Could A Human Consistently Win Against A Computer? - OzChess

**Gumshoe Mystery** · 01-04-2008, 08:04 PM

With technology as good as it is these days, a win against a computer chess program would be amazing but it could be seen as luck or some sort of mistake on the computers' part.

With the ability for the chess programs to analyse thousands of moves in the time it takes a human to analyse two or three, I feel it would be near on imposssible for a human to win 2 or more matches in a row against a computer.

What are your thoughts on the matter?

**Just2Good** · 01-04-2008, 10:36 PM

Originally Posted by Gumshoe Mystery

With technology as good as it is these days, a win against a computer chess program would be amazing but it could be seen as luck or some sort of mistake on the computers' part.

With the ability for the chess programs to analyse thousands of moves in the time it takes a human to analyse two or three, I feel it would be near on imposssible for a human to win 2 or more matches in a row against a computer.

What are your thoughts on the matter?

Actually Gumshoe, if memory serves even in the 1997 match between world Champion Gary Kasparov and Chess Computer Deep Blue, Deep Blue could calculate millions of moves per second! And Gary still won at least a game in that match.

**scheveningen** · 01-05-2008, 11:06 AM

Rybka is currently playing a series of matches against GM Joel Benjamin and is giving draw odds! Joel gets the win if he can draw. He lost the first and drew (won) the second.

[White "Benjamin"]
[Black "Rybka"]
[WhiteElo "2575"]
[BlackElo "3100"]
[Result "0-1"]
[GameID "235"]
[UniqID "63267"]
[WhiteClock "0:02:21"]
[BlackClock "0:54:06"]
[Stamp "1204"]
[LastMoves "40...a4 41.Kf2 -5.12"]

1.e4 {This is game 1 of draw odds match. Rybka must win to get a point.} e5 2.Nf3 d6 3.d4 exd4 4.Nxd4 Nf6 5.Nc3 Be7 6.Be2 O-O 7.O-O c5 8.Nf3 Nc6 9.Re1 Be6 10.b3 a6 11.Bb2 Re8 12.Bf1 d5 13.exd5 Nxd5 14.Nxd5 Qxd5 15.Qxd5 Bxd5 16.Rad1 Rad8 17.a3 Kf8 18.Bd3 g6 19.Kf1 b5 20.c4 bxc4 21.Bxc4 Bxc4+ 22.bxc4 Rxd1 23.Rxd1 Rb8 24.Rd2 f6 25.Rc2 Kf7 26.Ke2 g5 27.h3 h5 28.Kd1 g4 29.hxg4 hxg4 30.Nh4 Bd6 31.Nf5 Be5 32.Kc1 Kg6 33.Ne3 Bf4 34.g3 Bxe3+ 35.fxe3 Ne5 36.Bxe5 fxe5 37.Rf2 Rb3 38.Kd2 Rxa3 39.Rf8 a5 40.Ke2 a4 41.Kf2 { 0-1} Z0 ( 41...Ra2+ 42.Ke1 a3 43.Ra8 Ra1+ 44.Kd2 a2 45.Kc2 Rg1 46.Rxa2 Rg2+ 47.Kb3 Rxa2 48.Kxa2

**Fischer-Fan** · 01-05-2008, 01:13 PM

Originally Posted by scheveningen

Rybka is currently playing a series of matches against GM Joel Benjamin and is giving draw odds! Joel gets the win if he can draw. He lost the first and drew (won) the second.

in terms of strength how does rybka compare with the computer that kramnik drew a few years ago?

04-18-2008, 08:01 PM

Yes I have done it!

**vagrantnomad** · 04-19-2008, 03:24 PM

Originally Posted by lillywilliam

Yes I have done it!

same here

**vagrantnomad** · 04-19-2008, 03:27 PM

but i must confess, i had the computer set at the chess playing
ability of a monkey

**YumYum** · 04-20-2008, 03:31 PM

Some times....

