We all know that 1. A4 or 1. H4 are probably the most pointless opening first moves in chess... but have any of the computer whizzes here established which, on paper at least, is the BEST opening?
While we, as humans, like to have a repertoire, there must surely be one that 'rates' above others. I'd like to think the London System would rate highly but I don't know.
I suppose the only way to test this would be to see which opening as the highest computer evaluation after the end of that line, but I doubt there is an overwhelmingly superior opening - at least not to the extent that 1 a4 and 1 h4 are bad.
The only way to answer this is to look at a very large number of games and total up how many games white won with each given first move. That's been done; from wikipedia:
>New In Chess observed in its 2000 Yearbook that of the 731,740 >games in its database, White scored 54.8% overall; with the two >most popular opening moves, White scored 54.1% in 349,855 >games beginning 1.e4, and 56.1% in 296,200 games beginning >1.d4. The main reason that 1.e4 was less effective than 1.d4 was >the Sicilian Defence (1.e4 c5), which gave White only a 52.3% >score in 145,996 games.
You could do this with purely simulated games (i.e. an 'evaluation') but the answer should be the same.
I'm pretty confident that the 'perfect game' has d4 as its first move. Gotta wait till the game is solved to know for sure, though.
1. Nf3, (...)
2. h4!
I use to play this kind of openings for «tasting» my opponent repertoire and logic; being the idea to attack very quick the kingside. Because this rarely has been player along registered Chess History, I called it the “Zaitzev's attack”.
Not sure you should use the "!" though ;)
I think the "most perfect game" would start as a QGD..
Unless 1) chess is a draw, and only 1 move by white maintains the draw, or 2) chess is a win, and only 1 move by white maintains the win, then there is no "best" opening.
When players (human or computer) evaluate lines as "better" for one side without there being a forced win, that evaluation is completely relative. More precisely, because in one way or another that evaluation is based on the distribution of probabilities for winning/drawing/losing, that evaluation is relative to the particular skill of the player and that player's opponent.
Let's talk about engines for a second, since they correlate scoring chances with their evaluations (if not directly, then indirectly, since those weights are chosen based on testing, which is just seeing how the engine scores with different changes added/removed).
If between two playable positions, I have to choose one that is 0.6 according to the engine, and another that is 0.2, I don't actually have enough information to know which is the better choice.
Picking the line that is 0.6 is of no consolation if it's a position my opponent understands very well and can defend to a draw.
Conversely, the other line might "only" be 0.2 according to SF, but if it is only 0.2 if that same opponent finds 15 consecutive difficult moves, then that's a better try.
Flipping that scenario back on us, if the line that is 0.6 is 0.6 only if we are able to justify it with the tactical skill of Stockfish, and otherwise we lose, then it's a terrible choice.
It's all relative to our skills and our opponent's skills as long as there's no forced win. If I have to choose between going into an interesting (but drawn) pawn ending and a drawn version of R+B vs R, which is the better choice will depend on how well I and my opponent will play our respective sides.
If my opponent is someone who spends a lot of time looking at tricky pawn endings, I'm probably better off going into R+B vs R, while if my opponent has made it his life's goal to learn everything there is to know about R+B vs R, I'd better go into the pawn ending. Both of those are true regardless of what some statistical evaluation might be.
On the subject of statistics, it might end up that there is some line that is "best" on average. That would be a position/line that is more difficult than the alternatives for most players to meet, but that's just a generalization, and will have exceptions [especially once people start saying it's the "best" line, and everyone learns how to meet it :)].
That's one of the interesting things about chess. Even when perfect play is known, so long as we're playing against imperfect opponents, some choices will be better practically than others, and that's a relative measure that admits of different styles and approaches :)
play g3,
nf3 is good but looks difficult,
try g3 if you have not tried, its a solid opening and you make like it
Agree with #3.
1.d4 is probably the most promising move. Against it Gruenfeld/Nimzo/Queens Indian/(1...d5 maybe) are the best approaches, while Kings Indian/Dutch is so so (but funny).
1.c4 e5! (reversed Sicilian) causes no problems for black in the two mainlines ..Nc6,..d5 /..Bb4, ..c6.
1.Nf3 can transpose to 1.d4, question is if white has something against 1...c5.
All other first white moves are so so in the sense that white can not enforce draw.
if one is approaching the question from a theory standpoint, then there are only three possible results after a move - draw with perfect play, white wins with perfect play and black wins with perfect play.
so, it's probably the case that 1 d4 is just as good as 1 a3 :o (both moves lead to a draw with perfect play)
now, from a more practical standpoint, the most reliable method is probably recent statistics. and fashion will have its effects there...
computer evaluation is decent (computers are not weak in the opening nowadays), but the computer will eventually gravitate towards 0.00 if you are following some solid main line opening.
