Sports


Chess engines




Chess Piece
Bobby Ang


Posted on April 06, 2017


I believe it is generally accepted that Stockfish, Komodo and Houdini are the top three chess engines in the world. The Computer Chess Ratings List (CCRL) lists their ratings as:


Stockfish ...... 8 3390


Houdini ..... 5.01 3387


Komodo ...... 10.3 3380

Stockfish, Houdini and Komodo are bunched together at the top with only a few points separating them. The next in the list, Deep Shredder 13, is a hundred points behind. This, obviously, is quite a large gap.

At the latest TCEC (Top Chess Engine competition) Chess Championship where the three of them participated it was Stockfish 8 which won the crown. Clearly, therefore it is an engine everyone should have, especially since it is free. You can get it from stockfishchess.org -- it is a zip file which you simply unzip into one of your computer folders. There are two versions included -- 32-bit and 64-bit for use with whatever version of Windows you are on.

This is a Universal Chess Interface (UCI) engine, meaning most chess programs can be used together with this engine. For example, if you have chess arena (the one I told you about last Tuesday), you can go to the “engines” tab, click “install new engine,” point it at stockfish_8_x64.exe, double-click and you are done.

When you look at the results of the analysis you will see something like this:

23 01.30 265,728k 2.933k -0.04 then some chess moves.

“23” means the number of ply, or half-moves analyzed. 2 ply, of course, is equivalent to 1 move. 23 ply, as in the example above, means that the engine looked 11.5 moves ahead.

“01.30” means that so far the engine has consumed 1 minute and 30 seconds to analyze the position

“265,728k” it has looked at 265.7 million positions

“2.933k” is the speed of the engine -- in this example it looks at almost 3,000 positions per second.

“-0.04” is the evaluation of the position. 1 means one pawn, 2 means 2 pawns, etc. etc. The normal convention is that the evaluation is always from white’s point of view. 1.5 therefore means that white is the equivalent of a pawn and a half ahead. -1.5 means that Black is ahead by a pawn and a half. You have to take this figuratively. For example if white sacrifices a pawn Black should be ahead -1.00. If the evaluation reads only -0.65 that means White has compensation.

The moves given is currently what the computer considers the best line so far.

Word of warning -- just because Stockfish 8 is the highest rated engine and won the last world championship does not mean it is the best with which to analyze games. The ratings are based on computer vs computer competitions -- in many cases the computer need not find the best move -- all that is necessary is to play a move which does not lose. This is not compatible with game analysis where we have to find the truth in the position, the absolutely best move.

During the computer chess seminar conducted two Saturdays ago, I gave an example of this.

Lutikov, Anatoly S -- Silva, Fernando [B31]

Odessa, 1976

This game is from the tournament where Rosendo Balinas got his GM title. I wrote a series of articles for BusinessWorld entitled “The Conquest of Odessa” and put this together for my book Inside Philippine Chess Volume 1. I went through every single game of the tournament for this purpose and this one really captivated me.

GM Anatoly Lutikov is a fiery tactician who would have achieved a lot more if he were not best friends with the bottle. There is a very nice story associated with him.

Genna Sosonko wrote that back in 1958 Bobby Fischer was in Moscow, showed up in the Central Chess Club and started clobbering everyone in blitz. It became a matter of pride and the club officials called in their top GMs to defend their honor but the American kept winning and winning. Finally they brought in Lutikov and here Fischer met his match. They contested around 30 games and Lutikov won two thirds of them. Thus, he upheld the “superiority of Soviet chess.”

