< Earlier Kibitzing · PAGE 1 OF 2 ·
|Jun-11-10|| ||Phony Benoni: After 26.Rhe3:
click for larger view
Maybe the c5 pawn was indefensible in the long run, but why not at least try 26...Nd7? I suppose it was due to 27.Re6, and the g6 pawn is a goner. With 26...Rf8, Black can at least counter 27.Re6 with 27...Qf4.
Strange looking game. White's whole set-up seems awkward to the human eye, but it works out. Computers are changing high-level chess not by spewing out long volumes of analysis, but by indicating that there are principles to be considered which we haven't dreamed of or just reject outright.
|Jun-11-10|| ||Sneaky: Quite right Phony.
I've always considered that there might be something to those opening ideas that Steinitz experimented with, the ones that involved a lack of castling and the king rushing to the center in the opening (even with queens on the board!) The king bears down on a lot of central squares. If it wasn't for the concept of "king safety" it would be a perfectly crushing way to open a chess game.
Intuition tells us "don't do that--you'll get mated", and in the early days of chess computers they did exactly that, and sure enough, they got mated. So our remedy was to hardwire the machines to not play that way. That was our attempt at giving computers "a helping hand" so they could profit by human insight.
But maybe now that the student has surpassed the teacher, all of those helpful human heuristics are just holding them back?
|Jun-11-10|| ||Peter Nemenyi: You beat me to it. I was going to say that 14. Kc2 is a quintessential computer move--but one can imagine Steinitz playing it.|
|Jun-11-10|| ||Once: <Sneaky> <Phony Benoni> Excellent thought-provoking posts - thank you!|
It reminds me of a bit of management theory I was taught years ago. There are supposed to be four stages to acquiring a new skill or learning a new job:
1 - conscious incompetence. We are rubbish, and we know that we are.
2 - unconscious incompetence. We have gained a little knowledge, but we don't fully realise how much there is to learn. We think we have mastered the skill or job, but that's only our ignorance talking.
3 - conscious competence. We can do the job perfectly well, but we have to think about it.
4 - unconscious competence. We can do the job so well that it has become instinctive. We don't need to think about what we are doing - it just comes naturally.
In terms of chess, before computers I think we were in stage 2 - unconscious incompetence. We thought we knew it all, or at least most of what we needed to know. We pontificated general rules about king safety, good and bad openings, bad bishops, pawn structure, etc.
But then along came silicon, and we have to reappraise what we know. Some openings are not as bad (or as good) as we thought, some general principles have more exceptions than we had realised.
Perhaps we are beginning to move into stage 3?
|Jun-11-10|| ||Check It Out: It's amazing to watch a game between these "guys", knowing that they aren't going to make weak moves and commit tactical oversights that even super GMs make almost every game. So then it seems to come down to long term strategy, which I thought was the weaker element of computer engines. |
Rybka weakens black's queenside early on so his king has to go to the kingside, then walks his own king over to the queenside, proceeds to attacks black's kingside, and then cashes in with his queen rook pawn. That's a well executed strategy.
I had been thinking computer engines were killing chess for humans, but as Phony Benoni, Sneaky and Once seem to be saying, perhaps they will help us understand new heights that can be reached in the game. After all, when it comes down to it, games between two humans will always be the ultimate sporting event in chess, and if the super GMs are considering previously rejected concepts, due to advanced engines, it's possible that some of the best games in chess history lie before us.
|Jun-11-10|| ||Aniara: I learned a lot from the comments above. Some of the ideas reminded me of the following article by Kasparov, which most of you have probably read already: http://www.nybooks.com/articles/arc....|
|Jun-11-10|| ||Damianx: But how strong is Pandax has it ever beat Ryba can it beat Ryba?>|
|Jun-11-10|| ||OhioChessFan: I was curious what my Fritz10 thought of the game. To my surprise, he immediately went for 14. Kc2. 16. h4 has a real Alekhine look about it. Fritz thinks 18....Rfd8 is a blunder, at lower plies anyway, and I thought so when I first played through the game. I will let it sit on that position a while.|
|Jun-11-10|| ||mrsaturdaypants: Really interesting comments. The funny thing is, <Once>'s post reminds me of the analysis of human skill development worked out by the philosopher Hubert Dreyfus and his brother Stuart. And who is Hubert Dreyfus? He's the author of "What Computers Can't Do." And his work with his brother was titled "Mind over Machine." |
(It turns out that computers can, with incredible processing speeds and improved heuristics, do a great many things that Dreyfus didn't think they could do in the 1980s. But I think the Dreyfus brothers were nevertheless right about how humans learn.)
