Tactics cannot be very important for chess

[hgm]

"Tactics is what you do when there is something to do. Strategy is what you do when there is nothing to do."

BTW, the NN engines do not prove that tactics is not important. They prove they can avoid it against conventional engines. But the conventional engineswere never tuned to seek tactical complexity, and it is very questionable whether the NN engines could still avoid the tactics when a tactically strong opponent seeks it.

[Dann Corbit]

That is an interesting point
One of the complaints of the test I performed is that more than half of the positions are composed tactical problems

But if a human can compose a tactical problem, then a computer can do that too. Hence incredibly complicated tactical problems could be built by computers. Perhaps some of them are currently out of reach for a computer to build, but within ten years a computer could build tactical problems beyond human comprehension.

On the other hand, I think it is probably possible to train a net for tactical situations, just by feeding these positions to the computer over and over on both offense and defense. Eventually, the computer would learn every possible theme.

[Dann Corbit]

Test progress with the x256 net:

Code: Select all

Analysis of G:\chess\epd\ACT2.epd   
Analyzing engine: Lc0
10/29/2019 9:25:14 AM Level: 720 Seconds
   Source                    : List of EPD/PGN files
      G:\chess\epd\ACT2.epd
      G:\chess\epd\ACT1-B.epd
   Colours                   : White moves=True, Black moves=True
   Direction                 : forward
   Games (PGN)               : From 1 to 999999
   Moves                     : From 1 to 250
   Positions (EPD)           : From 1 to 999999 
   Use Engine(s)             : List
      Lc0
   Use seconds per move      : 720 Seconds per move
   Analysis Lines            : Minimum search depth=2
   Activate abort analysis   : False
  1) Nd5-b4             a3-a4    
  2) Ba8-h1             Ba8-h1  * 10 Seconds
  3) e6xd7              Bf1-a6   
  4) Rg1-h1             Rg1-h1  * 7 Seconds
  5) Nb7-d8             Nb7-d8  * 2 Seconds
  6) Qc8-c3             Be6xd5   
  7) Rf5-f8             g6xh7    
  8) Bb8xd6             Bb8xd6  * 166 Seconds
  9) b2-b3              Nd5-f6   
 10) b6-b7              b6-b7   * 4 Seconds
 11) Be2-c4             Be2-h5   
 12) Bc3-d2             Bc3-d2  * 4 Seconds
 13) Kg7-h8             Ba7-d4   
 14) Ke6-f5             Ke6-f5  * 0 Seconds
 15) b4-b5              b4-b5   * 1 Second
 16) Rc2-c5             Qh8-h7   
 17) Kb3-c3             Nd8-c6   
 18) Bf6-g7             Kh3-h4   
 19) Qc1-a3             Qc1xc2   
 20) Ne5-c4             Ne5-f7   
 21) Rh4-a4             Rh4-a4  * 0 Seconds
 22) Nb3-c1             Nf1-e3   
 23) Be5-b8             Be5-d4   
 24) .. Qd1-e1          Qd1-e1  * 6 Seconds
 25) Rb1-a1             Rb1-d1   
 26) Ne1-g2             Ne1xd3   
 27) Qa1-f6             Nc1-b3   
 28) f3-f4              Ng7-e6   
 29) Nh4-g6             Nh4-g6  * 65 Seconds
 30) Nd3-c5             Nd3-c5  * 46 Seconds
 31) Bb4xa5             Bb4-e1   
 32) d2-d3              g7-g8B   
 33) Rc1-c7             Rc1-c7  * 25 Seconds
 34) Bb1-a2             Nf5-d4   
 35) Be3xd2             Kh4-g5   
 36) g7xh8N             g7xh8N  * 252 Seconds
 37) Nd8-e6             Kd6-d5   
 38) Rg7-h7             Rg7-g3   
 39) Bb7-c6, c3-c4      Bb7-c6  * 2 Seconds
 40) f5xg6              f5xg6   * 16 Seconds
 41) e7-e8Q, e7-e8R     e7-e8Q  * 30 Seconds
 42) Qb6-d4             Qb6-e3   
 43) Ng8-h6             Ng8-h6  * 58 Seconds
 44) a5xb6              a5xb6   * 0 Seconds
 45) Re3-c3             Nh4-f5   
 46) Ne7-d5             Re8-f8   
 47) Rf6-h6             Rf6-f7   
 48) Qa2-f7             Qa2-e6   
 49) Be4-d5             Qe3xg5   
 50) Bg1-a7             Bg1-a7  * 2 Seconds
 51) Bg6-f5             Bg6-f5  * 494 Seconds
 52) Rf7-h7             Rf7-h7  * 5 Seconds
 53) Bd4xg7             Bd4xg7  * 6 Seconds
 54) Bb4-d6             Bb4-c5   
 55) Ne6-c7             Rh7-h8   
 56) Rh2-h1             h3-h4    
 57) Re3-e1             Re3-e5   
 58) f4-f5              Qa5-c5   
 59) Bg7xe5             Qh1-c6   
 60) Bh7-b1             a4-a5    
 61) Bg4-f5             Bg4-e2   
 62) Rh3-h2             Ba8xb7   
 63) Rh3xh6             Rh3xh6  * 1 Second
 64) Rg2-a2             Rg2-a2  * 31 Seconds
 65) Nh8-f7             Nh8-f7  * 21 Seconds
 66) .. h6-h5           Ne3-c4   
 67) c5-c6              c5-c6   * 0 Seconds
 68) Ba7-b6             Nc7xa8   
 69) Rc8-a8             Rc8-a8  * 1 Second
 70) Nc8-e7             Nc8-e7  * 80 Seconds
 71) Ka2-a3             Ka2-a3  * 0 Seconds
 72) a2-a3              a2-a3   * 148 Seconds
 73) Nd6-b5             Nb6xc8   
 74) Kd5-e6             Qd4-f2   
 75) Ra6-a7             Ra6-a7  * 685 Seconds
 76) Ne2-c1             b6-b7    
 77) Re8-a8             b3-b4    
 78) Bd2-e3             b3-b4    
 79) Bd4-h8             Bd4-g7   
 80) Ke1-d2             Qe6-c6   
 81) a7-a8N             a7-a8Q   
 82) Be3-d4             Be3-d4  * 104 Seconds
 83) Nb7-d6             Nb7-d6  * 12 Seconds
 84) Ne5-g4             Ne5xf7   
 85) Ne4-d6             Ne8-d6   
 86) Qg6-c6             Qg6-c6  * 33 Seconds
 87) Re1-d1             Bg8xd5   
 88) Bh4-f6             Bh4-g3   
 89) b2-b3              b2-b3   * 52 Seconds
 90) Nc7-a8             Nc7-a8  * 0 Seconds
 91) Kf8-g7             Kf8-g7  * 63 Seconds
 92) Nf2-g4             Nf2-g4  * 73 Seconds
 93) Ne7-g6             Kc8-b7   
 94) Kf4-e4             Kf4-e4  * 490 Seconds
 95) c2-c3              Nh5-f6   
 96) Ng3-e4             Ng3-e4  * 63 Seconds
 97) a3-a4              a3-a4   * 0 Seconds
 98) Bg5-f6             Bg5-f6  * 0 Seconds
 99) Qf2-f7             Ne4-g5   
100) Bg4-e6             Bg4-e6  * 575 Seconds
44 of 100 matching moves
10/30/2019 5:26:45 AM, Total time: 8:01:31 PM Rated time: 12:12:33 = 43953 Seconds

