Live engine tournament

[Lyudmil Tsvetkov]

And here's the same data sorted by opening name and variation:

Code: Select all

60	B13	Caro-Kann	"Exchange, 4.Bd3 Nc6 5.c3 Nf6 6.Nf3"
29	B13	Caro-Kann	"Panov-Botvinnik, 5...Nc6 6.Nf3"
3	 E08	Catalan	"Closed, 7.Qc2 c6 8.Bf4"
27	D05	Colle	"5.c3 Nc6"
66	C47	Four Knights	"Scotch, 4...exd4 5.Nxd4 d6"
19	C02	French	"Advance, Ruisdonk Gambit, 1.e4 e6 2.d4 d5 3.e5 c5 4.Nf3 cxd4 5.Bd3"
11	C14	French	"Classical, Steinitz, 7...a6 8.Nf3 c5"
71	C14	French	"Classical, Steinitz, 7...a6"
20	C01	French	"Exchange, 4.Nf3 Bd6 5.Bd3 Nf6"
23	C01	French	"Exchange, 4.Nf3 Bd6 5.Bd3 Nf6"
32	C01	French	"Exchange, 4.Nf3 Bd6 5.Bd3 Nf6"
48	C01	French	"Exchange, 4.Nf3 Bd6 5.Bd3 Nf6"
69	C01	French	"Exchange, 4.Nf3 Bd6 5.Bd3 Nf6"
42	C11	French	"Steinitz, Boleslavsky, 7...a6"
49	C08	French	"Tarrasch, Open, Main Line, 4.exd5 exd5 5.Ngf3 Nf6, 10.Nb3"
46	C09	French	"Tarrasch, Open, Main Line, 6.Bb5 Bd6 7.dxc5 Bxc5 8.O-O Ne7"
9	 C54	Giuoco Pianissimo	"5.d3 d6 6.O-O O-O"
6	 D95	Grünfeld	"5.e3 O-O 6.Qb3 dxc4 7.Bxc4 Nc6"
1	 D87	Grünfeld	"Classical Exchange, 8...c5 9.Be3 Nc6"
34	E65	King's Indian	"Fianchetto, Yugoslav, 7.O-O Nc6 8.h3"
64	E27	Nimzo-Indian	"Sämisch, 5...O-O 6.f3 d5 7.e3"
47	E24	Nimzo-Indian	"Sämisch, Botvinnik Variation"
72	E24	Nimzo-Indian	Sämisch, Botvinnik Variation
7	 D22	QGA	"Alekhine, 4.e3 e6 5.Bxc4"
43	D27	QGA	"Classical, 6...a6 7.Bb3 cxd4"
16	D27	QGA	"Classical, Rubinstein, 8.Nc3 Be7"
17	D37	QGD	"Classical, 5...O-O 6.e3 c5 7.dxc5 Bxc5 8.a3"
14	D37	QGD	"Classical, 5...O-O 6.e3 c5 7.dxc5"
65	D37	QGD	"Classical, 5...O-O 6.e3 Nbd7"
57	D35	QGD	"Exchange, 5.Nf3 Be7 6.Bf4 c6 7.Qc2 Nbd7"
21	D55	QGD	"Neo-Orthodox, Petrosian Variation"
53	D61	QGD	"Orthodox, Rubinstein, 7.Qc2 h6 8.Bh4 c5"
63	D63	QGD	"Orthodox, Swiss, Karlsbad, 9.Bd3 Re8"
22	D38	QGD	"Ragozin, 5.Bg5 h6 6.Bxf6 Qxf6 7.cxd5"
33	D40	QGD	"Semi-Tarrasch, 5.e3 Nc6 6.a3 a6 7.dxc5 Bxc5"
