Sorry to go off-topic, but there were some interesting points made on computer analysis that I would like to comment on.

Quote:


Not only can, but a team of average human + average engine will beat even the strongest engines in the vast majority of games, as shown in the ChessBase Freestyle tournaments where if I recall correctly, even Rybka and Fritz (without any human interference during the game) did not manage to qualify. I agree and have stated before that computer evaluations cannot be trusted above a human's evaluation, unless one side is winning. Even then, the computer may suggest continuations that are not the easiest and most practical way of winning the game.

Quote:


I'm not convinced. In the 2005 Freestyle Chess Tournament, two amateurs (one rated in the 1300s and one in the 1200s), playing under the handle 'ZackS', defeated teams of Grandmasters equipped with analysis engines to win first prize, defeating a team of 2600 GMs and 3000 strength engines in the final. Unfortunately I can't find the article explaining how ZackS won the tournament in their own words, but basically they used a superior analysis process with their engines than the other teams. And thus, two amateurs with a superior analysis process proved stronger than Grandmaster + engine centaur teams with an inferior analysis process. 

I looked at the analysis of one of ZackS's games from the final, and the analysis was of a decent quality. 

Quote:


13.Nd5 is the move Black is begging White to play - why play a move that solves all of Black's positional problems and simultaneously loses a tempo for no good reason?

SWJediknight wrote on 03/29/10 at 00:15:31:



From my correspondence experience , and from this particular doubt which  I had few months back in analysis i Asked GM Uri Blass to confirm this. And what i found is 
1. Rybka is not optimized to play before moves 10 (Vas)
2. Use Naum 4.x for 1e4 e5 as its the best.
3. Close Position ? Shredder is King
4. Attacking Moves ? Fritz 10 and Hiarcs PB 2007.
5 Opening Novelties Shredder 12,Naum 4.2  Stockfish 1.6s , Fritz 12 and Shreddy 12 boy.
6. More pieces  and material imbalance Rybka 3.
So i would suggest use Naum 4. If you need it plz PM  me.

Is there a specific engine which is supposed to be good in these Two Knights lines?  I note that Fritz tends to agree with most of the PGN's assessments in the 6.d4 lines, but regards some of the play as sub-optimal (e.g. Kaissiber 29's suggested improvement 12.Nc3! over the old 12.Na3 is given by Fritz as best after about five seconds' thought, assessing it as +=).  

But in contrast, Fritz goes further than my "roughly equal" assessment of the lines I gave after 9.Qe4 and 9.0-0 in the Fried Liver, often giving "+=" or even "+/-".

BPaulsen wrote on 03/28/10 at 04:11:26:



Let me explain you bit by bit as you dont seem to understand properly about engine analysis.
First the thing is For how much time you run an engine it evaluates the position as it was programmed. But The Analyser must know What and How to run them . There is no human present on earth that can beat a Computer program Tactics and for your kind information tactical ideaas comes from positional plans or strategical plans. If you think an engine is tacticaly stronger doesnot understand Strategy then it would be a dumb thing from you. That was the reason why Vas made 4 Engines  of Rybka just to use in analysis and GM Larry kaufman has put Positional ideas in Rybka 3 just for that and it is the best engine so far. So if you think Engines dont understand then that is foolish.
Second thing is all engines are unique to a particular position ( and that i can prove ) , While adding Strategical plans we need to proof that if we miss tactical shot or not.

Third thing is If you think Infinite analysis is what i run to find the analysis then you are 100% wrong . I use my own analysis method using Persistent hash and that too Using Backward Analysis. It is not automated one Human interact is necessary. This method is used by Free style Players only in which GM Kosten is so good.
And finally what it seems to be the truth is not actually . I have played many positions i shared in my analysis except  what bucker said upto 6 piece endgame and white has no chance of getting into the game in most positions.
And last but not the least It is Weakness in chess if your opponent can exploit.

Markovich wrote on 03/25/10 at 17:55:50:

Are you unaware of all this things. Big GMs buy big hardwares to analyse. Ivanchuk has two 32 cores in his chess school. Peter svidler has recently bought a 32 way (nehalem EP) , Anand has a 64 way and has the biggest Database collection to date apart from the free style winner  Nelson Hernadez. Whatever a human does he cant be strong tactically. The only chance that a human can prevail is in endgame and that too ore than 6 pieces on board.

TonyRo wrote on 03/24/10 at 18:43:13:

From my college.

TonyRo wrote on 03/24/10 at 18:43:13:


I have access to a 20 core server, but my employers would be a tad upset if I ran Rybka on it.