Chess Computers are they IM? or GM? or Super GM?

[Dirt]

Here an example for the middle game:
[d]r1b1N2k/1pBn2p1/p3Q2p/5n2/8/2q5/2PRB1PP/7K w KQkq -

Not the best example. Toga II 1.4 beta5c (the only engine I checked) finds Qxf5 in just over two minutes on my old and slow system. It should be well under one minute on modern hardware.

There are plenty of position where even the strongest programs have little chance of finding the best move, and a few of those are easily seen by GMs, or even much weaker players. Computers have different strength and weaknesses, but on average they are very strong.

By the way, what's the correct move for the second position? Toga got stuck on Kd6 in the first few seconds, but I doubt you would post the position if it were that easy.

[smirobth]

Hi,
Often they don't understand the middle game.

Here an example for the middle game:
[d]r1b1N2k/1pBn2p1/p3Q2p/5n2/8/2q5/2PRB1PP/7K w KQkq -

Some programs may have trouble with this position, but many do not. Hiarcs finds Qxf5, with a winning score for White, in just a couple seconds. I doubt many humans could "understand" it this well.

[Dann Corbit]

If a Chess Computer had the following:

1) one of the top five SSDF computer chess engines;

2) single Intel Xeon Quad core; and

3) eight gigs of RAM.

At Long Time Controls, would it be an IM or GM or Super GM?

At Long Time Controls, what would be the ELO range?

Thank you and comments are welcome

I guess that the programs will sometimes play like all of the above.
I guess that only very rarely will you see an outright blunder, like putting it's own piece in a closed container, but you will see brilliancies like a Morphy or Casablanca often as well.

Overall, I guess that the strongest chess engines on top hardware are about the same ability as humans. We have the recent Nickel match as a reference, but not enough games for any conclusion, of course.

[Leto]

Here an example for the middle game:
[d]r1b1N2k/1pBn2p1/p3Q2p/5n2/8/2q5/2PRB1PP/7K w KQkq -

Zanzibar finds Qxf5 winning in seven seconds in my quad.

Analysis by Zappa Mexico II x64 4CPU:

1.Bf4 Nc5 2.Qe5 Qxe5 3.Bxe5 Be6 4.Nc7 Ne4 5.Bd3 Nxd2 6.Nxe6
=/+ (-0.27) Depth: 10/33 00:00:00 601kN
1.Bf4 Nc5 2.Qe5 Qxe5 3.Bxe5 Be6 4.Nc7 Ne4 5.Bd3 Nxd2 6.Nxe6
=/+ (-0.27) Depth: 10/33 00:00:00 811kN
1.Bf4 Nc5 2.Qe5 Qxe5 3.Bxe5 Be6 4.Nc7 Rf8 5.Kg1 Bf7
=/+ (-0.39) Depth: 11/34 00:00:00 1153kN
1.Qd5 Ne3 2.Qd3 Qxd3 3.Bxd3 Nc5 4.Be5 Kg8 5.Nxg7 Nxd3 6.cxd3
=/+ (-0.29) Depth: 11/34 00:00:01 2623kN
1.Qd5 Ne3 2.Qd3 Qxd3 3.Bxd3 Nc5 4.Be5 Kg8 5.Nxg7 Nxd3 6.cxd3
=/+ (-0.29) Depth: 11/34 00:00:01 2847kN
1.Qd5 Qe3 2.Bf1 Qe1 3.Rd1 Qxe8 4.Qxf5 Nf6 5.Qa5 Be6
= (-0.10) Depth: 12/43 00:00:03 7340kN
1.Qd5 Qe3 2.Bf1 Qe1 3.Rd1 Qxe8 4.Qxf5 Nf6 5.Qa5 Be6
= (-0.10) Depth: 12/43 00:00:04 8199kN
1.Qd5 Qe3 2.Bf1 Qe1 3.Rd1 Qxe8 4.Qxf5 Nf6 5.Qf4 Be6 6.Bd3 b5 7.Re1
= (-0.17) Depth: 13/43 00:00:05 10864kN
1.Qxf5 Qxd2 2.Qf7 Qe1+ 3.Bf1 Qa1 4.Bd6 h5 5.c3 Qxc3 6.h3 Qa1 7.Kh2 Nc5 8.Nc7 Be6 9.Qxh5+ Kg8 10.Qxc5
+- (1.89) Depth: 13/45 00:00:07 15696kN
1.Qxf5 Qxd2 2.Qf7 Qe1+ 3.Bf1 Qa1 4.Bd6 h5 5.c3 Qxc3 6.h3 Qa1 7.Kh2 Nc5 8.Nc7 Be6 9.Qxh5+ Kg8 10.Qxc5
+- (1.89) Depth: 13/45 00:00:08 16694kN
1.Qxf5 Qxd2 2.Qf7 Qe1+ 3.Bf1 Qa1 4.Bd6 h5 5.h3 b5 6.c3 Qxc3 7.Qxh5+ Kg8 8.Qd5+ Kh8 9.Qxa8 Nb6 10.Qe4 Qa1 11.Qf4 Kg8
+- (4.00) Depth: 14/47 00:00:14 29332kN
1.Qxf5 Qxd2 2.Qf7 Qe1+ 3.Bf1 Qa1 4.Bd6 h5 5.h3 b5 6.c3 Qxc3 7.Qxh5+ Kg8 8.Qd5+ Kh8 9.Qxa8 Nb6 10.Qe4 Qa1 11.Qf4 Kg8
+- (4.00) Depth: 14/47 00:00:15 30830kN
1.Qxf5 Qxd2 2.Qf7 Qe1+ 3.Bf1 Qa1 4.Bd6 h5 5.h3 b5 6.c3 Qxc3 7.Qxh5+ Kg8 8.Qd5+ Kh8 9.Qxa8 Nb6 10.Qe4 Qa1 11.Qf4 Kg8
+- (4.00) Depth: 15/49 00:00:18 38703kN