30	D40	QGD	"Semi-Tarrasch, 5.e3 Nc6 6.a3 cxd4 7.exd4 Be7 8.c5"
61	D41	QGD	"Semi-Tarrasch, 6.e3 Nc6 7.Bc4"
62	D40	QGD	"Semi-Tarrasch, Symmetrical Variation"
58	D30	QGD	"Semmering Variation"
12	D30	QGD	"Spielmann Variation"
52	D32	QGD Tarrasch	"5.Nf3 Nc6 6.e3 Nf6"
41	E14	Queen's Indian	"4.e3 Bb7 5.Bd3 d5 6.O-O Bd6 7.Nc3"
40	E12	Queen's Indian	"Petrosian, 5.Nc3 Be7"
15	E12	Queen's Indian	"Petrosian, 5.Nc3 g6"
37	D02	Queen's Pawn	"2.Nf3 Bf5 3.Bf4 e6"
31	D02	Queen's Pawn	"London"
45	B01	Scandinavian	"Marshall, 4.Nf3 g6 5.Be2"
44	C45	Scotch	"4.Nxd4 Nf6 5.Nxc6 bxc6 6.Bd3"
36	D47	Semi-Slav	"Meran Variation"
55	D43	Semi-Slav	"Moscow, 6.Bh4 Gambit, Main Line, 9...Bb7"
38	B50	Sicilian	"2.Nf3 d6 3.c3 Nf6 4.h3 Nc6"
13	B52	Sicilian	"3.Bb5+ Bd7 4.Bxd7+ Nxd7 5.O-O Ngf6 6.Qe2"
54	B22	Sicilian	"Alapin, 2...e6 3.d4 d5 4.exd5 exd5 5.Nf3 Nc6"
26	B30	Sicilian	"Rossolimo, 3...e6 4.O-O Nge7 5.c3 a6"
39	B31	Sicilian	"Rossolimo, 3...g6 4.O-O Bg7 5.Re1 Nf6 6.c3 O-O 7.h3"
8	 B21	Sicilian	"Smith-Morra, 4.Nxc3 Nc6 5.Nf3 e6"
25	D10	Slav	"3.Nc3 Nf6 4.e3 a6 5.Qc2 b5"
18	C68	Spanish	"Exchange, 5.O-O Bg4 6.h3"
28	C68	Spanish	"Exchange, 5.O-O Bg4 6.h3"
59	C68	Spanish	"Exchange, 5.O-O Bg4 6.h3"
5	 C68	Spanish	"Exchange, Alapin Gambit, 7.d3 Qf6 8.Nbd2"
24	C68	Spanish	"Exchange, Alapin Gambit, 7.d3 Qf6 8.Nbd2"
51	C68	Spanish	"Exchange, Alapin Gambit, 7.d3 Qf6 8.Nbd2"
67	C68	Spanish	"Exchange, Alapin Gambit, 7.d3 Qf6 8.Nbd2"
2	 C69	Spanish	"Exchange, Gligoric, 6.d4 exd4 7.Nxd4 Ne7"
70	C69	Spanish	"Exchange, Gligoric, 6.d4 exd4 7.Nxd4 Ne7"
50	C69	Spanish	"Exchange, Gligoric, 6.d4 exd4 7.Qxd4"
68	C68	Spanish	"Exchange, Romanovsky Variation"
4	 A46	Torre Attack	"3...h6 4.Bxf6 Qxf6 5.e4 c5"
35	C58	Two Knights	"Morphy, Bogoljubow Variation"
10	C58	Two Knights	"Morphy, Kieseritsky Variation, 6...h6 7.Nf3 e4 8.Qe2"
56	C58	Two Knights	"Morphy, Kieseritsky Variation, 6...h6 7.Nf3 e4 8.Qe2"