|Jun-11-10|| ||zanshin: Better pun would have been Fish and Chips ;-)|
|Jun-11-10|| ||zanshin: <So then it seems to come down to long term strategy, which I thought was the weaker element of computer engines.>|
Interesting thoughts by many on this topic, but I'd like to add a comment to that of <CIO>. A defintion of 'strategy' might be "a plan of action designed to achieve a particular goal." (http://en.wikipedia.org/wiki/Strategy) Using this definition, an engine does not use strategy, unlike a human player. Instead, from a given position, an engine generates candidate moves and evaluates the resulting positions. The process is repeated iteratively for each half move. I think the engine is looking for the next best move based on its evaluation function, and not to implement a strategy.
However, I believe a human player can retroactively examine the line generated and infer a strategy. CC players are also proficient at using engines to test their own strategies - and that's where human chess can improve (imho).
|Jun-11-10|| ||kevin86: The finish:was it a pin or a skewer?|
|Jun-11-10|| ||playground player: If a computer can't foresee getting nailed by a rook-kebab, how smart is it?|
|Jun-11-10|| ||zanshin: <kevin86> I think it's a skewer.|
<PGP> I'm pretty sure Pandix saw the skewer, but the position is lost and it probably evaluated a8Q as being even worse.
|Jun-11-10|| ||Once: <kevin86: The finish: was it a pin or a skewer?>|
I think it is both. For my money, it's a pin when you win or threaten the first piece in a line. A skewer is when you win or threaten the second piece. In this case, both the first and second attacked pieces are the same value as each other, so black gets to choose which one he loses. And I think that makes it a combination of both a pin and a skewer.
|Jun-11-10|| ||Check It Out: <Damianx> Here's a game where Pandix drew Rybka,Pandix vs Rybka, 2009|
and here is some information about Pandix (along with other engines participating in the 10th Int'l tournament the above game comes from)http://www.csvn.nl/index.php?option...
Obviously Rybka's using more advanced technology than Pandix, so not too bad, but Pandix is no fish!
|Jun-11-10|| ||lzromeu: <once> I agree. But for me, Rybka makes so many wrong moves. I don't understand why he (it) win. Scotish is a decadent openning, He forgot castling, beyond other blunders.|
|Jun-11-10|| ||Check It Out: <However, I believe a human player can retroactively examine the line generated and infer a strategy.>|
<zanshin>, lol, that's exactly what I did: attribute a long term strategy to Rybka, where there was only good move followed by good move.
Best move versus strategical planning - it's hard to reconcile how the moves can fit together harmoniously if there is no overall guiding plan.
So Rybka simply sees a little further and a little faster into the future than other engines and by doing so sets a flexible plan into motion, move by move, that it doesn't actually envision?
It's hard for me to believe that Rybka would typically beat a super GM, and do it without a plan. But then, I've seen the weak moves (according to engines) those same GM's make game after game, and can see how engines would be able to capitalize on them.
Therefore, consistently strong moves trump a long term strategic plan executed with some strong and some weak moves. That makes sense.
Sorry for that run on post, I'm still sorting out how engines, the complexities of chess, and human thought and planning all fit together.
|Jun-11-10|| ||Once: Can a computer think strategically?
I think the answer, at least at the moment, is probably no. But there are circumstances where a computer can appear to have a long term strategy in mind.
Let's start with a definition. I would define a strategy as a long-term plan which determines a number of moves. So a strategy might be, for example, to gang up on a weak enemy pawn, to exchange pieces when ahead in material or to push a queenside pawn majority.
As far as I am aware, most software engines evaluate each position afresh. They do not usually carry information from one position to the similar position that arises the next move on. In this respect, computers are like hyper-smart goldfish. They are incredibly intelligent, but can only remember something for a short while. That is why the poor goldfish thinks he is swimming in the ocean rather than a tiny glass bowl. It all looks new to him.
So computers don't think in terms of long term strategies.
But most computer engines are programmed to take positional and strategical considerations into account when evaluating a position. So they value central pawns more than wing pawns in the opening and middlegame. A centre pawn might be valued at 1.02 and a wing pawn as 0.98, for example. Using similar evalution tricks, engines can also be programmed to put rooks on open and half open files, to centralise pieces, find knight outposts, castle and so on.
This means that computers will play moves that are strategically sound. And this can develop into a consistent approach over several moves. An engine that is programmed to exploit weak pawns will keep banging away at a backward enemy pawn, as if it was pursuing a deliberate strategy. It doesn't consitute a plan, although it might appear to do so. Each move is calculated on its own merits.
