|
On July 20 2006 05:54 Reason wrote:Show nested quote +On July 20 2006 01:51 a-game wrote:
yea but my point is who cares if computers can do things we cant, i don't think any humans play chess with 'beating computers' as a goal in their mind, just like runners don't aim to beat cheetahs, weightlifters dont aim to beat elephants, swimmers dont aim to beat dolphins, jumpers dont aim to beat kangaroos, etc etc. Yes but maybe your missing FA's point ?? (Or I misinterpreted it  ) Regardless I share that sentiment, if it is indeed shared. Cheetah's and many other animals can run faster than humans, always have done and always will. There's no point in getting upset about that. Elephants and many other animals are stronger than humans, always have been and always will be. There's also no point in getting upset about that. Chess is a game of intelligence, of using pure thought to defeat your opponent. Nothing in this universe, to our knowledge, ever has/will rival our intelligence, we are unique. Nothing is smarter than we are. Oh shit, something is, a computer. And we made it. DOH! Edit: Maybe that should be encouraging however, as the only thing better than us was made by us 
You cant say a computer is "smarter" than us. Computers do not think. They process information. Alot of information.
|
Is there anywhere i can play on free server/chess software?
|
On July 21 2006 10:32 decafchicken wrote:
Is there anywhere i can play on free server/chess software?
Yahoo games has a brand new interface and many people to play with.
|
I used to play Chess alot growing up...
Hmm I guess I'm kinda rusty now, if you want a game over yahoo just PM me over this. Give me a time, etc.
|
On July 18 2006 18:19 FrozenArbiter wrote:
Hey, I have a question to all of the chess players here; isn't it discouraging to know that a computer can play better than the best humans, than the best you could possibly play? It seems very discouraging to me somehow.
If you consider that a computer using raw calculation CANNOT beat you when you reach decent strength(1600 elo); it's not discouraging. Computers get to reference opening databases so they play the beginning of the game almost perfectly, then they have all type of programming to assess middlegame themes and this is where they shine. Though they still SUCK in closed positions. Lastly, they have tablebases which are huge files where simpler endgames have been calculated to infinity so that the computer plays them perfectly. Honestly, playing a computer is really about working on staying even tactically while positionally outmanouvering the computer.
Hawkeye WAS the most accomplished 1 minute player in the world. He was ranked like 166 in the world during that time. SMALLVILLE(icc handle), aka Hikaru Nakamura, has set damn near every ratings record on ICC, Nakamura(born in 1987) is currently ranked 4th in the USA and 80th in the World. His FIDE elo is 2632 while Vasily Topalov recently acheived an elo of 2813. These players gain appx 12-18 elo pts per win vs a like-rated opponent.
Here is a game played by Hikaru Nakamura versus a computer rated at blitz 3500+ (naka is 3300).
The time control is game in 3 minutes with 1 second added after each move. Nakamura finished the game with over 2:30minutes on his clock.
[Event "ICC 3 1"]
[Site "Internet Chess Club"]
[Date "2006.07.16"]
[Round "-"]
[White "ChessThinkerX"]
[Black "Smallville"]
[Result "0-1"]
[ICCResult "White resigns"]
[WhiteElo "3566"]
[BlackElo "3374"]
[Opening "Queen's pawn"]
[ECO "A40"]
[NIC "QO.17"]
[Time "17:09:22"]
[TimeControl "180+1"]
1. d4 e6 2. c4 g6 3. Nf3 Bg7 4. Nc3 d6 5. Bg5 Ne7 6. e3 h6 7. Bxe7 Qxe7 8.
Bd3 O-O 9. O-O c6 10. Qb3 e5 11. h3 Re8 12. Rfe1 Kh7 13. Qa3 Nd7 14. Ne4 Bf8
15. Ned2 f5 16. e4 f4 17. d5 c5 18. Nb1 g5 19. Nc3 h5 20. Nh2 Nf6 21. Bc2
Bh6 22. Ba4 Rg8 23. Bd1 g4 24. hxg4 hxg4 25. Nf1 Bd7 26. Ba4 Kg6 27. Bxd7
Qxd7 28. Rad1 Rh8 29. Ne2 Bg5 30. Nfg3 fxg3 31. Nxg3 Bf4 32. Kf1 Bxg3 33.
Qxg3 Rh1+ 34. Ke2 Nxe4 35. Qe3 Rxe1+ 36. Rxe1 Ng5 37. Rh1 Qf5 38. Qd3 Qxd3+
39. Kxd3 Rf8 40. Ke3 Nf7 41. Rh4 Nh6 42. Ke2 Kg5 43. Rh1 Rh8 44. a3 Nf7 45.
Rxh8 Nxh8 46. Ke3 Ng6 47. b4 b6 48. Ke4 Kf6 49. b5 Nf4 50. g3 Nh3 51. Ke3
Kf5 52. a4 e4 53. Ke2 Ke5 54. Ke3 Nxf2 55. Kxf2 Kd4 56. Ke2 Kxc4 57. Ke3
Kxd5 58. a5 {White resigns} 0-1
|
These players gain appx 12-18 elo pts per win vs a like-rated opponent.
