Fritz kills my joy of playing

Originally posted by diskamyl
I think opening theory could serve as a good reference point for comparison, I know there's no necessity or scientific laws yet, but it seems to me that a large portion of theory is close to perfection.

I don't think we have any indication of that. -it could easily turn out everything we now think we 'know' about opening theory is wrong. to pick a single random example, it might turn out that for some oscure geometrical reason 1.a3 lines hold all forced wins there are, and everything else loses. but as we've covered only the very beginning of any lines in any kind of rigour, we don't really know what lies beyond. we're just assuming things will continue working like they work in the first few moves, but we have absolutely no guarantee they will. it might turn out that everything we ever thought we knew about chess changes completely beyond some specific move.

I'm not saying that's a probable scenario, but it is possible, and we simply have no grounds to rule it out. everything beyond endgame tablebases might be wrong.

if you think of how huge difference it makes to look at different parts of a mandelbrot fractal, it's not a long stretch to think something similar might happen within the hypothetical solved multidimensional tree of possible lines. big areas might be simply barren (no wins nor draws), some areas won, and yet some others weaving an incredibly complex and beautiful pattern of wins, draws & losses intertwined. oh, and then there is the tree of illegal positions which can't be reached at all, sort of like a black void surrounding all legal lines.





you can probably tell I haven't slept much last night. 🙂

Originally posted by wormwood
I'm not saying that's a probable scenario, but it is possible.

if you think of how huge difference it makes to look at different parts of a mandelbrot fractal, it's not a long stretch to think something similar might happen within the hypothetical solved multidimensional tree of possible lines. big areas might be simply barren (no wins nor draws), ...[text shortened]... s weaving an incredibly complex and beautiful pattern of wins, draws & losses intertwined.

I'm not saying that's a probable scenario, but it is possible.

philosophically I know I can't argue with that, but anywhere outside the boundaries of philosophy (for example, a forum chat 🙂 ), I would call it impossible. For example, I will never take Berliner as a serious chess writer because he claims 1.d4 is a forced win for white. That's not because I believe I understand d4 openings better than him, but it's an educated speculation that I think I have enough resources to believe.

I don't see how big areas might be barren when we include the 50 move rule. sides would have to find a way to make non-losing pawn moves forever. maybe I'm thinking wrong?

Originally posted by diskamyl
[b]I'm not saying that's a probable scenario, but it is possible.

philosophically I know I can't argue with that, but anywhere outside the boundaries of philosophy (for example, a forum chat 🙂 ), I would call it impossible. For example, I will never take Berliner as a serious chess writer because he claims 1.d4 is a forced win for white. Tha would have to find a way to make non-losing pawn moves forever. maybe I'm thinking wrong?[/b]

yeah, categorical statements like the one about 1.d4 are obviously unfounded for the same reasons as all others. we simply don't know.

the 50-move draw lines are only a small subgroup of all possible (forced) draws. the number of different lines (10^120) exceeds the number of atoms in the known universe (10^80) with a mindboggling margin, and we have covered maybe hundreds of millions (10^8) positions during the existence of chess, including all patzer games. the number of fritzed positions isn't greater in any meaningful way.

-you could say that practically all of possible chess positions have still never been evaluated neither by a man nor a machine. making categorical claims based on that microscopic amount of known positions is the chess equivalent of thinking the sun orbits the earth. we have no idea whether there are galaxies and galaxy superclusters of legal chess positions out there, with vast empty space in between.

we don't know anything about the unknown search space, but for typical psychological reasons assume to be exactly like the 'known' space.



edit: I think I might've just killed fritz's joy of playing! 😵

I'm interested in the idea of different starting positions being an answer to chess being solved. It seems to me that most positions reached from a new starting position would be already covered if chess were solved. The only complicating factor is that castling might be possible in positions where it wouldn't be normally.

suppose the "perfect computer" is created that completely "solves" chess.

would it really change anything? -- unless someone could implant a computer into their own brains, he/she would still have to rely on their own thinking skills (and emotions) while playing a game.

I know what you mean, you play a game of chess that you thought was great and then Fritz rips it apart. In saying that tho, when fritz gives one of your moves '!!' That's lap of honor round the bedroom time 😀

just finished a game that was ideal for fritz: open board with few pawns, both kings exposed, rooks and queen on open files & diagonals, and advanced posted knights.

fed it to fritz and it showed me the light of day. (i missed so much,)

Wormwood's analogy to fractals is a good one. Google 'Mandelbrot Fractal' and watch some videos of them as they zoom in closer exposing much more weird, intricate, and beautiful patterns. It's not a stretch to believe that if using God's computer we could solve chess and visually map every possible line, that winning and losing chess lines would be as intricately woven through draws and illegal positions, and as we 'zoom' closer to the endgame the patterns would be extremely complex and just as unpredictable.

Bringing it back to good ol' Fritz. Yeah, chess engines are at the point we're they're on par with the top Super-GMs and possibly better, but they certainly are not the end-all of chess. In fact, engines have beat us at our own game by playing it the same way as humans -- with more emphasis on high-level pattern recognition and selective bias over brute-force calculation. In essence, they're mirroring us and just because they're better at it than us right now, that doesn't mean they are the *best*.