1.e4 c5 2.Nf3 Nc6 3.Bb5 g6 4.0 -- 0 Bg7 5.c3 Nf6 6.d4 Nxe4 7.d5 Nb8 8.Re1 Nd6 9.Bf4 Nxb5 10.a4 0 -- 0 11.axb5 d6 12.Qe2 e5 13.dxe6 Bxe6 14.Ng5 Re8 15.Ne4 Bd7 16.Qd2! Bxb5 17.Na3 Bc6 18.Nxd6 Rxe1+ 19.Rxe1 Na6 20.Nac4 Nc7 21.Ne5 Bxe5 22.Bxe5 Ne6 23.b4! cxb4 24.cxb4 a6 25.Qa2! Qd7 26.b5! Bxb5 27.Ne4 Qd3 28.Nf6+ Kf8 29.Nxh7+ Kg8 30.Nf6+ Kf8 31.h3 Rc8 32.Qa1 Rc2 33.Rd1 Qe2 34.Bg3 Bc6 35.Kh2 g5

Now comes the brilliant finish.

36.Rd8+!! Ke7

Mate follows 36...Nxd8 37.Bd6+ Kg7 38.Ne8+ Kg6 39.Qf6+ Kh5 40.Ng7+ Kh4 41.g3#

The position on the board now is the one I pointed out to the seminar participants. Stockfish has a curious blind spot here and insists that White should play 37.Rb8, which also wins. Komodo and Houdini both agree with Lutikov that there is a much quicker finish:

37.Rd7+! Bxd7 38.Nd5+ Kd8

[38...Ke8 39.Qf6]

39.Qf6+ Kc8 40.Nb6#! 1 -- 0

When it comes to long-range straightforward calculation (you know, if I go here you will go there type of analysis) Stockfish is without a doubt (for me, at least) the fastest and best. Here is what I mean.

* * *
Ivanchuk, Vassily (2715) -- Jobava, Baadur (2727) [B32]
77th Tata Steel GpA Wijk aan Zee NED (1.2), 10.01.2015

1.e4 c5 2.Nf3 Nc6 3.d4 cxd4 4.Nxd4 e5 5.Nb5 d6 6.c4 g6 7.Be3 Be6 8.N1c3 a6 9.Na3 Nf6 10.Be2 Bg7 11.Nc2 Rc8 12.Rc1 0 -- 0 13.0 -- 0 Ne7 14.Na3 Nc6 15.Re1 Nd4 16.f3 Qb6 17.Rb1 Nxe2+ 18.Rxe2 Qb4 19.Rd2 Rfd8 20.c5?!

In the after-game interview, Ivanchuk commented that 20.Qb3 might be better than the move he actually played.

20...Qa5 21.Rxd6 Bf8 22.Rxd8 Rxd8 23.Qe2 Bxc5 24.Nc2 Bd4 25.Qe1 b5 26.Nxd4! exd4 27.Bg5 dxc3 28.Qxc3 Qxc3 29.bxc3 Bxa2 30.Bxf6 Rd6 31.Be5 Re6 32.Ra1 Rxe5 33.Rxa2 Re6 34.Kf2 Kf8 35.Ke3 Ke7 36.g4 Kd7 37.f4 Rc6 38.Kd4 Kc7!?

The position is equal but as usual Jobava takes risks to get a win.

39.f5 Rc4+ 40.Kd5 Kb6 41.Ra3 a5 42.e5 Ra4?

A blunder, Black has to play 42...gxf5 43.gxf5 Rf4 with a probable draw.

43.Rxa4 1 -- 0

Most of the engines were showing that the position is equal, but then Jobava suddenly resigned. Ivanchuk went to the commentator booth and demonstrated the win, saying with a smile that this is “not so difficult for a good grandmaster.” -- 43.Rxa4 bxa4 44.Kc4 Kc6 45.Kd3 gxf5 (45...Kd5 46.e6 fxe6 47.f6! we have a similar position as in the actual game) 46.gxf5 Kd5 47.e6 fxe6 48.f6 Kd6 49.c4 e5 50.c5+ Ke6 51.c6 e4+ 52.Kc2 and the white pawns cannot be stopped.

Stockfish 8 finds the win after 43.Rxa4 in a split-second. Long-range calculation like that is completely its forte.

However, to get to the truth of a position, where straight calculation is not enough but “understanding” the subtle nuances of the position is important, I find that Komodo 10.4 is to be preferred. I have the perfect example here -- the famous 9th game of the 1st Karpov vs Kasparov world championship match.

