Rybka Chess Community Forum
What will the match score be (remember draws are wins for Rybka)? (Closed)
|Meyer by 4-0
|Meyer by 3-1
|Tie Match 2-2
|Rybka by 3-1
|Rybka by 4-0
Speaking of openings where one side is ahead in material, I'd like to see a match where Rybka takes the white side of the Blackmar-Diemer Gambit against a strong human player, to see if that gambit can be rehabilitated. I've tried to rehabilitate it myself but just can't seem to find the way. That might be a way to make these Rybka vs human matches interesting, trying to see if Rybka can rehabilitate dubious gambits.
Well, the problem here is that if GMs can't win a match with Rybka when they get an extra pawn for no compensation, how could they have any chance when there is substantial (if insufficient) compensation? What would happen would be that game after game, White would win, but the computer would show that Black should have had the advantage. Maybe if one of the top five in the world played this could have some interest, but they want too much money.
Maybe there are some really unsound gambits, which are still interesting or have some sort of historical significance.
Another idea along the same lines would be playing dubious opening variations which aren't gambits.
Yes, we could do this, especially against 2700 level GMs who might be a bit strong for pawn odds. Some examples, with a recent Rybka version eval: 1e4 d5 2exd5 c6? (+0.61); 1e4 e5 2Nf3 d6 3d4 Bg4? 4dxe5 Nd7!? (0.59); 1e4 e5 2Nf3 Nc6 3Nc3 Nf6 4Nxe5?! (-0.62); 1d4 Nf6 2g4? (-0.47); 1d4 Nf6 2e4? (-0.38). Or for a non-gambit sometimes played to insult one's opponent or just to show off, 1e4 f6? 2d4 Kf7? (+0.94). We could even do an eight game match with eight different bad openings like these. What the gambits all have in common is that they obtain some small but clearly insufficient compensation for the material lost.
If the match is in October, then 1e4 e5 2Nf3 Nc6 3Nc3 Nf6 4Nxe5?! could be played on Halloween. :-)
> If the match is in October, then 1e4 e5 2Nf3 Nc6 3Nc3 Nf6 4Nxe5?! could be played on Halloween. :-)
Reading the text of your quip, I immediately thought of the Frankenstein-Dracula variation, but when I looked it up, the moves didn't seem to match up, and I didn't see how to transpose them either. Is there another monster on your mind?
1. e4 e5
2. Nc3 Nf6
Alternatively 1. e4 e5 2. Bc4 Nf6 3. Nc3 is also possible
4.Nxe4 d5 is considered to give Black no problems; 4.Qh5 threatens Qxf7#, a threat which White continues to renew in this line
5.Qxe5+ is possible if White wants to avoid complications
It's the Halloween Gambit. :-p
I have never heard of the Halloween gambit. By the way, I really like the idea of "odds" games where the odds for the human opponent come in the form of Rybka playing a dubious or out of date opening. Of course the human might also need a time advantage over the computer as well. I think that some matches that bring out either the "Golden Oldies" of chess openings that are not currently seen as competitive or else are gambit-themed would be very entertaining and would draw a lot of positive attention.
One other question: would the human GM be told ahead of time which gambits were on the plate, or should he (she?) just be happy to know that Rybka will be flinging pawns across the table?
If we ever do such a match, it must be fair to both sides. That means that either the list of starting positions (openings) is known to both sides in advance, or else that they be selected at random from some larger list that is known to both sides in advance.
People who have more respect for it call it the Halloween Attack :-)
I think you are right about Qxd5 and many other moves, although in this case Rybka's reluctance to trade when behind is just too mild. I have proven with engine vs. engine matches that deeper search does generally help even when giving handicaps, but there may well come a point where it starts to hurt. It is definitely better to outsearch your opponent by 3 plies instead of 2, even when losing, but maybe outsearching him by 9 plies instead of 8 (maybe roughly like in this match) could actually hurt on balance in a knight odds match.
Rybka played in many games lately with very large pieces back, a bit on the defensive, does not take the initiative and let him attack the contrary. In addition, changes or fails to change the large pieces, with ease, losing potentiality to overcome. As for tactics and strategy to attack the king, but lacks knowledge.
Rybka, but tops the lists of engines, but does not know that the GM., as have other concepts to learn more about how memory plays an engine and attack their weaknesses.
To see the results of the poll are very interesting (after the match :-))! Now we all knew, why Rybka hadn´t any chance; but the result 4-0 for FM Meyer was fewest expected result! Are we all blind and dense because of the previous results in odds matches Rybka played? I think so!
None the less, FM Meyer´s playing style shown in this match was smart. Calmly and easily he played his moves! The prize money of $100 is well-earned. Maybe you can give him additionally a free copy of Rybka 3 for his great show?!
