (0:05) Qe4+ (0:04)
70. Kg3 (0:03) Qg1+ (0:11)
71. Kh3 (0:00) Qeg4# (0:03)
{White checkmated} 0-1
So superior strategic play of the international master
killed a dual 550 celeron getting in closed positions huge
search depths, even at blitz.
Simple said: the computer doesn't play with a plan. Plan
should be not to open the position, sometimes that's very
bad! Usual it's a good idea though against humans!
Following a single plan
Now we have had strategies that are objectively very good,
where some games get won by following a (better) plan,
there are also positions where incorrect plans, though consequently
executed, win the game.
Move blik Diepeveen
---- ---------------- ----------------
1. d4 (0:02) c6 (0:00)
2. e4 (0:00) d5 (0:00)
3. Nd2 (0:00) a6 (0:04)
4. Ngf3 (0:20) Bg4 (0:11)
5. h3 (0:11) Bh5 (0:04)
6. c4 (0:21) dxe4 (0:22)
7. Nxe4 (0:23) Bxf3 (0:03)
8. gxf3 (0:12) Nd7 (0:06)
9. Bg2 (0:20) Ngf6 (0:36)
10. Qb3 (0:00) Qb6 (0:03)
11. Qxb6 (0:20) Nxb6 (0:02)
12. b3 (0:08) Nbd7 (0:04)
13. Bf4 (0:19) h6 (0:29)
14. a3 (0:08) Nh5 (0:03)
15. Bc7 (0:10) Rc8 (0:04)
16. Bg3 (0:17) f5 (0:04)
17. Nc3 (0:11) f4 (0:03)
18. Bh2 (0:08) g5 (0:16)
19. O-O-O (0:10) Bg7 (0:02)
20. Rhe1 (0:06) Nhf6 (0:53)
21. Kb2 (0:13) Kf7 (0:10)
22. h4 (0:17) Rcd8 (0:20)
23. hxg5 (0:15) hxg5 (0:04)
24. Bg1 (0:03) Nf8 (0:09)
25. Re5 (0:13) Rh5 (0:10)
26. Re2 (0:12) Ng6 (0:08)
27. Rd3 (0:08) Rd7 (0:28)
28. Kc2 (0:11) Nh7 (0:11)
29. d5 (0:08) cxd5 (0:04)
30. Nxd5 (0:03) Rd6 (0:13)
31. Red2 (0:03) Nhf8 (0:14)
32. Kc1 (0:09) Nd7 (0:02)
33. b4 (0:02) Nde5 (0:02)
34. Rc3 (0:05) Nh4 (0:06)
35. Bh1 (0:13) Rh8 (0:09)
36. c5 (0:17) Rc6 (0:13)
37. Kd1 (0:03) Rd8 (0:06)
38. a4 (0:05) e6 (0:01)
39. Nb6 (0:04) Rxd2+ (0:01)
40. Kxd2 (0:06) Rc7 (0:08)
41. Nc4 (0:04) Nc6 (0:02)
42. Nd6+ (0:04) Ke7 (0:01)
43. Rb3 (0:03) Nd4 (0:03)
44. Rd3 (0:03) Nc6 (0:05)
45. b5 (0:01) Nb4 (0:04)
46. Rb3 (0:03) Nd5 (0:03)
47. bxa6 (0:04) bxa6 (0:01)
48. Ne4 (0:02) Bd4 (0:07)
49. c6 (0:15) Bf6 (0:06)
50. Rb7 (0:06) Kd8 (0:05)
51. Nxf6 (0:04) Nxf6 (0:04)
52. Rb6 (0:01) Ke7 (0:03)
53. a5 (0:05) Kd6 (0:05)
54. Rxa6 (0:01) Nd5 (0:03)
55. Ra8 (0:06) Rxc6 (0:01)
56. Rd8+ (0:01) Ke7 (0:03)
57. Ra8 (0:01) Rc4 (0:06)
58. Rg8 (0:06) Kf6 (0:03)
59. Rf8+ (0:03) Kg7 (0:01)
60. Re8 (0:04) Kf7 (0:01)
61. Rb8 (0:03) Ra4 (0:02)
62. Rb7+ (0:01) Kf6 (0:02)
63. Kc2 (0:03) Rxa5 (0:02)
64. Kb2 (0:04) Rc5 (0:02)
65. Rb8 (0:03) Rc4 (0:03)
66. Rf8+ (0:03) Kg7 (0:02)
67. Rb8 (0:03) Rd4 (0:02)
68. Kc2 (0:03) Ra4 (0:03)
69. Rb7+ (0:04) Kf6 (0:03)
70. Kb2 (0:02) Rb4+ (0:03)
71. Rxb4 (0:05) Nxb4 (0:02)
72. Kc3 (0:01) Nc6 (0:01)
73. Kd3 (0:03) Ke5 (0:01)
74. Ke2 (0:04) Nd4+ (0:01)
75. Kf1 (0:01) Ndxf3 (0:02)
76. Bxf3 (0:03) Nxf3 (0:01)
77. Kg2 (0:02) Nxg1 (0:04)
78. Kxg1 (0:06) g4 (0:00)
79. Kh2 (0:01) Kf5 (0:02)
80. Kg1 (0:04) Kg5 (0:00)
81. Kf1 (0:03) Kh4 (0:01)
82. Kg1 (0:02) f3 (0:03)
{White resigns} 0-1
Move crafty Diepeveen
---- ---------------- ----------------