[Albert Silver]

Super GM.

Of course, when you say Super GM, we mean how would they perform, not how well they 'understand' chess. If it has to do with 'understanding' then they 'understand' nothing at all. Nor do they need to 'understand' a thing. All they need to do is have results against human players equivalent to those of a Super GM. A Super GM in my book is any GM rated around 2700 or more.

I think that this is no longer in doubt.

Albert

[Tanya Deborah]

Though programs do not play too much in human tournaments (because they are fearful and have big ego among some of the negative things) you have some examples lately where machines have won convincingly.

Machines are very strong nowadays. In my opinion I think these could have an elo between 2800 and 3000 points (I am not sure if it can be more than 3000, but I have my doubts) under slow time controls and against the best players.

Of course, I am also assuming that these programs (5 Top) are using at least a Quad machine.

Regards,
Tanya Deborah

[Vinvin]

By the way, what's the correct move for the second position? Toga got stuck on Kd6 in the first few seconds, but I doubt you would post the position if it were that easy.

Second position is a draw, I think (white can make no porgress), but computers give advantage to black because of material.

[BBauer]

Hi,

my first position was not so hard. It could have arised after the opponent had resigned. Here a position from the same game. It should be solvable.
[d]r1b1r1k1/1pBnnpp1/p1q1p2p/8/4N3/6Q1/2PRB1PP/5R1K w - - 22 24


My second position arose in the game Anand-Carlsen in Linares 2008.
[d]8/4k3/5p2/6p1/4P1Bp/5PP1/1r5P/6K1 b - - 0 24
Anand komments 34...Re2!!
and says: 34...Kd6? 35. gxh4 gxh4 36. f4 Rd2 37. Bh3 Kc5 38. e5 Rd4
39. exf6 Rxf4 40. Kg2 Rxf6 41. Bc8 =
He says "when the white king can leave the 1. rank the position is a draw.
Anand won this. He is a GM. Engines seam to have no idea about this - and many other positions.
kind regards
Bernhard

[BBauer]

Here a better example from the same game
[d]r1b1r1k1/1p1nnpp1/pq1Bp2p/8/4N3/6Q1/2PRB1PP/5R1K w KQkq - 20 23

enjoy!
kind regards
Bernhard

[BBauer]

I agree, not many humans will "understand" this position, but an educated chess player will at least know the Saavedra position.
When I give the following study from Platow to Rybka 1.0Betax64 with a CEGT-rating of 2864 and a CCRL-rating of 2921 the engine gives

Code: Select all

  18	00:04	   4.276.673	19.651	+4,03	Kc3b4
  19	00:08	   7.786.554	446.523	+0,09	Kc3b4
  20	00:10	  10.135.652	183.446	+0,03	Kc3b4

So the Elo may be >2800 but the understanding is 0.
Here the position
[d]8/8/8/2P4P/8/2K5/7B/k4r2 w - - 0 1

kind regards
Bernhard