• Forums
• Chess.com Community

Human v Computer

• 

  Clavius
  #1 Jun 24, 2008

  Many computer team members playing the Human v Computer game are missing out.  The forum (below the game board) has a highly instructive daily discussion about the best move that can help you see how the high-rated players on our team, and the even higher rated software, analyze a position.  (Rybka, Hiarcs, Fritz and Fruit are software used by our team that are all rated close to or above Elo 2900, far ahead of other commercial chess software, and they play at an even higher level when we let them analyze for hours).

  
  
• 

  chessballer2
  #2 Jun 24, 2008

  WOOOOOOWWWWW!!!!!

  What was that supposed to mean? Those players are that high? That's amazizing.

  
  
• 

  Clavius
  #3 Jun 24, 2008

  I edited my post to clarify that Rybka, Hiarcs, Fritz and Fruit are computer programs used by team members.  Aside from Fritz, the programmers don't advertise them much but they are far superior to Chessmaster, The King and other better known programs. 

  By the way, it would help the team if members hold their votes until the last 12 hours so so they can benefit from the forum discussion and software analysis.  Too many team members are voting without benefit of seeing what the best software has to say. 

  
  
• 

  Ellbert
  #4 Jun 24, 2008

  Doctor Clavius, do you offer Chess as a way for people to learn how to relax ?

  
  
• 

  Clavius
  #5 Jun 24, 2008
  Chess is a great way to take your mind off everything else.  Consequently I recommend it as one of a number of stress relievers/self care ideas on page 114 of my book.  (See www.stressillness.com for more information).
  
• 

  Clavius
  #6 Jun 25, 2008
  By the way, being part of the computer team (and reading the forums to participate in consensus choice) will enable you to play chess at 3000 Elo given how long we let the software analyze.  Just to see a game played at that level against strong opposition should be fascinating.
  
• 

  Arcturus10
  #7 Jul 1, 2008
  These engines are very good, but just keep in mind that are aren't right all of the time; they tend to miss certain specific kinds of moves like piece sacrifices and more strategic moves.  Webgogs has an example of an exchange sac in his blog where most engines miss the correct move.
  
• 

  Clavius
  #8 Jul 4, 2008
  Top level software like Rybka, Hiarcs, Fritz, Shredder, Fruit can find correct exchange sacrifices nowadays.  The program I use, Hiarcs, found the exchange sac in Webgogs blog in less than a minute.
  
• 

  Hydrocannon
  #9 Jul 4, 2008
  Big words=Faint
  
• 

  MM78
  #10 Jul 4, 2008
  AS I'm on the human team I don't get to view that forum but I look forward to reading it after the game and learning about the discussions.  If the computers can truly operate at 3000 ELO we should have no chance...........
  
• 

  Clavius
  #11 Jul 4, 2008
  Unfortunately some computer team members are voting based on much weaker software and are not reading the forum to see what the top level software is recommending.  The computer team has played some 2nd-rate moves as a result.  Also, strong humans given 24 hours to consider a move will also undoubtedly play higher than their over-the-board elo rating.  The result so far is an even game played at a high and instructive level.
  
• 

  Arcturus10
  #12 Jul 6, 2008
  Clavius, on move 20 in the GM Nigel Davies game a blunder was made Nxe4.  The correct move would have been Ne5.  If you have time, can you analyze with Hiarcs 12 mp to see if it finds N35?  I tried with Hiarcs 11 and Rybka 2.3.2a 32 bit after the blunder had been made and they couldn't find it in 10 minutes.  Strangely enough, they picked Nxe4.
  
• 

  Clavius
  #13 Jul 6, 2008

  Hello Arcturus.  Hiarcs 12.1mp running on a Mac with an Intel Duo Core processor needed 4 minutes and 16 ply to decide that 20...Ne5 was a little better than 20...Nxe4.  After 17 ply (8 minutes) it made a slight modification and recommended either 20...Kh8  21.Nc3 Ne5 or 20...b6  21.Rad1 Ne5 but the Knight still goes to e5 in both of those lines.  The evaluation was 1.37 pawns in White's favor.  For 20...Nxe4 the evaluation was about 1.5 pawns in White's favor.

  How does Rybka do with Webgog's position (borrowed from Nunn) in his blog "Is your computer blind?" 

  
  
• 

  Arcturus10
  #14 Jul 6, 2008
  If I remember correctly (I can't run it right now unfortunatly), Rybka finds it in about a minute, although it's intresting to note that a tactical version of Rybka called Winfinder, meant to find deep tactical continuations that others may miss.  It found the correct move from Webgog's position instantly (ply 1), but none of my engines,  disappointingly, could find Ne5 as far as I can remember.  If I get a chance I'll try to analyze the position from GM Davies game again. My computer has only one processor, but I'm hoping to get a Duo Core Processor laptop soon.