Chess theory is all hand-waving and very little proof (other than it works), and it's not a stretch to believe there are chess singularities out there that may come along and make us completely change the way we play a game of chess. It's not a stretch in possibility to wonder if in a few decades from now, super-GMs and engines will have levels of theory and endgame tablebase analysis that would make current GMs and engines look like patzers.

In short, it's easy to think because someone knows more than you, they know everything, but we are still along way from any kind of proof that top super-GMs and computers "know" everything in chess. Nothing reminds us more of that than actual mathematically proven chess positions, which are only table-base endgames. For example, there's a known table-base seven-man win in 517 moves, and it's interesting to walk through the moves. To current chess analysis, many if not most of these moves are completely incoherent as to 'why' they are winning or losing moves, and 'progress' for humans and engines right now would be to recognize what lines are winning, let alone *how* they are winning.

Finally, I'd be leery of how you use Fritz to analyze your games. Personally I don't bother with Fritz and I'm happy to use a much weaker engine. The only use I have use for engines is to see what tactics I may have missed.

A bit too long of a rant here. 😛 It's an interesting subject though.

Fritz 11 is great, kills me all the time.

For a real 'show' though you have to see Deep Rybka 3 - it's unreal, wins 90% of the time vs Fritz, even using the same book the fact it uses all 4 of my CPU's agains Fritz's 1 cpu means it is a complete stomp (quad 3.25ghz + 8gb ram).



I like playing against them because it teaches me to play perfect openings and punishes me when i get them wrong.

I also like the fact with fritz you can tone it right down so it will make small tactical errors, like (some) humans, capitalise on them and you can win

Having said that i still havn't beat the damn thing lol

If anyone ever fancies taking on deep rybka 3 let me know 😀 i will play the moves for it, will need to be done using another site obviously (or could it be played on here as an unrated game?? obviously i wouldnt use a computer chess program normally, but if it's stated up front and unrated surely thats ok. maybe get a bunch of people to colude together to try and beat rybka?).

Will

Originally posted by 3ncrypt
Fritz 11 is great, kills me all the time.

For a real 'show' though you have to see Deep Rybka 3 - it's unreal, wins 90% of the time vs Fritz, even using the same book the fact it uses all 4 of my CPU's agains Fritz's 1 cpu means it is a complete stomp (quad 3.25ghz + 8gb ram).



I like playing against them because it teaches me to play perfect openin ...[text shortened]... thats ok. maybe get a bunch of people to colude together to try and beat rybka?).

Will

This could be the next forum vs. ?

I doubt the forum has a chance though.

Originally posted by ivan2908
This is a 3 minute RHP blitz. I thought my roc sac was sound. Everytime I am happy with my game, Fritz has to find 500 blunders in it. :'(

I just wanted to chime in this thread because I know exactly how ivan feels when you think you played pretty well and then the computer shows that those moves you thought were brilliant were blunders. Same thing happened in the "swindle" I just recently posted. My defensive moves were not rated well with fritz and one led to a direct mate. In may be one case when you make a blunder that leads to forced mate, but other moves you think are good that fritz thinks are bad, still may have been good for that particular game, at that particular time, against that particular player. That's what computers don't take into consideration, and really, that's fine. You shouldn't let that kill your joy of playing, but use it as a tool. If your outcome was the desired effect (a win, or draw in a bad position) then you did your job for that game. Sure, you need to anaylze to see the mistakes and in-accuracies, but that analysis should help to make you stronger, not bring your morale down. Nobody is going to play a perfect game, but you play for "that moment," and that's what matters.

As I stated in my thread, even though your game looks cut and dry in a Fritz analysis, during the game there are many factors and circumstances going on and even sub-par moves can have positive effects, that's what happens when humans play, and in the end, that's all that matters when we play chess. It's man vs man (or woman!) and all about the situation at the moment of that game. Fritz, computers, preparation, are all tools that help to strengthen the player. It's good that Fritz can give assessments to show that seemingly good moves are bad, because then it makes us stronger. You need to lift heavier weights to become stronger, and you always need to be humbled in order to stand tall.

Hope that makes sense and helps a bit!

This quote may help as well (I just read it on some site browsing about chess) I think it pertains to the topic...

"Play, analyze, study tactics and you will be a master. Study openings and you will be an erudite. It is easy to think that you know a lot about chess by reciting books. But remember that when a chessboard is between you and your opponents, THEN only sweat and blood will rule."

I make fritz play white with the "Douchebag Defense" :

1. F3 - Any move
2. King F2


And it still crushes me at 5 seconds/move.

Originally posted by Sam The Sham
I make fritz play white with the "Douchebag Defense" :

1. F3 - Any move
2. King F2


And it still crushes me at 5 seconds/move.

try the "Worse than douchebag defense":

1. f3 - Any move
2. King f2 - Any move
3. King e3