* * *
Karpov, Anatoly -- Kasparov, Garry [D34]
Moscow-Wch Moscow (9), 05.10.1984

1.d4 d5 2.c4 e6 3.Nf3 c5 4.cxd5 exd5 5.g3 Nf6 6.Bg2 Be7 7.0 -- 0 0 -- 0 8.Nc3 Nc6 9.Bg5 cxd4 10.Nxd4 h6 11.Be3 Re8 12.Qb3 Na5 13.Qc2 Bg4 14.Nf5 Rc8 15.Bd4 Bc5 16.Bxc5 Rxc5 17.Ne3 Be6?!

[17...d4! 18.Rad1 d3!]

18.Rad1 Qc8 19.Qa4 Rd8 20.Rd3 a6 21.Rfd1 Nc4 22.Nxc4 Rxc4 23.Qa5 Rc5 24.Qb6 Rd7 25.Rd4 Qc7 26.Qxc7 Rdxc7 27.h3 h5 28.a3 g6 29.e3 Kg7 30.Kh2 Rc4 31.Bf3 b5 32.Kg2 R7c5 33.Rxc4 Rxc4 34.Rd4 Kf8 35.Be2 Rxd4 36.exd4 Ke7 37.Na2 Bc8 38.Nb4 Kd6 39.f3 Ng8 40.h4 Nh6 41.Kf2 Nf5 42.Nc2 f6?!

The sealed move. Nikitin: “Everyone thought that the adjourned position was drawn, but the secret move sealed by Kasparov turned out to be bad and Black unexpectedly found himself with problems. All night the trainers looked for a draw, without finding an accurate way to achieve one, or a clear win for Karpov. This was a genuine brain-storming session, but in the critical match situation the analysis was extremely uneven, unsystematic and therefore ineffective. Garry went along to the resumption in an angry mood and he played weakly that evening.”

43.Bd3 g5 44.Bxf5 Bxf5 45.Ne3 Bb1 46.b4 <D>

POSITION AFTER 46.B4


Kasparov commented here that the queenside is already locked up and the only way White can penetrate into the position is via the kingside, but it is not clear how he is to do it.]

46...gxh4 47.Ng2!!

A shocker and completely overlooked by Kasparov and his seconds. The move is counter-intuitive but the only way to make inroads into Black’s position. Recapturing with the pawn puts it on h4 -- White needs that square to penetrate. Komodo takes only seconds to find this move, while the other engines are completely blind to this possibility.

47...hxg3+ 48.Kxg3 Ke6 49.Nf4+ Kf5 50.Nxh5 Ke6 51.Nf4+ Kd6 52.Kg4 Bc2 53.Kh5 Bd1 54.Kg6 Ke7 55.Nxd5+ Ke6 56.Nc7+ Kd7 57.Nxa6 Bxf3 58.Kxf6 Kd6 59.Kf5 Kd5 60.Kf4 Bh1 61.Ke3 Kc4 62.Nc5 Bc6 63.Nd3 Bg2 64.Ne5+ Kc3 65.Ng6 Kc4 66.Ne7 Bb7 67.Nf5 Bg2 68.Nd6+ Kb3 69.Nxb5 Ka4 70.Nd6 1 -- 0

As I explained to the audience, Stockfish 8 is the reigning world champion and very strong and it is free, so everyone must go and get it. Now, if you have some money set aside you should invest $60 and buy Komodo 10.4. Having both engines you can be reasonably confident that your chess engine stable is not inferior to anyone, even the World Chess Champion.

On Tuesday, let us talk about the endgame tablebases.

Bobby Ang is a founding member of the National Chess Federation of the Philippines (NCFP) and its first Executive Director. A Certified Public Accountant (CPA), he taught accounting in the University of Santo Tomas (UST) or 25 years and is currently Chief Audit Executive of the Equicom Group of Companies.

bobby@cpamd.net