Is that so different from the way that humans think? We might have a long term plan to blockade a passed pawn, but we would abandon that plan if the chance to win material or give mate presented itself.
|Jun-11-10|| ||chrisowen: Looks like a whistling dixie Rybka plays a deadpan game. The trick is taking into account rumbling flashy queenside holdings commodoring space. Then jacking the pawn triangle trumping any rough attempt of black's knight to bridge the gap. Pan dashes any hope nearing Rf8, Ne6 reflexing's white and black is picking up sticks. Crab it on now reversing black in your headlights Rybka five piece band hands a great whipping. O catch the ten individual black pieces or deliver mate. No need the 35th's queen take scotches his resistance tumbling a nice pin shake down in the end.|
|Jun-11-10|| ||zanshin: <but Pandix is no fish!>|
No, Rybka is the fish ;-)
<that's exactly what I did: attribute a long term strategy to Rybka, where there was only good move followed by good move.>
Because you think like a human <CIO>!
<Best move versus strategical planning - it's hard to reconcile how the moves can fit together harmoniously if there is no overall guiding plan.>
The only 'plan' of the engine is to seek the line yielding the highest numerical evaluation, asssuming best play from both sides. You can see the effect of lacking a strategy when you have quiet, non-tactical positions. Typically, an engine will start shuffling pieces (often Rooks) from one square to another. This is a dead giveaway that the engine 'does not know what to do'. But come to think of it, you could say the same for a human player!
<So Rybka simply sees a little further and a little faster into the future than other engines>
I think she sees more effectively rather than faster or deeper. It's really hard to get an idea of how fast she is evaluating or how deep the plies are running because Vas Rajlich (her programmer), is famous for trying to mislead competitors by under-reporting node counts and ply depths.
<consistently strong moves trump a long term strategic plan executed with some strong and some weak moves.>
I wouldn't set this is stone. I think the trick to playing an engine is to steer it into a type of position where engines are known to be weak (check my forum header for examples). However, I also think this is much easier said than done!
|Jun-11-10|| ||ajk68: My guess is that computers have developed enough tactical depth to easily play humans and disregard "king safety" and exploit the activity of the king.|
With computer versus computer, I'm not sure this would always work. Rybka gets away with it here because it is the stronger machine. My guess is that weaker computers that ignore king safety will get crushed by the stronger ones.
|Jun-11-10|| ||zanshin: <As far as I am aware, most software engines evaluate each position afresh. They do not usually carry information from one position to the similar position that arises the next move on.>|
<Once> Actually, they do. Numerical evaluations of positions are stored in the hash (or transposition) tables. This can save time because a position that has arisen previously, e.g. by transposition, does not have to be evaluated again.
More importantly, evaluations obtained further down a line can significantly affect analysis of an earlier move. In fact, that's the reason auto-analysis modes start at the end of a game, and move towards the start. The engine analyses after having seen the outcome of the game! This technique is also useful in analyzing a strategy. You can effectively 'convince' an engine that a move is better by showing its outcome.
Also, engines store promising moves in 'killer move' tables. The idea is that if a move yields a high evaluation in a given position, it should be considered a strong candidate in the next few moves. So, killer moves are often evaluated first because the sequence that moves are evaluated also has profound effects on the strength of play. After all, the engine does not know how much time it has before it is forced to move or stop analysis.
|Jun-11-10|| ||ajk68: <Once:> "As far as I am aware, most software engines evaluate each position afresh. They do not usually carry information from one position to the similar position that arises the next move on. In this respect, computers are like hyper-smart goldfish."|
I may be wrong, but I think Rybka may only flush branches of its tree once a contrary move is made. Even then, it seems to keep some calculations around.
Also, I've noticed when doing analysis with Rybka that is sometimes indicates a given line of play at a certain ply. If I force a contrary move that I think is superior, it will then evaluate that line. If I then back up, it often changes its assessment of the original position to the move I suggested. So it seems to keep some of the calculation around once they are made.
|Jun-12-10|| ||shach matov: Before we jump to conclusions and hail Rybka's decision to keep it's king in the center as the beginning of the revolutionary era of chess strategy, is it not possible that Rybka's king in the center represents some sort of a programmed complacence of a chess program which is tactically head and shoulders above most other programs? It almost looks like Rybka arrogantly challenges it's opponent by not castling, basically declaring that I am so tactically superior that I'll keep the king in the center and you can't punish me for that. In other words: was this really part of a brilliant strategic plan or just a display of the tactical superiority which enables it to make objectively inferior moves with impunity? It seems strategy as we understand it, is practically unattainable for programs.|
< Earlier Kibitzing · PAGE 1 OF 2 ·