No way, not if you're talking about elo and not the ICC rating. There is no way that elo can fluctuate nearly that much for a single game.
If you consider that a computer using raw calculation CANNOT beat you when you reach decent strength(1600 elo); it's not discouraging. Computers get to reference opening databases so they play the beginning of the game almost perfectly, then they have all type of programming to assess middlegame themes and this is where they shine. Though they still SUCK in closed positions. Lastly, they have tablebases which are huge files where simpler endgames have been calculated to infinity so that the computer plays them perfectly.
Almost all computers use raw calculation; while all programs have some weight placed on certain values (ie: isolated pawn, doubled pawns, control of the center, sacrificing pawns, etc), it's almost entirely based on a move-by-move computation of the position. And let's keep in mind what we mean by a computer sucking at closed positions; top of the line computers will beat anyone not a GM and some of the GMs as well.
I personally don't really care that a computer can beat a human. It's just like a car being able to outrace an athlete; it's something that we can't really compete with.
|
You're mistaken. Yes elo will fluctuate 12 points from beating an opponent with the same rating. No, computers don't use almost entirely raw calculation. Lastly, I can beat Fritz (elo 2600) in a number of closed positions.
Don't state your ideas as fact.
|
Yes elo will fluctuate 12 points from beating an opponent with the same rating.
According to: http://www.fide.com/ratings/calculator_rtd.phtml , the change will only be 5 (I used Rc:2600, W-1, N-1, Ro-2600, K - 10 as settings for one and Rc:2800, W-1, N-1, Ro-2800, K - 10 for the other). Maybe it's wrong or maybe not, but I'll take its word on it.
Lastly, I can beat Fritz (elo 2600) in a number of closed positions.
Care to share some games with Fritz 9 in which you do so (regular time controls, opening book and endgame tablebases included)? I'm sure that Kramnik would love your help in preparing for his match. It's not so easy to beat a toplevel computer...
No, computers don't use almost entirely raw calculation.
Why do you think this? AFAIK, computers calculate variations, with weight given to certain factors in the position and tactics. However, most of their candidate moves will be based on calculating a line, then thinking that at the end, that line will give the computer the most advantage or decrease the opponent's advantage the most (giving it the horizon effect, where the computer can't properly evaluate the resulting position because it lies outside its calculating ability). Sure, a program can put priority on different aspects of the position. But it doesn't have... say, the strategical depth that a human does (it doesn't play 'with a plan').
|
Deep Blue 200 Million moves a second. Kasparov 3 moves a second.
Maybe technically true, however.
Every human has the ablility to throw out hundreds of millions of bad lines without even thinking about it. Humans just analyse a handful of canidate moves (GMs can spot these candiate moves very quickly) that help their plan. After some analysis on each canidate move, the make the move that helps them most - or the one that helps their plan while also hindering the opponents plan to a degree. Usally, they pick the double-edged one that helps them, and hinders the opponent.
|
FIDE uses a different elo formula than USCF and I didn't realize the K was so low for 2400+.
You're right about the rating changes being less for those players. For the majority of us elo points will change more significantly with wins or losses.
I cannot beat Fritz from start to finish with opening book on because it will avoid a number of anti-computer lines. But, my point was that there are a number of closed positions that I as a only fairly decent tournament player can beat the computer in. I beat Fritz in the position in the Nakamura game above when Fritz's eval was in half a pawn in it's own favor.
I guess I didn't quite know what you were getting at with saying 'raw calculation.' You have to program more and more components into how the computer assesses a position to make it stronger, it has much less to do with pure brute calculation in determining how strong the computer is. So in that sense I don't consider it raw calculation. Furthermore, they use their opening books and tablebases. There are a number of lines that you can get Fritz or any other engine to slip into when they are without any opening book and these are just simple technical wins.
|
Didnt Kasparov have a rematch and eventually beat the computer at least in 1 game? And while we are on the subject, who do you guys think was the greatest chess player ever?
|
On July 22 2006 23:29 Slaughter)BiO wrote:
Didnt Kasparov have a rematch and eventually beat the computer at least in 1 game? And while we are on the subject, who do you guys think was the greatest chess player ever?