1. d4 (0:02) d5 (0:01)
2. Nf3 (0:00) c6 (0:01)
3. Bf4 (0:00) Nf6 (0:01)
4. Nbd2 (0:00) Bf5 (0:02)
5. e3 (0:00) h6 (0:01)
6. Bd3 (0:11) e6 (0:04)
7. Bxf5 (0:11) exf5 (0:02)
8. Nh4 (0:00) g6 (0:28)
9. O-O (0:11) Na6 (0:05)
10. Nhf3 (0:11) Bg7 (0:04)
11. c4 (0:05) O-O (0:04)
12. Be5 (0:05) Nc7 (0:03)
13. Qb3 (0:13) Qc8 (0:15)
14. c5 (0:01) Ne6 (0:13)
15. Bd6 (0:09) Re8 (0:04)
16. Qb4 (0:05) Ne4 (0:15)
17. Nxe4 (0:14) fxe4 (0:02)
18. Ne5 (0:06) Bxe5 (0:38)
19. Bxe5 (0:14) Nf8 (0:26)
20. Qc3 (0:08) Nd7 (0:03)
21. Bf4 (0:08) g5 (0:06)
22. Bd6 (0:09) Re6 (0:13)
23. b3 (0:09) Nf8 (0:05)
24. Rac1 (0:07) Ng6 (0:02)
25. Rc2 (0:05) Qd7 (0:16)
26. Re1 (0:07) f5 (0:07)
27. Qd2 (0:07) Qf7 (0:07)
28. g3 (0:07) Kh7 (0:07)
29. Rcc1 (0:07) a6 (0:05)
30. Rc2 (0:06) Rg8 (0:01)
31. Kh1 (0:06) Rf6 (0:02)
32. Qb4 (0:06) Rg7 (0:05)
33. Qc3 (0:07) h5 (0:09)
34. Rg1 (0:06) g4 (0:05)
35. Rcc1 (0:05) Re6 (0:02)
36. Rcd1 (0:03) Rg8 (0:01)
37. Qd2 (0:05) Qd7 (0:02)
38. Qc2 (0:05) Nh8 (0:01)
39. Qe2 (0:05) Nf7 (0:01)
40. Bf4 (0:02) Ng5 (0:05)
41. Rg2 (0:04) Nf3 (0:02)
42. Qf1 (0:01) Kg6 (0:03)
43. Bd6 (0:04) Ree8 (0:04)
44. Qe2 (0:04) Rg7 (0:01)
45. Rf1 (0:04) Rh7 (0:02)
46. Bf4 (0:03) Kf6 (0:01)
47. Rd1 (0:02) Ke6 (0:01)
48. Bd6 (0:02) h4 (0:02)
49. gxh4 (0:02) Rxh4 (0:02)
50. Bg3 (0:03) Rh3 (0:03)
51. Rc1 (0:02) Reh8 (0:02)
52. Qd1 (0:02) R8h6 (0:03)
53. Bf4 (0:03) R6h5 (0:02)
54. Bg3 (0:02) Qh7 (0:01)
55. Qg1 (0:01) Nxg1 (0:08)
56. Kxg1 (0:00) Qd7 (0:02)
57. Be5 (0:02) Rh7 (0:02)
58. Kf1 (0:02) Rf7 (0:01)
59. Ke2 (0:02) Rfh7 (0:01)
60. Kf1 (0:02) Rf7 (0:01)
61. Ke2 (0:02) Qd8 (0:01)
62. Rc2 (0:00) Rf3 (0:02)
63. Kd1 (0:02) f4 (0:01)
64. Bxf4 (0:02) Rf7xf4 (0:02)
65. exf4 (0:02) Rxf4 (0:01)
66. Rd2 (0:02) Qh4 (0:02)
67. Ke2 (0:00) Qh3 (0:03)
68. Kf1 (0:00) Qf3 (0:07)
69. Rg3 (0:03) Qh1+ (0:03)
70. Rg1 (0:01) Qxh2 (0:01)
71. Rg2 (0:00) Qh1+ (0:03)
72. Rg1 (0:00) Qf3 (0:01)
73. Ke1 (0:03) Kd7 (0:03)
74. Re2 (0:03) Qd3 (0:02)
75. Rd2 (0:02) Qb1+ (0:01)
76. Rd1 (0:00) Qxa2 (0:01)
77. Rg2 (0:02) Qxb3 (0:01)
78. Rg3 (0:02) Qb4+ (0:03)
79. Kf1 (0:03) a5 (0:01)
80. Kg1 (0:02) a4 (0:01)
81. Kg2 (0:00) a3 (0:02)
82. Re3 (0:02) Rf3 (0:01)
83. Re2 (0:01) Qb3 (0:03)
84. Rde1 (0:02) g3 (0:00)
85. Rf1 (0:01) gxf2 (0:01)
86. Rexf2 (0:03) Rxf2+ (0:02)
87. Rxf2 (0:02) a2 (0:00)
88. Rf7+ (0:03) Ke6 (0:02)
89. Rf1 (0:05) Qb2+ (0:02)
90. Kg3 (0:05) a1=Q (0:01)
91. Rxa1 (0:00) Qxa1 (0:02)
92. Kf2 (0:00) Qxd4+ (0:01)
93. Kg2 (0:03) Qxc5 (0:00)
94. Kf1 (0:04) b5 (0:01)
95. Kg2 (0:01) b4 (0:01)
96. Kf1 (0:01) Qc2 (0:02)
97. Kg1 (0:00) b3 (0:00)
98. Kf1 (0:00) b2 (0:00)
99. Ke1 (0:00) b1=Q# (0:00)
{White checkmated} 0-1
Obviously the last strategy will lose MOST games, not in
the last place because games are so long if you seem to be
successful, but assuming computer MUST win from me,
it's fun to do, as one has nothing to lose!
Vincent Diepeveen
diep @ xs4all.nl
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.