74 ( +1 | -1 ) The SicilianAfter 1. e4 c5 2. Nf3, the 2,000,000-game database at -> chesslab.com appears to rank Black's candidate moves this way: #1) 2...g6: White won 34%, Black won 33%, Draw 33% #2) 2...e6: White won 35%, Black won 33%, Draw 32% #3) 2...d6: White won 37%, Black won 32%, Draw 31% #4) 2...Nc6: White won 38%, Black won 33%, Draw 29% With 2. ...g6, as you can see, the difference between White victories and Black victories is only 1%. * After more than 9 hours of analysis, Fritz8 appears to rank Black's candidate moves this way: #1) 2...Nc6: +0.41 #2) 2...g6: +0.49 #3) 2...d6: +0.50 #4) 2...e6: +0.50 Which source is more reliable, the 2,000,000-game database or one of the most credible chess software programs on the market? Perhaps it's a moot point. While the two appear to diverge re: the merits of 2...Nc6 and 2...e6, both seem to agree that 2...g6 is a strong candidate.
84 ( +1 | -1 ) Which just goes to show...... that White's initiative at the outset of the game does seem to be worth half a pawn, or maybe slightly less. The database stats seem to indicate that there is little to choose from the 4 moves suggested for Black (2...d6, 2...Nc6, 2...g6, 2...e6). The worst "performing" according to this database, 2...Nc6, the results can be expressed this way: 52.5%-47.5% in favour of White, a difference of 5%, or, if you like, 1 game in 20. Not exactly overwhelming, is it? But, in evaluating these kinds of stats, overall numbers might need to be taken into account as well. Now, suppose this database had, say, 80,000 games with 2...d6 and a similar number with 2...Nc6, but only 20,000 games with 2...g6. However significant you think the performance stats to begin with, you might conclude that the difference in numbers of games played renders problematical any statistical conclusions you might be inclined to draw. H'mmm
91 ( +1 | -1 ) While Fritz is an excellent program, there's a reason why computers use opening books. Pure analysis simply isn't enough in the early stages of the game. I certainly wouldn't draw any conclusions based on a few hundredths of a pawn difference at move 2.
It's also hard to draw any conclusions from game database statistics unless you know more about both the distribution of games and the strength of the players. However, I would expect the game database to be a better predictor of results in actual games between humans, because those are the games it models. For example, it might be that the advantages that Fritz sees in 2... Nc6 are difficult to actually convert for players who are not blessed with perfect tactical vision. Or (returning to my previous point) it might be that 2... Nc6 offers a short term tactical edge, but at a strategic cost that Fritz is incapable of calculating.
83 ( +1 | -1 ) Some other things to think about...How many of these games were Open Sicilians? How many Bb5 Sicilians showed up, Moscow after d6 and Rossolimo after Nc6 (or even sometimes after 2...e6 and 2...g6)? How many e6 games went into the Scheveningen (also a d6 line)? How many Nc6 games were Accelerated Dragon lines (which is basically what 2...g6 will be)? How many 'Accelerated Dragons' were demoted to Dragons (another 2...d6 line)? Another line, the Sicilian 4 Knight's variation (though not particularly popular) can occur after either 2...e6 or 2...Nc6. So, although White no longer has the option of the Grand Prix Attack, (s)he can still play one of the lesser Sicilian lines - a version of the Alapin, Rossolimo, Moscow, Chekhover, or (s)he can wait with 3. Nc3 to see what the opponent has in mind. So, I think you may be better off researching more specific lines, if you are stuck on using statistics.
52 ( +1 | -1 ) I dunno ...... Stats can be interesting, but arichallan does raise questions that are pertinent to whatever meaning one might be able to draw from these. Much also depends on what the stats are being used for. I would be disinclined to base my choice from the 4 2nd moves for black suggested in the opening post on the statistical results. Like spijker, I've always been a 2...d6 man... ;-)
64 ( +1 | -1 ) This is just too funnydid you really run Fritz for 9 hours on second move and hoped for any reliable results from that? Why didn't you start the beast on move one? Maybe it would have given you the holy truth about which is better. 1. e4 or 1. d4 :-) Guys, chess would be long dead if it was that simple. Whether you play 2. .. g6, 2. ..e6, 2. ..d6, or 2. ..Rc6 is a matter of personal choice and style of play you like. First try to understand the theory behind each of these moves and following variants, then choose one you like, .. and stick to it.
hehe... 9 hours, im still smiling. Poor old silicon beast, I can almost feel your pain! :-)) cheers magna68
65 ( +1 | -1 ) The trouble witha computer ranking things early: It doesnt know what the next move is. If running outside its opening book it will look for activity and play towards the center most likely. Perhaps Fritz sat there dreading 3.Nc3 for hours and gives a -.03 or worse to every response he can make!? (just Eg numbers here now). But then WT plays 3.d4 and Fritz goes "OMG Thanks Be! He didnt see Nc3, wow, oh wowsa wow ... I'm not losing anymore! Now I'm +.07 ... its almost over! ...and I can go back in the closet where it is safe and wont fear being used for drop-kick practice anymore today!" (Say Fritz ... who can cross the kitchen and out the back door faster?! heheh Yes, it IS YOU }B-)
100 ( +1 | -1 ) Hi i_play_slowly, however if you want to research database for opening, you need to research specific line and not only position after first two-three moves. Because after 1. e4 c5 2. Nf3 g6 there is a lot of possible transpositions, so your results are NOT from SICILIAN games only. For example after 1. e4 c5 2. Nf3 g6 3. c4 Bg7 4. d4 Qa5+ we are in Queen pawn game (A40) or after 1. e4 c5 2. Nf3 g6 3. d4 Bg7 3. d5 d6 4. c4 Nf6 5. Nc3 we are in Kings indian defence (E90). So if you want to explore sicilian defense you need to input few more "sicilian" moves. I try 1. e4 c5 2. Nf3 g6 3. d4 cd 4. Nd4 Nc6 (the position is knewn like starting position of accelerated dragon, but also with some transposition posibilities) and results are white won: 35%, black won 31%, draw 34%. Also the move 2...Nc6 is "worst", only because transposition problems. There are wery few sicilian main variatons where Knight does not go to c6. The true is if you play 1. e4 c5 2. Nf3 g6 3. d4 cd white can capture with queen 4. Qd4 and black rook is attacked, so black must play ...Nf6. This is probably not so dangerous for black but it is extra option for white, and black need to learn more theory.
36 ( +1 | -1 ) lol. I agree with magna, you may as well put Fritz 8 to work for 9 hours on whether e4 or d4 is best.
You can get a decent game from all of those moves. You are much better off playing a move which suits your style as well as knowing the theory behind it rather than playing a move because it is 0.01 of a pawn better than another move.
30 ( +1 | -1 ) magna68 and far1eyPerhaps more people would participate in this forum if there was less chance of being ridiculed--just a thought. I do appreciate your points, however, about the validity of all orthodox variations and the importance of simply sticking with a variation that matches one's temperament.