All data generated by Arena 3.0

Many French?, a fair portion of QGD, just a few Sicilians??, many Ruy Lopez Exchange??, a single KID??, a single Catalan?, just 2 Caro-Kann?

Who can claim that engines have even the slightest clue of opening understanding after that?

Instead, well informed engines would play: just a few French, many Sicilians, very few if at all Ruy Lopez Exchange, many many KIDs, many Catalans, and many Caro-Kann.

They got it all the wrong way.

[kranium]

Many French?, a fair portion of QGD, just a few Sicilians??, many Ruy Lopez Exchange??, a single KID??, a single Catalan?, just 2 Caro-Kann?

Who can claim that engines have even the slightest clue of opening understanding after that?

Instead, well informed engines would play: just a few French, many Sicilians, very few if at all Ruy Lopez Exchange, many many KIDs, many Catalans, and many Caro-Kann.

They got it all the wrong way.

Hi Lyudmil,

I hope you're not basing your conclusions on results from some mega database of GM games, which are rife with 'human error'.

But I do agree with you that some openings seem superior to others, and I believe that eventually, this will be empirically proven by computers.
However, this work has not yet been done...and will likely require the resources of super computers that don't even exist yet.

In the meantime, top engine's are no match for any human, no matter which opening they choose.

We all know that the human race has a deep need to believe they're superior to machines...
but, in chess all that's left for them to cling to is old error ridden opening theory than can maybe steer them into closed positions!


PS-
Established opening theory is getting better and better, but only thanks to diligent application of the very engines you reference above!

Regards-
Norm

[Lyudmil Tsvetkov]

Many French?, a fair portion of QGD, just a few Sicilians??, many Ruy Lopez Exchange??, a single KID??, a single Catalan?, just 2 Caro-Kann?

Who can claim that engines have even the slightest clue of opening understanding after that?

Instead, well informed engines would play: just a few French, many Sicilians, very few if at all Ruy Lopez Exchange, many many KIDs, many Catalans, and many Caro-Kann.

They got it all the wrong way.

Hi Lyudmil,

I hope you're not basing your conclusions on results from some mega database of GM games, which are rife with 'human error'.

But I do agree with you that some openings seem superior to others, and I believe that it will be empirically proven by computers.
However, this work has not yet been done...and will likely require the resources of super computers that don't even exist yet.

In the meantime, top engine's are no match for any human, no matter which opening they choose.

We all know that the human race has a deep need to believe they're superior to machines...
but, in chess all that's left for them to cling to is error ridden opening theory than can maybe steer them into closed positions!

Hi Norman.

Typical computer programmer statement.

As said, engines can not surpass the human race that has studied those openings for decades and centuries.

I am basing my assertions on the openings Kasparov and Fischer have played, no magabases involved.

Some openings are really better than others, this certainly is so, as the initial position is asymmetric in nature.

Engine search is simply too limited to see why some openings are better than others. Engines are stronger than humans, but only at shallow depths.

At bigger depths, with closed positions, humans are stronger.

Btw., I used to play a lot with Fire in the past, what I remember is the nice positional style it had. Unfortunately, not much time to play latest Fire now...

[kranium]

Lyudmil,

Well, I respectfully disagree with most of what you're saying here...

I am basing my assertions on the openings Kasparov and Fischer have played, no magabases involved.

vs Deep Blue, Kasparov managed to win one match and lose the next, but those 2 matches were almost 20 years ago and Deep Blue was quite a "primitive" chess program (in today's standards).

As said, engines can not surpass the human race that has studied those openings for decades and centuries.

I'm familiar with some of the centuries old opening book theory you refer to...
didn't they play things like Scotch, Giuoco Piano, Petroff, Bishop's, Damiano's, Philidor's, Ponziani, etc?
(none of which has been played so far in the tournament which this topic concerns, BTW)

Engine search is simply too limited to see why some openings are better than others. Engines are stronger than humans, but only at shallow depths.
At bigger depths, with closed positions, humans are stronger.

Is there any human that searches deeper than today's engines?
(is it even humanly possible?)

Norm

[Lyudmil Tsvetkov]

Lyudmil,

Well, I respectfully disagree with most of what you're saying here...

I am basing my assertions on the openings Kasparov and Fischer have played, no magabases involved.

vs Deep Blue, Kasparov managed to win one match and lose the next, but those 2 matches were almost 20 years ago and Deep Blue was quite a "primitive" chess program (in today's standards).

As said, engines can not surpass the human race that has studied those openings for decades and centuries.

I'm familiar with some of the centuries old opening book theory you refer to...
didn't they play things like Scotch, Giuoco Piano, Petroff, Bishop's, Damiano's, Philidor's, Ponziani, etc?
(none of which has been played so far in the tournament which this topic concerns, BTW)

Engine search is simply too limited to see why some openings are better than others. Engines are stronger than humans, but only at shallow depths.
At bigger depths, with closed positions, humans are stronger.

Is there any human that searches deeper than today's engines?
(is it even humanly possible?)

Norm

Norm, what is certain is that the opening is the most complex part of the game.
If engines misplay on occasions even simple endgame positions, why should not they do so in the opening?

You can easily compare with endgames: top engines might beat any human in endgames, but they still play this part of the game worse, when there are specific patterns to recognise. The human will make a msitake and the computer will win, but still the human approach will be sounder.

This is even truer of openings.

Believe me, engines play openings awfully.

[kranium]

Believe me, engines play openings awfully.

If that's true, in the most recent Human vs Computer matches, it didn't seem to hurt them...
and conversely, pretentious 'superior' human opening knowledge didn't matter either:

"In October 2004, Ruslan Ponomariov, Veselin Topalov and Sergey Karjakin played against computers Hydra, Fritz 8, and Deep Junior.
Ponomariov and Topalov were FIDE world chess champions. Sergey Karjakin at 12 was the youngest Grandmaster.
The computers won 8.5 to 3.5. "

"In November 2005, 3 former FIDE world chess champions, Alexander Khalifman, Ruslan Ponomariov and Rustam Kasimdzhanov played against computers Hydra, Junior and Fritz.
The computers won 8 to 4."

"In 2005, Hydra, a dedicated chess computer with custom hardware and sixty-four processors and also winner of the 14th IPCCC in 2005,
crushed seventh-ranked Michael Adams 5½–½ in a six-game match."

"The chess engine Hiarcs 13 runnning inside Pocket Fritz 4 on the mobile phone HTC Touch HD won the Copa Mercosur tournament in Buenos Aires, Argentina
with 9 wins and 1 draw on August 4–14, 2009."

"Pocket Fritz 3 using version 12.1 of Hiarcs won the same event the previous year with six wins and four draws, running on a 624 MHz iPaq hx2790.
The 2008 Mercosur Cup was a category 7 tournament."

http://en.wikipedia.org/wiki/Human%E2%8 ... ss_matches

and these matches were played against (relatively) weak engines on slow hardware and mobile phones.

But, that being said, I have one question for you...
well prepared and armed with all the best human opening book knowledge,
what might be the result of an OTB match Lyudmil Tsvetkov vs Stockfish 6 (no book) today?

[Matthias Gemuh]

...
Believe me, engines play openings awfully.

Didn't Stockfish, playing without book and a pawn less, crush mighty Nakamura in 2014 ?

Do you define "awfully" as "very well" ?

[Lyudmil Tsvetkov]

Believe me, engines play openings awfully.

If that's true, in the most recent Human vs Computer matches, it didn't seem to hurt them...
and conversely, pretentious 'superior' human opening knowledge didn't matter either:

"In October 2004, Ruslan Ponomariov, Veselin Topalov and Sergey Karjakin played against computers Hydra, Fritz 8, and Deep Junior.
Ponomariov and Topalov were FIDE world chess champions. Sergey Karjakin at 12 was the youngest Grandmaster.
The computers won 8.5 to 3.5. "

"In November 2005, 3 former FIDE world chess champions, Alexander Khalifman, Ruslan Ponomariov and Rustam Kasimdzhanov played against computers Hydra, Junior and Fritz.
The computers won 8 to 4."

"In 2005, Hydra, a dedicated chess computer with custom hardware and sixty-four processors and also winner of the 14th IPCCC in 2005,
crushed seventh-ranked Michael Adams 5½–½ in a six-game match."

"The chess engine Hiarcs 13 runnning inside Pocket Fritz 4 on the mobile phone HTC Touch HD won the Copa Mercosur tournament in Buenos Aires, Argentina
with 9 wins and 1 draw on August 4–14, 2009."

"Pocket Fritz 3 using version 12.1 of Hiarcs won the same event the previous year with six wins and four draws, running on a 624 MHz iPaq hx2790.
The 2008 Mercosur Cup was a category 7 tournament."

http://en.wikipedia.org/wiki/Human%E2%8 ... ss_matches

and these matches were played against (relatively) weak engines on slow hardware and mobile phones.

But, that being said, I have one question for you...
well prepared and armed with all the best human opening book knowledge,
what might be the result of an OTB match Lyudmil Tsvetkov vs Stockfish 6 (no book) today?

The statistics you provide are from long ago: humans have improved over the last years.

I never use opening books when playing against engines.

I am certain I will win an OTB match against latest SF, well, the margin may be small, but I will still win.

[Lyudmil Tsvetkov]

...
Believe me, engines play openings awfully.

Didn't Stockfish, playing without book and a pawn less, crush mighty Nakamura in 2014 ?

Do you define "awfully" as "very well" ?

You need a positional player for that kind of combat. Tactics do not help against engines.

The pawn more was actually an edge h pawn in one case, giving almost no advantage, some 20cps maybe, and a b file pawn in the other case, giving 70-80cps advantage, but not a full pawn.

[Matthias Gemuh]

...
I am certain I will win an OTB match against latest SF, well, the margin may be small, but I will still win.

Stop assuming that your chess strength is above Elo 3000.