--------------------------------------------------------------------------------

Analysis of G:\chess\epd\ACT1-B.epd   
Analyzing engine: Lc0
10/30/2019 5:27:01 AM Level: 720 Seconds
   Source                    : List of EPD/PGN files
      G:\chess\epd\ACT2.epd
      G:\chess\epd\ACT1-B.epd
   Colours                   : White moves=True, Black moves=True
   Direction                 : forward
   Games (PGN)               : From 1 to 999999
   Moves                     : From 1 to 250
   Positions (EPD)           : From 1 to 999999 
   Use Engine(s)             : List
      Lc0
   Use seconds per move      : 720 Seconds per move
   Analysis Lines            : Minimum search depth=2
   Activate abort analysis   : False
  1) Nc1-b3             Nc1-b3  * 10 Seconds
  2) Ra3-a8             Bb8xc7   
  3) Ne4-c5             Ne4-g5   
  4) Kd3-c4             Kd3-e3   
  5) Nd2-f3             Nd2-f3  * 56 Seconds
  6) Kd3-e2             Kd3-e2  * 0 Seconds
  7) a5-a6              Ra8xb8   
  8) Nd5-f4             Nd5-e7   
  9) Rb7-a7             Rb7-a7  * 0 Seconds
 10) b5-b6              b5xa6    
 11) g7xf8N             g7xf8N  * 0 Seconds
 12) Ra4-a8             Ra4-a8  * 134 Seconds
 13) b5xc6              b5xc6   * 1 Second
 14) h5-h6              h5-h6   * 5 Seconds
 15) b5-b6              b5-b6   * 15 Seconds
 16) Bg1-d4             Qg5xf6   