In my own case I was more influenced by the results of quick/blitz knight odds matches. I knew that longer time limits favor the human, but I underestimated the degree to whcih this was so. We'll probably have a rematch before too long at more reasonable odds, probably pawn and three moves (no f7 pawn, White gets e4 and d4 and still moves first). I don't think John even has a home computer, and probably can't put chess software on his office computer.
That's interesting that John doesn't have a home computer. It reminds me of how Kramnik prepared for the WC match he won vs. Kasparov--he specifically avoided using computers.
Larry and Vas: thanks for doing this match. I was quite interested to see this result. I look forward to future odds matches as I enjoy playing a computer in that way and am interested to see what types of algorithms evolve to deal with the handicap situation. This is all part of the approach to allowing lower rated players such as myself to have a meaningful experience using a computer chess program.
Where did you get that myth from - that Kramnik avoided using computers?? This would certainly have been madness. I saw a video report about the match. You could see Kramnik's seconds (Lautier, Bareev, Illescas) with their laptops, engines switched on virtually all day long.
Maybe what you wanted to say that back then, he avoided using computers during his games. :)
(Disclaimer: That last sentence was just a joke. I don't believe any of the nonsense spread by Danailov and Topalov.)
I seem to remember an interview in which Kramnik stated that he thought Kasparov used a computer too much; so Kramnik took the opposite approach, to work less with the computer and think more on his own.
Here are some comments I found that seem similar, but I thought I read (probably in 2000/2001) a statement by him about Kasparov, not just a general statement:http://www.chessbase.com/newsdetail.asp?newsid=1912
How important are the computer chess programs in the preparation and during a match?
Kramnik: On the one hand it is very important to use the computer in your team for getting a certain basic level in your preparation. On the other hand it is not good to use it all the time. In a way you have to become independent from it so as not to affect your own play. Let me put it like this: my trainers are working with the computer permanently – I do not.
It was often said that his decision to play the Berlin in the match was inspired by the thought that computers are less useful in the resulting positions - Kramnik didn't want to be 'outprepared' by Kasparov's engine based analysis. I'm not sure whether this is a good explanation, though. Kasparov was a fantastic attacker and most likely a better tactician than Kramnik. The queenless middlegame or endgame positions in the Berlin just suited his style better.
> The prize money of $100 is well-earned. Maybe you can give him additionally a free copy of Rybka 3 for his great show?!
Actually, he's going to have to collect his prize from CSS :)
I think that handicapped matches are interesting to some ... but I am finding it hard to understand a lot of the criticism here for an engine not knowing how to win a lost game from move 1. I mean would people think differently had the human blundered badly in every game and lost all the games? The point is that it has nothing to do with Rybka's play and has everything to do with the human play. Trying to make Rybka play "weaker" moves because the human might have more chances to blunder seems silly to say the least and would probably backfire ... Rybka's strength is to play as strong as possible all the time.
Anyway I think there is way too much time being spent on handicapped play of Rybka ... I mean it is interesting and might give some people some fun in feeling they have "won" a game against Rybka ... but really there are much more important things that need improvement in Rybka and I don't think handicapped play improves the overall strength of Rybka. Already there are a lot of more chess like options for handicaps to choose from ... most reasonable seems to play a lobotomized Rybka with a fixed depth. I wish more work would be spent on improving endgame play of Rybka by introducing endgame modules that would use information of bitbases to narrow its search. Endgames are still an area where engines can improve greatly.
As I was one of the 4 (5%) who went for a 4-0 to the human do we win any prizes hahahahhaha ... We (3) have played loads of these Knight odds games and the machine does really poor ... Maybe they are a waste of time or maybe not ... My son with a moddest ELO of 1800 beat Rybka 3-0 with 15 mins each on the clock
Shouldn't be a short notice on http://www.rybkachess.com/
in Section [Man vs Machine] about the knight-odd-match between Rybka and FM Meyer?
I think your handicap matches are very interesting, also this one.
And the task was too difficult this time, even for Rybka.
But it shouldn't look like:"We present the nice results of Rybka on the rybkachess-page, and we 'hide' the other events in the Threads of the forum."
Shouldn't be, will there be a short note about Meyer-Rybka 4:0 on ://www.rybkachess.com/ soon?
Yes, of course. This always needs a couple of days.
I think a couple of days have now passed...
Sure, but why do today what can just as easily be done tomorrow? :)
Damn... using my own life-philosophy against me.
For the record, the 3-0 playoff victory at pawn and move (f7) odds over GM Dzindzichashvili was left off of rybkachess.com as he was quite ill on the playoff day; some judgment calls are needed as to what should be included.
aaaah, yes, shouldn't be a link to the blitz chess tournaments of the months june and may on the website of my chess club? shouldn't be.... :)
So much work :)
> So much work
Your odd-Matches are very interesting, and they are observed with rather much interest.
I thought, it would look 'funny', if you only present those results which are complimentary for Rybka.
Such a good Product als Rybka shurely isn't in need of such a strategy.
'It needs a couple of days' shurely ist OK.
And now there ist the messages.
Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill