Humans v Houdini chess engine (Elo 3300)

The Houdini chess engine has an Elo grade of 3300. Has a Grandmaster ever beaten it in any game? If so which grandmaster and when? 

I'd be kind of surprised if humans do better in correspondence chess against a computer than in a 'normal' timed (but not rapid/blitz) game, because it can spend every second calculating (unless restricted in some way) whereas a human has the matter of life to attend to.

I guess you're surprised then.

Michael Adams lost 5.5-0.5 to a computer quite recently. I know that Kramnik lost 2-0 to Fritz in 2006. Houdini is a lot stronger than Fritz was then -  Houdini's grade of 3300 is 500 points more than the grade of even a super-grandmaster. To beat a player 500 points higher is very rare. 

I wondered if anyone had any concrete evidence of a GM beating Houdini (or Stockfish or any of the other super-engines) recently in any game at any time control - maybe they could post the score. 

When they'll make the chess enginge that makes best moves possible?

Hmm, but does strategy matter much when you can see 64 moves ahead in every single line?  It seems to me that strategy is essentially a bunch of thinking shortcuts that humans use because we're so intellectually limited.

It's funny how many of you downplay/misunderstand the importance of positional thinking.

Think of it this way, computers, although they calculate truely massive amounts of lines unthinkably quickly, they too don't begin to scratch the surface of the game tree complexity.

Computers have gained tremendous strength this past decade due to improved algorithms that are better at taking into account positional elements.  It's a shortcut for us both.

Even at my meager rating, I can suggest moves to Houdini that it doesn't like at first, but 5-10 moves down the road it decides are better (I go back to that move and suddenly my move is preferred).  Obviously this is not the norm or I'd be much stronger, but to see it happen with any frequency brings you face to face with their limitations.

Computers can't be beaten in classical time controls because their tactics are so damn consistent.  (And of course they make tactical/positional mistakes too, otherwise they'd never beat eachother).  Humans can't match this kind of consistent quality OTB.

Local expert who is ~2400 USCF correspondence talks about how easy it is to beat the idiots who take Houdini or Rybkas top few moves and only ever play them.  Correspondence chess is a totally different matter.

The link thet IM Pfren  gives is from 2005 -

• that was the year that Hydra beat M Adams 5.5 - 0.5. 
• Kramnik didn't play correspondence against Fritz he played OTB for $500,000 - with an extra $500,000 if he won. I think he was quite well motivated! He lost 2-0.
• In neither of these did the human win

Things have moved on even from 2005 - the engines are better the humans are about the same.

Does anyone have evidence (rather than just opinion) of a GM beating a top engine (Elo > 3000)?

To my knowledge no such match (correspondence type) has ever been organized.  So if you're looking for a game then you wont find one.

If you mean classical time controls, then of course the human will lose.

Correspondence chess is not really typical chess, since as IM pfren pointed out, the humans can use computers to help them anyway so I don't see how you can tell if humans can beat engines in correspondence chess.

My original question concerned normal OTB chess with any of the normal time controls.

I agree with padman. 

Quote from wikipedia (under Houdini) -

When GM Peter Svidler was asked which one player he would choose to represent Earth in a hypothetical match against aliens, he answered "Houdini"

GM Svidler seems to agree too!

Okay.  Then, surely, being able to see 64 moves ahead in all lines with 100% accuracy, plus having some positional understanding, is superior to being able to see a handful of moves ahead in some lines with reasonable accuracy, and having some position understanding.

It seems to me that strategy is simply long-term tactics, and engines are wonderful at tactics, whether in the short term or the long term.

I don't think the claim "engines are stronger than humans" entails that a beginner using an engine will play just as well as a GM using an engine.

It could be that engines have a few key weaknesses, and GMs are really good at filling in the gaps here, but at the same time engines are so strong that, even with these weaknesses, they can beat any human player.

For once, pfren is well off target.

Humans have no chance against good software/good hardware combinations, at any time control.

All this talk that engines don't understand positional play is simply out of date nonsense.

p.s. I'm willing to back up my assertion.... how about this: I play you (pfren) two games of centaur chess, in which we both use computers. I'm a patzer compared to you OTB but suspect my rig (hardware/software) is superior. Bet you don't even come close to winning a game. Not even close.

Id like to see a human player take the top ranking on chess.com. 

 I dont have an opinion on the correspondence chess question.

However this is not correct. No computer today will look 64 moves ahead no matter how much time you give it.

Well, maybe I am being a little too doctrinaire by saying "no matter how much time". Ok, if the computer has like 100,000 years or so, then yes, perhaps it will look ahead that many moves. But I doubt it. I don't think 100,000 years will be enough time. Perhaps someone has a way to calculate how long Houdini would take, starting from the first move of a game, to look ahead at all possibilities for a 64 move horizon?

Bear in mind it also must store its computational results somewhere, since what is the best move in a 10 move horizon may or may not be the best in a 12 move horizon. And since we are talking about a 64 move horizon, that will be untold trillions and trillion and trillions of positions to store as it tries to tally up the best move. I cannot imagine the size of hard drive needed for this, but since the estimated number of possible games of chess is larger than the number of particles in the known universe, we may have to look outside our universe to find enough raw materials to build this hard drive.

Also, the hard drive, being so massive, might experience relativistic time dilation effects due to the effect mass has on time. This could further slow the calculation process down to the point where the protons making up the atoms of the hard drive would begin to decay.

In other words, a backup hard drive would be handy.

Computers are dumb. That  is how they became computers...

It's weird the way they always win though . . .

Most of you seem to massively overestimate the abilities of chess engines. You treat them like chess gods. In the ICCF, all of the good players play as centaurs, not just engines, and with good reason. Prawn, you say that you could be pfren in a centaur match. I highly doubt this. Centaurs > computers, and the greater player will be able to use the engine more effectively, and make know when and when not to trust the engines evaluations.

It's perfectly reasonable that a human could be a computer in very long time controls. The extra time benefits the human much more than the computer.

There was a human vs. computer correspondance match where Arno Nickel beat the Hydra supercomputer. 

Seems like there are two conversations going on here.

Nobody disputes that human+computer > computer (given more or less equal hardware+software). Arno Nickel beat Hydra with the help of a computer so it is no wonder he won that match.

What is disputed is whether a GM without the aid of a computer beats today's engines like Houdini. Common sense tells me this: Kasparov lost >15 years ago his match against Deepl Blue 2. Deep Blue 2 would be thrashed by today's engines. Ergo: GM's don't stand much of a chance against today's engines.

pfren comments that Kramnik would win hands down in such a match therefore sounds strange to me unless I'm missing something.