Rest i can

**Chessic-Adventures** · 04-23-2008, 04:32 AM

I think it is easy to get swept away in the media hysteria surrounding these man .v. computer bouts. Presenting the idea of Kasparov/Kramnik doing battle to save mankind's honour against a whirring silicon foe probably sells quite a few extra magazines and newspapers, but in reality it isn't quite like that. The algorithms used are developed by teams of computer scientists and academics often in consultation with GMs. In earlier days, Kasparov was able to identify and expose weaknesses in the algorithms. With continuing refinements and greater firepower to barge their way through more calculations per second, this becomes less likely in future duels.

For me, what is far more remarkable and infinitely more commendable than the computer's power and increasing ability is that a single individual whether it be Kramnik, Kasparov, whoever is able to beat the best efforts of a large academic community armed with all that fire power.

**sheltered** · 04-29-2008, 07:28 AM

Well it depends, I think if you have the computer playing with some sort of artificial intellegence that will look at the players moves from previous matches, then no I do not think anyone can continuously beat a computer because at one point, it's bound to catch up with them and understand from prior moves where it would go

**ADO_SL** · 04-29-2008, 12:30 PM

One point to make is that with today's engines, since they are essentially just doing search, once you manage to beat an engine once, and you have the game recorded, if you were to replay the game move for move, leaving the engine's settings unchanged, you will beat it again and again forever!

I imagine all of the leading engines today use the same approach:

1. some kind of min-max (minimax, negamax, negascout, mtd) with alpha-beta pruning

2. iterative deepening

3. transposition tables

4. quiescence search; capture search

5. null move heuristic; killer move heuristics

6. parallel computing (multi-thread, multi-process)

7. end game tablebases

The engines differ only in the efficiency of the code (chess board representation, moves, etc), in the speed of the platform (C vs .net vs java) and in the evaluation function (I suppose Rybka's strength -- ie, Vas is an IM)

I'm trying to find a different approach. But it is easier said than done.

**Black-Wolf** · 04-29-2008, 08:00 PM

Originally Posted by ADO_SL

One point to make is that with today's engines, since they are essentially just doing search, once you manage to beat an engine once, and you have the game recorded, if you were to replay the game move for move, leaving the engine's settings unchanged, you will beat it again and again forever!

I imagine all of the leading engines today use the same approach:

1. some kind of min-max (minimax, negamax, negascout, mtd) with alpha-beta pruning

2. iterative deepening

3. transposition tables

4. quiescence search; capture search

5. null move heuristic; killer move heuristics

6. parallel computing (multi-thread, multi-process)

7. end game tablebases

The engines differ only in the efficiency of the code (chess board representation, moves, etc), in the speed of the platform (C vs .net vs java) and in the evaluation function (I suppose Rybka's strength -- ie, Vas is an IM)

I'm trying to find a different approach. But it is easier said than done.

It is well known fact that the Siberian Chess Tiger perverted the nature of chess study by relying too heavily on super computer!

**YumYum** · 08-16-2008, 12:22 AM

Like I said before I usually can. But then again, it is set at easiest level.

**Gumshoe Mystery** · 08-16-2008, 09:41 AM

Originally Posted by YumYum

Like I said before I usually can. But then again, it is set at easiest level.

True, however on an "easy" setting, the computer is programmed to make mistakes to give the user a simple victory. You should challenge yourself by moving up to a harder setting until you find that to be "simple"

**YumYum** · 08-17-2008, 12:39 AM

I've been doing that lately. Last week I started playing at intermediate.

01-04-2008, 08:04 PM	#1 (permalink)
Gumshoe Mystery Senior Member Join Date: Jan 2008 Location: Bundaberg Posts: 141	Could A Human Consistently Win Against A Computer? With technology as good as it is these days, a win against a computer chess program would be amazing but it could be seen as luck or some sort of mistake on the computers' part. With the ability for the chess programs to analyse thousands of moves in the time it takes a human to analyse two or three, I feel it would be near on imposssible for a human to win 2 or more matches in a row against a computer. What are your thoughts on the matter?

