3300+ engine ratings may be justified ( NakaCyborg - Stockfish 1 - 3 )

I have always thought that yeah engines are stronger than humans but they mostly play among themselves and 3300+ engine ratings are the product of the narcissism of programmers.

But the outcome of yesterday's match between the 5th strongest player in the world,  Hikaru Nakamura, known for his anti-engine practice and skills, partly aided by the assistance of a 2011 version of Rybka ( for blunder check) and pawn odds when not aided by Rybka, and a free engine, Stockfish, running on a commercial computer stripped of its opening book and tablebases has shaken my confidence.

Now I tend to think that 3300+ engine ratings are not far fetched or lofty, on the contrary: they reflect the said reality that engines have become that much stronger than humans.

This setup was really good because it addressed almost all of the complaints that can be raised against engines, namely:

a) humans have a good positional understanding, but they make one move blunders

Since Hikaru was aided by Rybka, this possibility is ruled out.

b) engines have access to opening books and tablebases, which is an unfair advantage

In the match Stockfish stripped of its bases had to invent opening theory created by humans in hundreds of years during a 45 min match and it did so in an excellent way.

I also have to add that being up a pawn on the super grandmaster level is just an enormous advantage. The fact that even a pawn advantage is not enough to compensate for the skill difference is striking and just underlines the correctness of 3300+ engine ratings.

I have to note that I have only watched the first game live and I have only anecdotal evidence of the result. I could not find any official statement or access the actual moves of the games.

Does anybody know what is an estimated rating of IBM 97?

You mean Deep Blue? The Wikipedia article on Deep Blue says that no rating was established for Deep Blue, but this statement lacks quotation.

However what is known is that Deep Blue examined two hundred million positions per second, which is a high speed ( 200 Mnodes / s ) even by today's standards ( at the unofficial engine world championship at TCEC on high end commercial computers cca. 50 Mnodes / s is reached ).

Dave?  What are you doing, Dave?

We may as well fondly remember the days of being more capable at chess than computers or go find some retro ones we can beat.

The disadvantage for Nakamura was the time control. 45 min. + 30 sec is way too fast even with pawn advantage. Nevertheless, I'm surprised that he lost with Rybka assistance...

 No pawns odd in the first two games.

After 2 games Stockfish is winning 1.5-0.5.

Stockfish won the second game because Nakamura refused to draw. Had it been an official match, he would have conceded the draw in the second game. Stockfish didn't do any better than Nakamura, in fact, it wanted a draw in the second game.

In the first two games, Nakamura used Rybka 3 from 2008 running on a 2008

MacBook.

In the third and fourth games, Hikaru will play unassisted, with the white

pieces, against Stockfish with a random black pawn removed from the board.

http://www.chess.com/forum/view/chess-buzz/where-are-the-nakamura-vs-stockfish-games

It is an interesting argument that Naka got bored in the second game, pressed too hard and lost. In fact I also got bored and went to sleep watching this game ( because it was getting late night in my country ), but I'm low rated. If Carlsen deserves his high rating for playing drawn endgames to the death waiting for his opponent to 'get bored' and make a mistake, then Stockfish also deserves his 3300+ rating for playing a drawn position to the death. This is how ratings are defined.

The argument is about whether engines deserve their high rating and this match was proof that they do.

If the question is whether a team of humans have the collective knowledge to be on par with Stockfish when they play a correspondence game this is a different story ( but even there I have concerns about giving Stockfish 3 days thinking time for a move ... ).

I'm not sure if those are the algorythms that have become so strong ( since basically they are the same: tree search with alphabeta algorythm ). It is a suspect that simply sheer computing power has become so strong ( this was the point Daniel Rensch was making all along the broadcast ). On the other hand it is true that if you can cut agressively branches of the search tree you can go deeper and depth is life and death in chess. The success of Stockfish is due to its deep search and the secret must be in how exactly the pruning is done. So it is both the improvement in computing power and pruning strategies that made possible for engines to reach this strength.

Game 2 was a loss.

...but now we know the narcissism was actually yours ;).  Humans always want to romanticize their capabilties and pretend like it's a poignant struggle of man vs. machine.  This is the equivalent of getting upset about the fact that a can opener opens a can better than a man with a flat rock...

I hear the computer was actually teased at the hairdressers for nearly losing to a human.  Told it should get more silicon implants.  Humiliating.  Poor thing is in therapy.

btickler is right, computers and chess engines are not an alien race that we must compete with, they are simply tools that we humans have created and we should be proud of this technological achievement. 

notting new here in this article

and beside we all know the word 99% in chess is tacties 

Wait, what?  Maybe the way you play, bro.  Use good positional understanding and better tactics magically appear.  Good thinking makes better positions.  Even Tal couldnt summon a win out of a locked down position.