For me the greatest player are :
1. Alekhine
2. Fischer
3. Kasparov
|
1. Kasparov
2. Fischer
3. Morphy
|
intrigue
Washington, D.C9934 Posts
i don't play chess and i don't pretend to understand it (but i'm interested in it =P) , and i want to ask a question to you guys that do: top starcraft pros are distinguished by their style, superior micro/macro/timing sense, etc. - do GMs and the top players of chess also have flairs and different strengths? can they win by mind games, make up 'builds,' go in with certain strategies, or is each game determined by pure stamina, memorized moves, intuition and analytical power? if given the transcript (is this what 'replays' of chess are called?) of a match, can you recognize instantly who is playing? for example one can tell ipx apart by his macro habits and godly muta control, and in_dove by his sick skterran.
in short, what makes a rank #1 player better than another GM if both have enormous amounts of practice, theory, and experience?
|
i don't play chess and i don't pretend to understand it (but i'm interested in it =P) , and i want to ask a question to you guys that do: top starcraft pros are distinguished by their style, superior micro/macro/timing sense, etc. - do GMs and the top players of chess also have flairs and different strengths? can they win by mind games, make up 'builds,' go in with certain strategies, or is each game determined by pure stamina, memorized moves, intuition and analytical power? if given the transcript (is this what 'replays' of chess are called?) of a match, can you recognize instantly who is playing? for example one can tell ipx apart by his macro habits and godly muta control, and in_dove by his sick skterran.
Usually, unless it's someone that has a distinctive style like Karpov, Kasparov, Tal, Capablanca, Fischer or Morphy, it isn't that easy to tell the difference. They have distinctive style that they have made their 'own'. Nowadays, you have to be a universal player (good at attack, defense, tactics, strategy and endgame technique) to succeed at a top level, which means that no one really has a distinct style of their own. The games themselves are a mixture of stamina (6 hour games can really, really tire you out physically as well as mentally- top chess players have to workout to prepare for a match), experience, intuition, analyzing and opening theory. Each player has his own little 'criteria' for subtly evaluating the position in one way or another, even if certain factors (ie: material) are usually always given the same weight.
Basically, psychology plays a much larger role in chess than something liek Starcraft, I think. In chess, doubt and other emotions can creep in, while in Starcraft, there's always action, stuff to do. But in chess, what can you do while the opponent is pondering his move? Nothing, except just look at the board and try to guess what his move will be and what you will do, or to reanalyze a certain line on the board. Psychology also has to do with opening theory; you want to try and draw the opponent into a position that he's uncomfortable and one in which you've analzed. While it is possible for a top player to beat another in a line he's worked out at home (Kramnik-Leko 0-1 in their world championship match comes to mind, along with the famous Capablance-Marshall 1-0 game in which Marshall uncorks his Marshall gambit for the first time but Capa manages to beat him, a game which is considered by many to be one of the best of all time), it's a hard way to earn your bread money.
in short, what makes a rank #1 player better than another GM if both have enormous amounts of practice, theory, and experience?
The way they evaluate the position of each candidate move. Some GMs can pick out the best lines immediately, and some GMs take some more time. The better they are at picking out the key moves in a position, the stronger they will be, as they waste less time looking at inferior prospects. Naturally, the stronger player will sometimes also 'see' a move that another GM completely misses.
You're right about the rating changes being less for those players. For the majority of us elo points will change more significantly with wins or losses.
Even for us (even taking it down to 1000), it's only 7.5 for the rating change at K=15. AFAIK, the change is always the same, even if you're a 1000 taking on a 1000, or a 2500 taking on a 2500. But yeah, most chess federations and FIDE have different formulas.
I guess I didn't quite know what you were getting at with saying 'raw calculation.'
All I'm saying is that they evaluate positions entirely based on variations, not for anything subtle like 'ideas' or 'plans' in the position. They'll look at the line and assess the end. It's a result of pure number crunching, not a 'feel' for a position.
|
intrigue
Washington, D.C9934 Posts
=) thanks goldrush good post
|
The Elo (FIDE) development coefficient for rating change is:
25 for a player new to the rating list until he has completed events with a total of at least 30 games.
15 as long as a player`s rating remains under 2400.
10 once a player`s published rating has reached 2400, and he has also completed events with a total of at least 30 games. Thereafter it remains permanently at 10.
Which means a player with less than 30 Elorated games, gains 12,5 point winning against an opponent with the same rating.
|
Someone had requested to be on my MSN list but they are never on MSN. If you are actualy on MSN then my address is jose@m7financail.com
|
it's cool to see there are some starcraft players that also like chess; I personally feel that getting better at chess helped my starcraft and vice-versa (to an extent). I noticed that playing oodles of chess has improved other things I do also (strategy games, sports).
sounds like we have some good chess players here from the quality of the posts
|
Out of the six of you I've played, winning has been about as difficult as figuring out what to do after 1.g4 e5; 2. f3.
If perhaps there is actually another Starcraft player inherently good enough at chess to beat this one, I'll still be playing often for the next two months.
|
|
|
|
|
|
EPT