[Dann Corbit]

Another important point, I guess that a few LC0 answers will also be new solutions. I did not analyze for that yet

[towforce]

"Tactics is what you do when there is something to do. Strategy is what you do when there is nothing to do."

I like that: I've added it to my archive!

BTW, the NN engines do not prove that tactics is not important. They prove they can avoid it against conventional engines. But the conventional engineswere never tuned to seek tactical complexity, and it is very questionable whether the NN engines could still avoid the tactics when a tactically strong opponent seeks it.

If one has built a fast-searching engine, one jolly well SHOULD tune it to seek tactical complexity!

[Dann Corbit]

If one has built a fast-searching engine, one jolly well SHOULD tune it to seek tactical complexity!

I don't think so. Or, at least, that experiment has been tried. Tuning a chess engine to solve tactical test sets definitely lowers the strength in game play. At least, that is what has happened when it was tried before.

It is also interesting that Komodo is not a tactical monster, yet it plays about as well as Houdini which is a tactical monster.

But your point is well taken, and it seems that logically it should work. After all, that is about how we play. After we develop, get space, and achieve some strategic aims, then we look for mating plans and tactical victories. Why shouldn't chess engines do the same thing.

I guess that maybe tactical tuning makes engines forget Sierawan's warning: "Don't search for tactical shots before you have completed development."

[Ovyron]

Aha, so maybe the problem has been that people tune their engines to be great at tactical tests, and the problem is that they use the tuning the whole game, while what would be better is to start without it and turning it on at an appropriate time. Maybe even gradually.

[hgm]

"Tactics is what you do when there is something to do. Strategy is what you do when there is nothing to do."

I like that: I've added it to my archive!

It is also one of my favorites. It is a quote from GM Tartakower.

[jhellis3]

I would say NNs have the *most* tactical potential due to superior move ordering & eval. The issues present in the most prevalent networks are more the result of search, training method(s), and input parameters (lack thereof) than any inherent deficit in NNs themselves. Just my 2cp anyway.....

[duncan]

"Tactics is what you do when there is something to do. Strategy is what you do when there is nothing to do."

BTW, the NN engines do not prove that tactics is not important. They prove they can avoid it against conventional engines. But the conventional engineswere never tuned to seek tactical complexity, and it is very questionable whether the NN engines could still avoid the tactics when a tactically strong opponent seeks it.

Not heard of Tartakower
https://en.wikipedia.org/wiki/Savielly_Tartakower


Tartakower has made some comments about chess, some of which are cynical:

"It's always better to sacrifice your opponent's men."
"An isolated pawn spreads gloom all over the chessboard."
"The blunders are all there on the board, waiting to be made."
"The winner of the game is the player who makes the next-to-last mistake."
"The move is there, but you must see it." (Horowitz 1971:137)
"No game was ever won by resigning."
"I never defeated a healthy opponent." (This quotation refers to players who blame an illness, sometimes imaginary, for their loss.)
"Tactics is what you do when there is something to do; strategy is what you do when there is nothing to do."
"Moral victories do not count."
"Chess is a fairy tale of 1001 blunders."
"The great master places a knight on e5; checkmate follows by itself."
"A master can sometimes play badly, a fan never!"
"A match demonstrates less than a tournament. But a tournament demonstrates nothing at all."
"Chess is a struggle against one's own errors."
"Every chessplayer should have a hobby."
"A game of chess has three phases: the opening, where you hope you stand better; the middlegame, where you think you stand better; and the ending, where you know you stand to lose."
"As long as an opening is reputed to be weak it can be played."
"Stalemate is the tragicomedy of chess."
"Erro ergo sum."