01-04-2008, 10:36 PM	#2 (permalink)
Just2Good Tin Cup Champ 2004 Join Date: Jul 2007 Location: Cairns Posts: 6,233 ICC Handle: Advantage FICS Handle: Advantage	Originally Posted by Gumshoe Mystery With technology as good as it is these days, a win against a computer chess program would be amazing but it could be seen as luck or some sort of mistake on the computers' part. With the ability for the chess programs to analyse thousands of moves in the time it takes a human to analyse two or three, I feel it would be near on imposssible for a human to win 2 or more matches in a row against a computer. What are your thoughts on the matter? Actually Gumshoe, if memory serves even in the 1997 match between world Champion Gary Kasparov and Chess Computer Deep Blue, Deep Blue could calculate millions of moves per second! And Gary still won at least a game in that match. __________________ . "The fox knows many things, but the hedgehog knows one big thing." ~ Isaiah Berlin ~

01-05-2008, 01:13 PM	#4 (permalink)
Fischer-Fan Senior Member Join Date: Jul 2007 Location: Perth Posts: 152	Originally Posted by scheveningen Rybka is currently playing a series of matches against GM Joel Benjamin and is giving draw odds! Joel gets the win if he can draw. He lost the first and drew (won) the second. in terms of strength how does rybka compare with the computer that kramnik drew a few years ago? __________________ Patzer see check, patzer give check! - Bobby Fischer

04-18-2008, 08:01 PM	#5 (permalink)
lillywilliam Posts: n/a	Hi! Yes I have done it!

04-19-2008, 03:24 PM	#6 (permalink)
vagrantnomad Senior Member Join Date: Aug 2007 Location: qld at the moment Posts: 272	Originally Posted by lillywilliam Yes I have done it! same here

01-05-2008, 11:06 AM	#3 (permalink)
scheveningen Member Join Date: Jul 2007 Posts: 32	Rybka is currently playing a series of matches against GM Joel Benjamin and is giving draw odds! Joel gets the win if he can draw. He lost the first and drew (won) the second. [White "Benjamin"] [Black "Rybka"] [WhiteElo "2575"] [BlackElo "3100"] [Result "0-1"] [GameID "235"] [UniqID "63267"] [WhiteClock "0:02:21"] [BlackClock "0:54:06"] [Stamp "1204"] [LastMoves "40...a4 41.Kf2 -5.12"] 1.e4 {This is game 1 of draw odds match. Rybka must win to get a point.} e5 2.Nf3 d6 3.d4 exd4 4.Nxd4 Nf6 5.Nc3 Be7 6.Be2 O-O 7.O-O c5 8.Nf3 Nc6 9.Re1 Be6 10.b3 a6 11.Bb2 Re8 12.Bf1 d5 13.exd5 Nxd5 14.Nxd5 Qxd5 15.Qxd5 Bxd5 16.Rad1 Rad8 17.a3 Kf8 18.Bd3 g6 19.Kf1 b5 20.c4 bxc4 21.Bxc4 Bxc4+ 22.bxc4 Rxd1 23.Rxd1 Rb8 24.Rd2 f6 25.Rc2 Kf7 26.Ke2 g5 27.h3 h5 28.Kd1 g4 29.hxg4 hxg4 30.Nh4 Bd6 31.Nf5 Be5 32.Kc1 Kg6 33.Ne3 Bf4 34.g3 Bxe3+ 35.fxe3 Ne5 36.Bxe5 fxe5 37.Rf2 Rb3 38.Kd2 Rxa3 39.Rf8 a5 40.Ke2 a4 41.Kf2 { 0-1} Z0 ( 41...Ra2+ 42.Ke1 a3 43.Ra8 Ra1+ 44.Kd2 a2 45.Kc2 Rg1 46.Rxa2 Rg2+ 47.Kb3 Rxa2 48.Kxa2

04-19-2008, 03:27 PM	#7 (permalink)
vagrantnomad Senior Member Join Date: Aug 2007 Location: qld at the moment Posts: 272	but i must confess, i had the computer set at the chess playing ability of a monkey

04-20-2008, 03:31 PM	#8 (permalink)
YumYum Senior Member Join Date: Apr 2008 Location: Melbourne, Australia Posts: 112	Some times.... Rest i can __________________ Panda is what the world was waiting for!

04-23-2008, 04:32 AM	#9 (permalink)
Chessic-Adventures Senior Member Join Date: Sep 2007 Location: Sydney Posts: 221	I think it is easy to get swept away in the media hysteria surrounding these man .v. computer bouts. Presenting the idea of Kasparov/Kramnik doing battle to save mankind's honour against a whirring silicon foe probably sells quite a few extra magazines and newspapers, but in reality it isn't quite like that. The algorithms used are developed by teams of computer scientists and academics often in consultation with GMs. In earlier days, Kasparov was able to identify and expose weaknesses in the algorithms. With continuing refinements and greater firepower to barge their way through more calculations per second, this becomes less likely in future duels. For me, what is far more remarkable and infinitely more commendable than the computer's power and increasing ability is that a single individual whether it be Kramnik, Kasparov, whoever is able to beat the best efforts of a large academic community armed with all that fire power.

04-29-2008, 07:28 AM	#10 (permalink)
sheltered Junior Member Join Date: Apr 2008 Posts: 14	Well it depends, I think if you have the computer playing with some sort of artificial intellegence that will look at the players moves from previous matches, then no I do not think anyone can continuously beat a computer because at one point, it's bound to catch up with them and understand from prior moves where it would go __________________ Noob Forum – That's right, you read correctly...we love newbies! Register Today – and become part of the action!

04-29-2008, 12:30 PM	#11 (permalink)
ADO_SL Junior Member Join Date: Apr 2008 Posts: 7	One point to make is that with today's engines, since they are essentially just doing search, once you manage to beat an engine once, and you have the game recorded, if you were to replay the game move for move, leaving the engine's settings unchanged, you will beat it again and again forever! I imagine all of the leading engines today use the same approach: 1. some kind of min-max (minimax, negamax, negascout, mtd) with alpha-beta pruning 2. iterative deepening 3. transposition tables 4. quiescence search; capture search 5. null move heuristic; killer move heuristics 6. parallel computing (multi-thread, multi-process) 7. end game tablebases The engines differ only in the efficiency of the code (chess board representation, moves, etc), in the speed of the platform (C vs .net vs java) and in the evaluation function (I suppose Rybka's strength -- ie, Vas is an IM) I'm trying to find a different approach. But it is easier said than done.

04-29-2008, 08:00 PM	#12 (permalink)
Black-Wolf Member Join Date: Mar 2008 Location: Omsk, Siberia Posts: 40	Originally Posted by ADO_SL One point to make is that with today's engines, since they are essentially just doing search, once you manage to beat an engine once, and you have the game recorded, if you were to replay the game move for move, leaving the engine's settings unchanged, you will beat it again and again forever! I imagine all of the leading engines today use the same approach: 1. some kind of min-max (minimax, negamax, negascout, mtd) with alpha-beta pruning 2. iterative deepening 3. transposition tables 4. quiescence search; capture search 5. null move heuristic; killer move heuristics 6. parallel computing (multi-thread, multi-process) 7. end game tablebases The engines differ only in the efficiency of the code (chess board representation, moves, etc), in the speed of the platform (C vs .net vs java) and in the evaluation function (I suppose Rybka's strength -- ie, Vas is an IM) I'm trying to find a different approach. But it is easier said than done. It is well known fact that the Siberian Chess Tiger perverted the nature of chess study by relying too heavily on super computer! __________________ Повинуйтесь и черный волк не съест вас!

08-16-2008, 12:22 AM	#13 (permalink)
YumYum Senior Member Join Date: Apr 2008 Location: Melbourne, Australia Posts: 112	Like I said before I usually can. But then again, it is set at easiest level. __________________ Panda is what the world was waiting for!

08-16-2008, 09:41 AM	#14 (permalink)
Gumshoe Mystery Senior Member Join Date: Jan 2008 Location: Bundaberg Posts: 141	Originally Posted by YumYum Like I said before I usually can. But then again, it is set at easiest level. True, however on an "easy" setting, the computer is programmed to make mistakes to give the user a simple victory. You should challenge yourself by moving up to a harder setting until you find that to be "simple" __________________ “When we remember we are all mad, the mysteries disappear and life stands explained.”

08-17-2008, 12:39 AM	#15 (permalink)
YumYum Senior Member Join Date: Apr 2008 Location: Melbourne, Australia Posts: 112	I've been doing that lately. Last week I started playing at intermediate. __________________ Panda is what the world was waiting for!

Thread Tools
Show Printable Version Email this Page
Rate This Thread
Rate This Thread: